[apps/web]: Plan

# AI Assisted 3D and AR Product Visualization for E Commerce

## Executive summary

This report assumes an initial **seller/provider-first** strategy: target **SMB Shopify merchants by default**, use a **SaaS + credits** revenue model by default, keep the AI generation layer **provider-agnostic**, and focus the first category wedge on **furniture, home, and accessories**. All four are assumptions, not fixed requirements. The most defensible hackathon product is not “the most advanced 3D generator”; it is the **fastest path from existing product images to compliant, mobile-safe, channel-ready product visualization**, with a human-review backstop and clear ROI instrumentation. That matters because Shopify installation and theme placement are still operationally non-trivial, Google’s 3D/AR surface is category- and country-limited, Amazon’s 3D/AR surfaces remain product-type constrained and mostly app-centric, and Trendyol’s public partner docs are still centered on standard product-image listings rather than first-party 3D/AR listing workflows. citeturn34view0turn34view1turn26view0turn26view1turn31view0turn32search0turn31view5

The strongest business case is “reduce uncertainty in online product evaluation.” Baymard found that **42% of users try to judge product size from product images**, and poor scale cues can lead to wrongful product rejection and abandonment. Shopify’s Rebecca Minkoff case study found that shoppers were **44% more likely to add to cart after interacting with a 3D model**, **27% more likely to place an order after interacting in 3D**, and **65% more likely to place an order after interacting in AR**. Amazon’s official 3D/AR program markets the same value proposition: higher confidence through “View in 3D,” “View in Your Room,” and virtual try-on experiences. citeturn15search0turn16search0turn31view0

For a hackathon, the right MVP is therefore narrow and disciplined: **catalog ingest, image-to-3D generation through an abstraction layer, human QA, a lightweight 3D viewer with hotspots and AR launch, Shopify publish flow, and analytics proving uplift**. The biggest mistakes to avoid are broad category support, perfect-realism ambitions, and heavy on-page payloads. Shopify still recommends small GLB assets in practice; its partner guidance targets roughly **4 MB** and says files **must not exceed 15 MB** for standard review scenarios, while the platform auto-optimizes above 15 MB and supports GLB/USDZ. Meanwhile, Google’s Core Web Vitals guidance still recommends **LCP within 2.5 seconds**, **INP below 200 ms**, and **CLS below 0.1**, so performance cannot be treated as a post-launch concern. citeturn35view0turn34view3turn5search1turn5search2

The clearest competitive gap is between enterprise-grade visualization vendors and raw AI generation tools. Enterprise tools are powerful but expensive, services-heavy, and often overbuilt for SMBs; AI generation tools are fast but weak on commerce workflow, QA, compliance, analytics, and channel publishing. The exploitable gap is: **“commerce-native, measured, compliant, and fast enough to deploy in a day.”** citeturn10search0turn10search1turn10search3turn19search7turn21view0turn9search2

## Assumptions and strategic framing

| Assumption | Default position | Why this is sensible for a hackathon |
|---|---|---|
| Target market | SMB Shopify merchants | Shopify provides a large merchant base, native product media handling, and a mature app-extension model, but still leaves enough integration pain for a differentiated app. citeturn33search3turn33search14turn34view0 |
| Pricing model | SaaS + credits | Generation cost is variable; subscription-only models struggle when 3D workloads spike per SKU. Meshy and Tripo both expose credit/pay-as-you-go logic, which matches a credit-based COGS model on your side. citeturn21view0turn9search2 |
| AI provider | Unspecified | Provider abstraction is essential because quality, latency, and cost vary by category and vendor, and no single provider is clearly dominant across all ecommerce product types. citeturn21view0turn9search2 |
| Initial category focus | Furniture/home + accessories | These categories benefit directly from scale, placement, and “view in room” behavior. They also align with Google’s current 3D/AR support for home goods and Amazon’s “View in Your Room” positioning. citeturn26view0turn26view1turn31view0turn32search0 |
| First channels | Shopify PDP + Google Merchant export | This combines owned PDP monetization with a discoverability channel that already supports 3D model links in limited markets/categories. citeturn34view0turn26view0turn26view1 |

The seller/provider-first strategy should follow a strict priority order:

| Priority | Who benefits first | Why this is the right sequence |
|---|---|---|
| First | Merchant/store owner | If merchants cannot publish quickly and measure ROI quickly, the product dies before shopper delight matters. Shopify merchants still need guidance to activate app blocks and supported themes. citeturn34view0turn34view1turn34view2 |
| Second | Platform/channel teams | Providers care about media quality, policy compliance, payload size, and support burden. Google, Amazon, Shopify, and Trendyol each impose different constraints. citeturn26view0turn26view4turn31view0turn31view4turn31view5 |
| Third | Shopper | A better viewer, AR, and hotspot storytelling increase confidence, but only after asset quality, speed, and compatibility are solved. citeturn15search0turn16search0turn31view0 |

```mermaid
erDiagram
    MERCHANT ||--o{ STORE : owns
    STORE ||--o{ PRODUCT : publishes
    PRODUCT ||--o{ VARIANT : has
    PRODUCT ||--o{ SOURCE_IMAGE : ingests
    PRODUCT ||--o{ THREE_D_ASSET : generates
    THREE_D_ASSET ||--o{ QA_REVIEW : undergoes
    THREE_D_ASSET ||--o{ CHANNEL_EXPORT : syndicates
    STORE ||--o{ WIDGET_INSTALL : configures
    SHOPPER ||--o{ ANALYTICS_EVENT : triggers
    PRODUCT ||--o{ ANALYTICS_EVENT : receives
    MERCHANT ||--o{ EXPERIMENT : runs
    EXPERIMENT ||--o{ ANALYTICS_EVENT : evaluates
```

## Pain point matrix

The matrix below enumerates the major direct and indirect pains across all requested domains. Closely related sub-pains are grouped into single rows so the list stays operational.

| ID | Domain | Pain point | Symptoms | Business impact | Root causes | How your product addresses it | Evidence |
|---|---|---|---|---|---|---|---|
| P1 | Merchant / Store Owner | 3D asset creation is too slow and too expensive | Merchants rely on photo shoots, agencies, or manual modeling; long turnaround before publish | Launch delays, higher content cost, missed campaigns; Shopify’s own partner workflow expects multiple quality photos and dimensions in mm, which is a high-friction input burden | Manual 3D production, lack of standardized inputs, no fast self-serve workflow | Image-to-3D generation from existing SKU media, optional dimension capture, and provider abstraction with human QA fallback | citeturn35view0turn35view1turn16search2 |
| P2 | Merchant / Store Owner | Unclear ROI makes adoption difficult | “Looks cool” but merchant cannot tie it to conversion or AOV | Budget objections; deprioritized initiative | Missing instrumentation and weak before/after reporting | Built-in experiment mode and ROI dashboard for interaction rate, ATC uplift, conversion uplift, and per-SKU impact | Rebecca Minkoff saw 44% higher add-to-cart after 3D interaction, 27% higher orders after 3D interaction, and 65% higher orders after AR interaction. citeturn16search0 |
| P3 | Merchant / Store Owner | Theme integration is still operationally messy | Merchant installs app but widget is not visible, misplaced, or unsupported by theme structure | Support tickets, churn during onboarding, poor install-to-publish conversion | App blocks are not present by default after install; themes must support `@app` blocks and JSON templates; statically rendered sections do not support app blocks | Guided Shopify onboarding, theme compatibility check, deep links, fallback script embed, and automatic “safe location” recommendations | citeturn34view0turn34view1 |
| P4 | Merchant / Store Owner | Variant-heavy catalogs break media logic | Color/material/size variants do not map cleanly to media states | Rework, shopper confusion, poor merchandising | Shopify variant media still supports only images; 3D models/videos cannot be used as variant media | Variant-aware asset mapping, grouped media states, material/color toggles inside viewer, and “best available variant” fallback | citeturn34view2turn34view4 |
| P5 | Merchant / Store Owner | Channel mismatches multiply asset ops | One asset works on Shopify but not on Google, Amazon, or marketplace surfaces | Extra operational work, duplicated assets, asset waste | Different file, product-type, policy, and surface constraints per channel | Channel capability matrix and export rules: Shopify-native, Google `virtual_model_link`, Amazon 3D upload path, Trendyol image-first fallback | citeturn26view0turn26view1turn31view0turn31view4turn31view5 |
| P6 | Merchant / Store Owner | Generated assets often fail quality review | Wrong scale, poor materials, distorted logos, bad UVs, blurry textures | Revisions consume time; lower trust in AI output | Weak mesh/material standards, lack of commerce-focused QA, no real-world scale checks | Human-in-the-loop QA workbench with scale, texture, material, and logo checks based on commerce asset standards | citeturn35view1turn35view0turn3search0 |
| P7 | Platform Providers | Platforms bear media-policy and quality enforcement costs | Feed disapprovals, listing errors, broken rich results, unsafe or misleading media | Support burden, lower marketplace quality, lower buyer trust | Low-quality images, conflicting data, missing attributes, unsafe URLs, unsupported categories | Export validation, preflight linting, metadata preservation, and safe-host checks before publish | Google says inaccurate or missing product data can cause disapprovals or display issues; safe-search standards also apply to `virtual_model_link`. citeturn26view4turn1search16 |
| P8 | Platform Providers | App/channel integration heterogeneity increases support load | Merchants need different instructions for Shopify themes, Amazon listing APIs, Trendyol uploads | Higher integration cost and longer time to value | No common surface for 3D/AR commerce export | Adapter layer per channel plus a single merchant-facing publishing workflow | Shopify app blocks/theme support, Amazon Listings Items API/media submissions, and Trendyol Partner API all differ materially. citeturn34view0turn31view4turn18search4turn18search3 |
| P9 | Platform Providers | Marketplace surfaces are limited or category-specific | 3D/AR works in some geographies or product types only | Coverage gaps reduce merchant confidence | Channel programs are staged and constrained | Capability-aware publishing; when unsupported, show enhanced 2D/lifestyle/hotspot output instead of a failed 3D promise | Google limits 3D/AR experiences to specific countries and primarily home goods/shoes; Amazon limits “View in Your Room” and other AR surfaces by type; Trendyol public docs remain image-centric. citeturn26view0turn26view1turn32search0turn31view5 |
| P10 | Shopper / Customer | Shoppers cannot judge size and context from photos alone | Questions about fit, scale, room placement, material detail | Wrong rejection of relevant products, hesitation, lower conversion | Static images lack “in-scale” context and spatial understanding | AR placement, dimension overlays, scale-calibrated viewer, and environment shots generated from the same asset | Baymard found 42% of users try to infer size from images, and poor “in scale” imagery causes misinterpretation and abandonment. citeturn15search0turn15search3 |
| P11 | Shopper / Customer | Shoppers distrust AI-enhanced media | “Will it really look like this?” skepticism | Lower conversion and potential complaints or regulatory exposure | Lack of disclosure, over-polished visuals, mismatch between render and delivered product | Trust microcopy, AI-assisted media disclosure, reference-photo view, and hotspot callouts labeled as illustrative vs exact | FTC requires advertising claims to be truthful and not misleading; Google requires AI-generated image metadata preservation in Merchant surfaces. citeturn4search2turn4search10turn26view2 |
| P12 | Shopper / Customer | Mobile/device support is inconsistent | AR button appears but fails, opens native app unexpectedly, or is unavailable | Frustration, bounce, lower engagement | ARCore device dependence, iOS/Android viewer differences, mixed in-browser vs native flows | Device detection, ar-capability checks, native Quick Look/Scene Viewer launch where appropriate, and graceful fallback to interactive 3D/poster | Apple Quick Look uses USDZ in Safari and other built-in apps; Google AR requires ARCore support; `<model-viewer>` supports WebXR/Scene Viewer/Quick Look modes. citeturn36view0turn2search1turn2search5turn25search2turn25search4 |
| P13 | Product Team / Operations | Asset governance becomes chaotic at catalog scale | Multiple versions, uncertain source of truth, no clean publish/unpublish history | Operational waste, bad exports, duplicated generation costs | No asset registry, no QA states, no versioning | Central asset registry with status, channel exports, and approval workflow | VNTANA’s product messaging exists largely because 3D workflows become unmanageable without purpose-built orchestration. citeturn19search7turn19search15 |
| P14 | Product Team / Operations | Feed and site content drift over time | Storefront, Google, marketplace, and asset metadata stop matching | Listing errors, customer confusion, policy risk | Manual field edits and siloed systems | Canonical product-to-asset mapping with export jobs and reconciliation checks | Google Merchant explicitly warns that conflicting data between feed and website, missing variant attributes, and low-quality images cause issues. citeturn26view4 |
| P15 | Product Team / Operations | Teams lack instrumentation across the 3D funnel | Can measure page views but not 3D load success, hotspots, AR launch, or fallback rate | Impossible to improve the product with confidence | Standard ecommerce analytics stops at product/cart/checkout | Viewer event schema mapped to GA4, Shopify pixels, and custom app events | Shopify exposes standard customer events like `product_viewed`, `product_added_to_cart`, and `checkout_started`; custom app events must be published separately. GA4 supports ecommerce plus custom events. citeturn7search0turn7search1turn7search8turn7search9turn7search24 |
| P16 | Legal / Privacy | Consent, tracking, and AI-content handling can become non-compliant | Pixel data collected without proper notice; AI images stripped of metadata | Regulatory risk and merchant distrust | Weak privacy design and careless asset transformation pipeline | Consent-aware analytics, minimal data collection, retention controls, DPA, and metadata-preserving media pipeline | GDPR applies to personal data including online identifiers; CCPA requires notice at or before collection; Google requires AI-generated image metadata tags to be preserved. citeturn30search7turn4search12turn4search1turn4search9turn26view2 |
| P17 | Legal / Privacy | Ownership and authorship of AI-generated 3D outputs is unclear | Merchant asks: “Do I own this asset? Can I reuse it? Can I copyright it?” | Contracting friction, enterprise security review delays | AI output law and vendor contracts remain nuanced | Provider-agnostic contract layer, provenance logs, and explicit merchant terms that distinguish AI output from human-authored edits | The U.S. Copyright Office states generative AI outputs are protectable only where sufficient human authorship exists; prompts alone are generally insufficient. citeturn17search0turn17search1turn17search7 |
| P18 | Performance / Infrastructure | 3D payloads can break mobile performance | Slow LCP, janky interaction, GPU overload, failed low-end device experience | Lower SEO resilience, lower conversion, more abandonment | Large GLB textures/meshes, no lazy load, weak compression, over-eager hydration | Poster-first loading, click-to-load option, KTX2/Draco/Meshopt where beneficial, dynamic scaling, CDN edge delivery, strict asset budgets | Google’s good thresholds remain LCP ≤2.5s, INP <200ms, CLS <0.1; `<model-viewer>` recommends poster/lazy loading and warns that DRACO’s decoder is large enough to be a net loss on small files. citeturn5search1turn5search2turn36view2turn3search1turn3search2 |
| P19 | Performance / Infrastructure | Generation pipelines are asynchronous and bursty | Jobs queue up; cost and latency spike during bulk uploads | Merchant frustration and runaway COGS | Third-party generation APIs, rate limits, per-call pricing, unpredictable SKU bursts | Queue-based architecture, retry policy, budget caps, and progressive publish states | Meshy prices generation per call/credit; Tripo uses pay-as-you-go credits; Cloudflare Queues are purpose-built to buffer asynchronous work. citeturn21view0turn9search2turn23search4turn23search0 |
| P20 | Business / Go to Market | Buyer education is still required | Merchant understands photos and video, not “why 3D now?” | Longer sales cycle, especially above SMB | Market still spans enterprise 3D DAM, configurators, and AI novelty tools | Position as conversion-confidence infrastructure, not “3D for 3D’s sake” | Shopify frames 3D commerce around better confidence/conversion, not novelty; average product-page UX is still mediocre, leaving room for measurable improvement. citeturn16search1turn16search6turn15search10 |
| P21 | Business / Go to Market | Platform-provider partnerships require proof, not just a demo | Agencies/platform teams care about support cost, compatibility, and merchant outcomes | Hard to win app-store traction or referrals without hard data | Few startups bring provider-facing analytics and compliance reporting | Add provider-ready reporting: activation, asset pass rate, performance delta, merchant ROI snapshots | This is an inference from the reviewed Shopify/Amazon/Google integration surfaces and their operational constraints. citeturn34view0turn26view4turn31view4 |
| P22 | Hackathon constraints | You cannot solve perfect image-to-3D quality in eight weeks | Inconsistent results across reflective, transparent, deformable, or highly detailed products | Demo risk, overpromising risk | Model generation quality varies by category and provider; asset cleanup still matters | Limit MVP to furniture/home/accessories, add manual QA, and publish “AI-assisted” rather than “fully automatic” | Shopify and Khronos guidance both emphasize real-world scale, clean UVs, optimized textures, and model review; these are not reliably auto-solved in all categories today. citeturn35view1turn3search0turn3search4 |
| P23 | Hackathon constraints | Lack of real merchant data makes prioritization noisy | Team builds impressive demo features with low operational value | Lower odds of post-hackathon traction | No pilot stores or limited SKU corpus | Ship with 2–3 pilot merchants and a repeatable measurement plan | This is an execution inference, but it follows directly from the need to prove merchant ROI and channel compatibility early. citeturn16search0turn26view0turn34view0 |

## Feature priorities and competitive landscape

The feature roadmap should begin with merchant activation and provider/channel reliability, then expand into shopper delight.

| Feature | Maps to pain IDs | Why it matters | Complexity | Est. effort | MVP or post-MVP |
|---|---|---|---|---:|---|
| F1 Merchant onboarding wizard with catalog import | P1, P3, P20, P23 | Reduces install-to-first-value time | Medium | 1.5 pw | MVP |
| F2 Provider-agnostic image-to-3D generation adapter | P1, P19, P22 | Avoids lock-in and lets you swap quality/cost profiles | High | 2.0 pw | MVP |
| F3 Human QA workbench with approve/reject reasons | P5, P6, P13, P22 | Prevents low-trust AI output from reaching shoppers | Medium | 1.5 pw | MVP |
| F4 Shopify app block + fallback script embed | P3, P8, P21 | Gives the fastest path to deploy on most merchant themes | Medium | 1.5 pw | MVP |
| F5 Asset optimization pipeline with poster-first loading | P18, P19 | Protects mobile UX and budget | Medium | 1.0 pw | MVP |
| F6 3D viewer with hotspots, dimensions, and AR launch | P10, P11, P12 | Core shopper value proposition | Medium | 1.5 pw | MVP |
| F7 Device/channel capability detection and fallback UX | P5, P9, P12, P18 | Prevents broken AR promises | Medium | 0.75 pw | MVP |
| F8 Analytics schema + experiment mode + ROI dashboard | P2, P15, P20, P23 | Converts “cool demo” into measurable product value | Medium | 1.5 pw | MVP |
| F9 Google Merchant export for `virtual_model_link` | P5, P7, P9 | Strong seller/provider differentiator for supported categories | Medium | 1.0 pw | MVP |
| F10 Variant-aware asset mapping and material toggles | P4, P14 | Important for real catalog utility beyond single-SKU demos | High | 1.5 pw | Post-MVP |
| F11 Marketplace adapters for Amazon and Trendyol | P5, P8, P9, P21 | Needed for provider-first GTM beyond Shopify | High | 2.0 pw | Post-MVP |
| F12 Compliance layer for consent, AI metadata, provenance logs | P16, P17 | Needed for trust and enterprise readiness | Medium | 1.0 pw | MVP-lite, full post-MVP |
| F13 Bulk ingestion, queue controls, retries, and budget guardrails | P19, P22 | Required once real usage appears | Medium | 1.0 pw | Post-MVP |
| F14 Merchant benchmark insights and category templates | P2, P20 | Strong long-term moat if you can compare “3D-ready PDP quality” across stores | High | 2.0 pw | Post-MVP |

A pragmatic MVP architecture should use open, commerce-friendly formats and thin viewer layers.

| Layer | Recommended choice | Pros | Cons | Cost / commercial note |
|---|---|---|---|---|
| Canonical 3D format | **glTF/GLB** | Royalty-free, efficient for runtime delivery, interoperable across web engines; effectively the default web-commerce target | Needs optimization discipline for mobile | Use GLB as canonical master; generate USDZ for Apple surfaces. citeturn2search3turn2search6turn3search9 |
| Apple AR format | **USDZ** | Native Quick Look support in Safari, Messages, Mail, Notes, iPhone/iPad/Vision Pro | Apple-specific; not ideal as canonical authoring format | Necessary for iOS-native AR surface. citeturn36view0 |
| Viewer | **`<model-viewer>`** | Web-friendly, AR-capable, supports WebXR/Scene Viewer/Quick Look patterns, hotspots/annotations, poster/lazy-load techniques | Requires careful compression and fallback logic | Best fit for hackathon speed and embeddability. citeturn25search1turn25search2turn36view2 |
| Compression / optimization | **glTF-Compressor, KTX2, Meshopt or Draco** | Official Khronos ecosystem, geometry + texture compression | Decoder overhead can outweigh file-size savings on small assets; needs tuning | Use per-asset heuristics, not blanket compression. citeturn3search1turn3search2turn36view2 |
| Generation provider | **Meshy primary, Tripo alternative** | Fast API integration, credit-based economics, good for hackathon velocity | Output quality varies by product type; cleanup may still be needed | Meshy exposes per-call credit pricing; Tripo uses pay-as-you-go credits. citeturn21view0turn9search2 |
| Edge compute / orchestration | **Cloudflare Workers + Queues** | Good async pattern, low-friction edge execution, no egress charge on R2, request-based pricing | More low-level than full-stack managed app platforms | Workers Paid plan starts at $5/month; Queues bill by operations. citeturn23search1turn23search0turn23search4 |
| Object storage / CDN | **Cloudflare R2** | S3-compatible, zero egress, low storage cost | Need your own metadata/index logic unless paired with DB | Published pricing includes $0.015/GB-month standard storage and no egress fees. citeturn21view2 |
| Dashboard hosting | **Vercel** for merchant/admin UI | Extremely fast FE iteration, CI/CD, spend controls | Serverless compute overages can climb if backend logic is heavy | Pro starts at $20/month plus usage. citeturn21view3 |
| App backend alternative | **Supabase** | Auth, Postgres, storage, edge functions, S3-compatible storage, fast setup | More centralized than edge-first, can feel heavier for simple widgets | Pro starts at $25/month; edge function invocations are $2 per 1M above quota. citeturn22search0turn22search1turn22search6turn22search3 |
| Product analytics | **PostHog + GA4 + Shopify Web Pixels** | PostHog for product loops, GA4 for commerce reporting, Shopify events for storefront truth | Requires event governance to avoid duplication | PostHog has a generous free tier incl. 1M analytics events; GA4 ecommerce and Shopify standard events are well documented. citeturn21view4turn7search1turn7search0 |
| Platform integrations | **Shopify first; Google Merchant second; Amazon/Trendyol next** | Maximizes speed to value and control | Marketplace surfaces are more constrained and slower to operationalize | Aligns with the strongest SMB motion. citeturn34view0turn26view0turn31view4turn31view5 |

A useful rule for the stack decision is simple: keep the **viewer simple**, the **pipeline asynchronous**, and the **data model canonical** around product → asset → export → event.

### Competitive audit

| Competitor / tool | Core motion | AI image→3D | Viewer / AR | Configurator | Channel / commerce ops | Likely weakness you can exploit |
|---|---|---|---|---|---|---|
| **Shopify native 3D support** | Native product media on Shopify | No native generation pipeline | Yes, via Shopify product media and compatible themes | Limited | Good for Shopify-only media handling | No end-to-end generation, onboarding guidance, QA, or ROI analytics. citeturn34view2turn34view3turn20search6 |
| **Amazon 3D & AR** | Marketplace-native 3D/AR surfaces | Upload/create within Amazon flow | Yes, but mostly app-based customer surfaces | Try-on niche | Strong marketplace surface where supported | Not an owned-store solution; product-type constraints; weak multi-channel story for merchants. citeturn31view0turn32search0 |
| **Zakeke** | 3D customization and configuration | AI creative layer, but generation is not its main moat | Yes | Strong | Integrates with major commerce platforms | Best for custom/configurable products, less focused on fast 2D-catalog-to-3D onboarding for SMB merchants. citeturn10search0turn10search4turn10search18 |
| **Threekit** | Enterprise visualization + configuration + guided selling | Not primarily AI image→3D self-serve | Yes | Very strong | Deep enterprise workflows | Powerful, but enterprise-heavy and services-oriented for SMB hackathon positioning. citeturn10search1turn10search5turn10search19 |
| **Emersya** | Interactive 3D/AR authoring platform | Not core | Yes | Strong | Rich storytelling and no-code authoring | Strong platform, but your opening is faster AI-assisted asset creation plus merchant ROI instrumentation. citeturn10search3turn10search7 |
| **VNTANA** | 3D orchestration / optimization / DAM | Not core generation moat | Yes | Moderate | Strong transformation/governance | Excellent enterprise orchestration; opportunity is SMB simplicity and self-serve workflow. citeturn19search7turn19search12turn19search14 |
| **Meshy** | AI 3D generation platform | Strong | Limited commerce workflow | No | API-first creation | Great raw generation, weak commerce-native publishing/governance/analytics. citeturn21view0turn9search11 |
| **Tripo** | AI 3D generation platform | Strong | Limited commerce workflow | No | API/pay-go generation | Similar gap: generation is strong, commerce operations are not the main product. citeturn9search2turn9search6turn9search10 |

### Gaps you can exploit

| Gap | Why it matters |
|---|---|
| **Self-serve SMB onboarding from existing catalog images** | Most enterprise vendors assume existing 3D/CAD maturity or services budgets; raw AI vendors assume creator workflows, not merchant workflows. citeturn10search1turn19search7turn21view0 |
| **Channel compliance + fallback publishing** | Merchants do not want to think about GLB vs USDZ, app block support, Google country limits, or marketplace media constraints. citeturn34view0turn26view0turn31view0turn31view5 |
| **Commerce-specific QA and asset governance** | “Looks plausible” is not good enough for product detail pages; scale, material truthfulness, and logo fidelity matter. citeturn35view1turn3search0 |
| **ROI analytics baked into the product** | Every merchant asks the same question: “Did this increase add-to-cart, conversion, or reduce confusion?” Many competitors mention outcomes; fewer operationalize them in the day-to-day product. citeturn16search0turn10search1turn19search7 |
| **Provider-first reporting** | Agencies, platform teams, and merchant-success teams need fewer tickets, faster activation, and fewer broken embeds. Your product can become “safe to recommend.” citeturn34view0turn31view4 |

### Suggested positioning statements

1. **Turn existing product images into commerce-ready 3D and AR in a day, not a quarter.**
2. **Built for merchants first: publish fast, prove ROI, and stay within platform rules.**
3. **Not just AI 3D generation—AI-assisted asset production with QA, channel exports, and measurement.**
4. **The simplest path from static PDPs to measurable confidence commerce.**
5. **Provider-agnostic 3D commerce infrastructure for SMB stores, not enterprise services bloat.**

## Delivery, integrations, and go to market

### Phase plan

| Phase | Purpose | Milestones | Deliverables | Success metrics | Main risks | Mitigation |
|---|---|---|---|---|---|---|
| Discover and constrain | Lock the wedge and merchant story | 3 pilot merchants, 20–50 SKUs, one supported theme family, one category template | Problem briefs, sample dataset, success criteria, event schema v1 | 3 merchants agree to pilot; one real SKU ready end-to-end | Scope sprawl | Freeze category + channel list early |
| Build the production-shaped MVP | Deliver the shortest path to publish | Image ingest, generation job, QA, viewer, Shopify publish, analytics | Working Shopify app/widget, asset registry, viewer, metrics dashboard | <30 min from install to first live model for demo merchant, excluding human QA | Generation quality inconsistency | Human approval gate; use best-effort category filters |
| Add channel credibility | Show provider-first readiness | Google export, performance budget, compliance basics | `virtual_model_link` export, AI-content handling, CWV audit | Asset publish succeeds on Shopify; Google-ready feed object produced | Channel-specific edge cases | Make exports explicit and capability-gated |
| Pilot and investor readiness | Turn demo into believable business | Case-study data, onboarding docs, GTM assets, pitch narrative | Pilot report, deck, KPI baseline, pricing tests | 1–3 pilots, measurable interaction and conversion proxy data | No real uplift data yet | Use interaction/ATC leading metrics if purchase volume is low |

### Sprint-level eight-week plan

| Week | Goal | Key tasks | Owners |
|---|---|---|---|
| Week one | Frame the wedge | Pilot merchant interviews, SKU collection, architecture, quality rubric, success metrics | PM, Design, Growth, ML/3D |
| Week two | Build ingestion and generation path | Product import, image upload, provider adapter abstraction, generation queue, DB schema | BE, ML/3D, PM |
| Week three | Build merchant review flow | QA workbench, approve/reject states, asset versioning, thumbnail/poster generation | BE, FE, Design |
| Week four | Build the storefront experience | `<model-viewer>` integration, hotspots, dimensions overlay, AR entry, mobile fallback UX | FE, Design, ML/3D |
| Week five | Shopify-first publishing | Theme app extension, app block placement, fallback embed, product mapping, install guidance | FE, BE, PM |
| Week six | Measure and harden | PostHog + GA4 + Shopify events, experiment mode, performance budgets, compression pipeline | FE, BE, Growth |
| Week seven | Add channel credibility | Google export, compliance checks, AI disclosure UX, Amazon/Trendyol adapter stub or export plan | BE, PM, Growth |
| Week eight | Pilot, polish, pitch | Pilot deployment, bug bash, case study draft, pricing page, investor memo, demo script | PM, FE, BE, Design, Growth |

```mermaid
gantt
    title Eight-week hackathon and pilot plan
    dateFormat  YYYY-MM-DD
    section Discovery
    Merchant interviews and SKU selection :a1, 2026-05-19, 7d
    KPI and quality rubric                :a2, 2026-05-19, 7d
    section Core pipeline
    Ingest + provider abstraction         :b1, 2026-05-26, 14d
    QA workbench + asset registry         :b2, 2026-06-02, 14d
    section Storefront
    Viewer + hotspots + AR                :c1, 2026-06-09, 14d
    Shopify app block + embed fallback    :c2, 2026-06-16, 14d
    section Measurement and channels
    Analytics + experiments + budgets     :d1, 2026-06-23, 7d
    Google export + compliance basics     :d2, 2026-06-30, 7d
    section Pilot and pitch
    Pilot rollout and bug bash            :e1, 2026-07-07, 7d
    Deck, memo, demo, GTM polish          :e2, 2026-07-07, 7d
```

### Seller onboarding flow recommendation

| Step | Merchant experience | Product requirement |
|---|---|---|
| Install | “Connect your store and select products to 3D-enable.” | Shopify app install; theme support check; fallback embed available if app blocks are unsupported or inconvenient. citeturn34view0turn34view1 |
| Select products | Merchant chooses 1–20 SKUs | Pull existing Shopify media and variant data |
| Generate | “We’ll turn your existing images into a draft 3D asset.” | Queue job, provider selection, budget estimate |
| Review | Merchant approves geometry/material/scale | QA panel with yes/no checklist based on Shopify/Khronos best practices |
| Publish | “Add to product page” | App block insertion instructions or auto-safe embed install |
| Measure | “See interaction, AR launch, add-to-cart lift” | Dashboard with 7-day and 28-day windows |

### Platform provider integration checklist

| Platform | Checklist | Critical constraints |
|---|---|---|
| Shopify | Theme app extension, app block support, product→asset mapping, Web Pixel event plan, 3D model/file storage policy, supported theme check | App blocks are not added by default; theme must support `@app`; 3D/video not valid as variant media; product media limits and plan-specific upload/storage limits apply. citeturn34view0turn34view1turn34view2 |
| Google Merchant | Produce hosted public model URL, populate `virtual_model_link`, keep AI metadata intact, validate feed/site consistency | Current 3D/AR support is limited by geography and category; AR in Shopping ads remains constrained; AI-generated image metadata must be preserved. citeturn26view0turn26view1turn26view2turn26view4 |
| Amazon | Use Seller Central 3D upload/media flow or listing workflows, handle product-type restrictions, ensure mobile-app expectation is clear | Listings Items API is 1x1 oriented for many tasks; product type support varies; customer 3D/AR surfaces are primarily app-based and category-limited. citeturn31view0turn31view4turn32search0 |
| Trendyol | Use Partner API auth, HTTPS image URLs, barcode/product mapping, image-size compliance; treat 3D as off-platform/PDP enhancement until first-party support is confirmed | Public docs reviewed are image/listing-centric: max 8 images, HTTPS URLs, 1200×1800 at 96 dpi; no equivalent public first-party 3D/AR seller workflow was identified in the reviewed docs. This last point is an inference from the documentation reviewed. citeturn18search3turn18search4turn31view5 |

### Pricing model options

| Model | Best fit | Pros | Cons | Recommendation |
|---|---|---|---|---|
| Flat SaaS only | Low-volume merchants | Easy to explain | Mismatched to variable generation cost | Not ideal as default |
| Credits only | Creator or API users | Matches COGS | Can feel punitive for merchants | Good for API-only add-on |
| **SaaS + credits** | SMB and agencies | Predictable base + variable usage | Slightly more complex | **Default recommendation** |
| White-glove setup + SaaS | Agencies / mid-market | Faster ACV growth | Services-heavy, less scalable | Use selectively post-MVP |

### Partner outreach tactics

| Target partner | Why they matter | Outreach angle |
|---|---|---|
| Shopify agencies / partners | Control store setup and theme placement | “Reduce custom theme work and show measurable PDP uplift.” |
| Product photography studios / 3D freelancers | Already serve the same merchants | “Your service becomes faster with AI-assisted first draft + QA workflow.” |
| Feed / PIM / DAM consultants | Own catalog and channel pain | “Add 3D/AR readiness as a measurable channel export layer.” |
| Marketplace integrators in Türkiye | Useful for Trendyol/marketplace expansion | “Offer image-first fallbacks now, marketplace-rich media later.” |
| Merchant success teams at platforms | Provider-first wedge | “Fewer broken embeds, better media quality, documented adoption metrics.” |

## UX, analytics, compliance, and investor brief

### UX copy examples

**Merchant onboarding**

> **Turn your existing product images into interactive 3D.**  
> Start with 1 product. Review the draft. Publish when it looks right.

> **What you’ll need**  
> At least 4 clear product images. Front, back, side, and a detail shot if possible.  
> Optional: dimensions for more accurate scale.

> **Quality note**  
> This asset is AI-assisted and reviewed before publishing. You stay in control of what goes live.

> **Publish options**  
> Add to your Shopify product page now.  
> Export a Google-ready model link later.

**Embed snippet**

```html
<script
  src="https://cdn.yourapp.com/widget.js"
  data-store-id="store_123"
  data-product-id="gid://shopify/Product/1234567890"
  data-mode="3d-ar"
  data-hotspots="true"
  data-poster="auto">
</script>
```

If you want a direct viewer pattern, a `<model-viewer>`-based embed is the fastest path for a hackathon because it already supports interactive 3D, AR paths, annotations, and poster/lazy-load patterns on the web. citeturn25search1turn25search2turn36view2

```html
<model-viewer
  src="/assets/chair.glb"
  ios-src="/assets/chair.usdz"
  ar
  camera-controls
  poster="/assets/chair-poster.jpg"
  alt="Interactive 3D model of the oak lounge chair">
  <button slot="ar-button">View in your space</button>
  <button class="hotspot" slot="hotspot-1" data-position="0 0.5 0.1">
    Solid oak frame
  </button>
</model-viewer>
```

**Shopper-facing hotspots and trust microcopy**

- **Solid oak frame** — “Real-world dimensions preserved from the product spec.”
- **Seat depth** — “Tap to compare to your room or existing furniture.”
- **Fabric texture** — “Close-up view for weave and finish detail.”
- **Assembly note** — “Shown fully assembled for visualization.”
- **Color note** — “Lighting and screens may affect perceived color slightly.”
- **Trust microcopy** — “Interactive model for exploration. Check listed dimensions and materials before purchase.”
- **Fallback microcopy** — “AR not available on this device. You can still rotate the product in 3D.”

### KPI framework and analytics schema

The KPI stack should separate **merchant activation**, **shopper engagement**, **commercial outcome**, and **operational efficiency**.

| KPI group | Core KPI | Why it matters |
|---|---|---|
| Merchant activation | Install→first published model time | Measures real setup friction |
| Merchant adoption | % of installed stores publishing at least one model | Measures actual value, not vanity installs |
| Shopper engagement | 3D interaction rate, hotspot open rate, AR launch rate | Measures whether the experience is discoverable and useful |
| Commercial outcome | Add-to-cart uplift, PDP conversion uplift, assisted revenue | Ties the product to spend justification |
| Operations | Approval rate, revision rate, generation latency, cost per approved SKU | Protects margins and team bandwidth |
| Performance | 3D-enabled PDP CWV pass rate | Prevents the feature from harming the store |

| Event | Trigger | Key properties | GA4 mapping | Shopify mapping |
|---|---|---|---|---|
| `product_viewed` | Shopper visits PDP | product_id, variant_id, category | Standard GA4 ecommerce context | Native Shopify standard event `product_viewed` citeturn7search0turn7search1 |
| `app_3d_widget_loaded` | Viewer successfully initializes | product_id, asset_id, load_ms, fallback_state | Custom event | Custom app pixel event; prefix with app name per Shopify guidance citeturn7search9turn7search24 |
| `app_3d_interaction_started` | First rotate/zoom interaction | product_id, device_type, dwell_ms | Custom event | Custom Shopify pixel event |
| `app_hotspot_opened` | Shopper opens hotspot | hotspot_id, product_id | Custom event | Custom Shopify pixel event |
| `app_ar_launched` | Shopper taps AR | platform, ar_mode, product_id | Custom event | Custom Shopify pixel event |
| `product_added_to_cart` | Add to cart from PDP | product_id, variant_id, quantity | GA4 `add_to_cart` | Native Shopify `product_added_to_cart` citeturn7search2turn7search1 |
| `checkout_started` | Shopper starts checkout | cart_value, item_count | GA4 `begin_checkout` | Native Shopify `checkout_started` citeturn7search8turn7search1 |
| `purchase` | Order complete | revenue, items, model_interacted_before_purchase | GA4 `purchase` | Native purchase pipeline / order reconciliation citeturn7search10turn7search1 |
| `app_asset_generation_started` | Merchant starts generation | provider, sku_count, estimated_cost | Internal ops event | N/A |
| `app_asset_generation_completed` | Job done | provider, latency_s, quality_score | Internal ops event | N/A |
| `app_publish_clicked` | Merchant publishes asset | channel, product_id, asset_id | Internal ops event | N/A |
| `app_quality_rejected` | QA rejects asset | reject_reason, provider, category | Internal ops event | N/A |

**Suggested dashboard mockups**

| Dashboard | Primary widgets | Suggested charts |
|---|---|---|
| Merchant activation | installs, published stores, time to first publish | Funnel: install → select products → generate → approve → publish |
| Content ops | generation volume, approval rate, rejection reasons, cost per approved SKU | Stacked bar by rejection reason; heatmap by provider × category |
| Shopper engagement | viewer loads, interaction rate, hotspot usage, AR launches | PDP engagement timeline; device split donut; AR launch rate by product |
| Commercial impact | ATC uplift, conversion uplift, assisted revenue, AOV delta | A/B uplift chart; cohort line chart by product category |
| Performance | LCP/INP/CLS for 3D vs control PDPs | Box plot by device class; origin-level CWV pass rate |
| Channel exports | Shopify published, Google exported, marketplace-ready backlog | Status board by channel |

### Privacy, accessibility, and mobile performance release gates

| Area | Requirement | Concrete implementation check | Release gate |
|---|---|---|---|
| Privacy | Data minimization | Do not collect more than needed for analytics and QA; separate merchant/admin data from shopper event data | Pass only if event catalog and retention policy are documented. GDPR’s data framework and online identifiers are relevant here. citeturn30search7turn4search12 |
| Privacy | Notice and consent | Respect consent before pixels/session replay; provide notice at or before collection where applicable | Pass only if consent management and event suppression are tested. citeturn4search1turn4search9turn7search6 |
| AI content | Metadata preservation | Preserve IPTC `DigitalSourceType` tags for AI-generated commerce imagery; do not strip tags during optimization | Pass only if media pipeline preserves metadata for applicable image outputs. citeturn26view2 |
| Truthfulness | Do not imply exactness where approximate | Mark AI-assisted visuals clearly; do not overclaim material/fit/size precision | Pass only if trust microcopy and QA thresholds are enabled. FTC truth-in-advertising applies. citeturn4search2turn4search10 |
| IP / authorship | Asset provenance logging | Record provider, prompt/input images, human edits, publish approver | Pass only if provenance exists for every published asset. citeturn17search0turn17search7 |
| Accessibility | Provide non-3D equivalent information | Keep product images, dimensions, and text specs available even if viewer fails or is skipped | Pass only if PDP remains usable without 3D |
| Accessibility | Keyboard and pointer alternatives | If rotation or hotspot access depends on dragging, provide button or single-pointer alternatives | Pass only if viewer controls and hotspot navigation work without drag. WCAG 2.5.7 requires a non-drag alternative where dragging is non-essential. citeturn27search1turn27search8 |
| Accessibility | Touch target sizing | Hotspots and control targets at least 24×24 CSS px or valid spacing exception | Pass only if mobile hotspot targets meet WCAG 2.5.8. citeturn28search0turn28search1 |
| Accessibility | Focus visibility | Hotspots/buttons visibly focused and not obscured | Pass only if keyboard QA on mobile/desktop is clean. citeturn4search3turn27search11 |
| Performance | Asset budgets | Default target GLB around 4 MB when possible; avoid routine publish above 15 MB for Shopify-first workflow unless aggressively optimized | Pass only if asset budget report is clean. citeturn35view0turn34view3 |
| Performance | Lazy load and poster | Do not auto-hydrate heavy 3D before user intent on weak/mobile contexts | Pass only if poster-first path exists. citeturn36view2 |
| Performance | CWV thresholds | Monitor LCP, INP, CLS on 3D-enabled PDPs | Pass only if origin/page remains within acceptable thresholds in real-user monitoring. citeturn5search1turn5search4 |

### Investor-readiness brief

**One-page thesis**

The company should be pitched as **confidence-commerce infrastructure** for merchants, not as a stand-alone 3D novelty tool. Online shoppers routinely struggle to infer real-world size, placement, and product detail from static media, while merchants still face high cost and slow workflows if they want to deploy 3D and AR well. Existing enterprise visualization platforms are powerful but expensive and operationally heavy; pure AI 3D generators are fast but not commerce-native. The opportunity is a product that turns existing catalog media into measurable, channel-ready visualization with human QA, performance safety, and revenue instrumentation. The strongest early wedge is SMB Shopify merchants in furniture/home/accessories, where scale and room placement are most valuable and where Shopify, Google, and Amazon already provide enough official surface area to validate demand. citeturn15search0turn16search0turn33search3turn26view0turn31view0

**Illustrative TAM / SAM / SOM assumptions**

These are explicit modeling assumptions, not official market counts.

| Layer | Assumption logic | Illustrative figure |
|---|---|---:|
| TAM | Commerce merchants who could benefit from confidence-centric 3D/AR product media across owned stores and major channels. Shopify alone says **millions of businesses** use the platform in **175+ countries**, so the relevant merchant universe is clearly in the millions even before marketplaces are counted. citeturn33search3turn33search14 | 2.0M merchants (illustrative subset assumption) |
| SAM | SMB Shopify merchants in furniture/home/accessories, with enough image quality and SKU density to justify 3D/AR, initially concentrated in Shopify-friendly and Google-supported markets | 60k merchants |
| SOM | Merchants reachable within three years through Shopify app distribution, agencies, and pilot-led referrals | 1.5k merchants |

**Revenue model options**

| Revenue stream | Comment |
|---|---|
| Core SaaS subscription | Access to onboarding, publishing, analytics, QA workflow |
| Usage credits | Charged on generation/export volume; aligns to underlying provider cost curve |
| Agency / partner seats | Multi-store workflows, centralized billing, white-label reporting |
| Professional services | Optional onboarding, catalog cleanup, template setup, and asset QA for larger accounts |

**Three-year roadmap summary**

| Horizon | Product focus | Commercial focus |
|---|---|---|
| Year one | Shopify-first workflow, furniture/home templates, QA workbench, ROI analytics, Google export | Pilot merchants, Shopify agencies, early app-store motion |
| Year two | Variant logic, Amazon/marketplace adapters, bulk ops, stronger governance, richer experimentation | Mid-market brands, feed/PIM partners, agency channel expansion |
| Year three | Multi-modal input pipeline, stronger automated QA, deeper DAM/PIM integrations, category-specific models | Platform/provider partnerships, enterprise upsell, API revenue |

**Open questions and limitations**

- The exact public **Grand View Research** market-size figure was not re-opened in this finalized pass, so the investor section uses a conservative **bottom-up merchant assumption model** instead of quoting a third-party dollar-market number.
- In the reviewed public documentation, **Trendyol** exposes strong listing/image APIs but no comparable first-party 3D/AR seller workflow; treat that as a documentation-based inference, not a definitive statement that internal/private support does not exist. citeturn31view5turn18search4
- Actual quality, latency, and margin depend heavily on the chosen **image→3D provider** and the category wedge; reflective, translucent, deformable, and heavily branded products remain the highest-risk asset classes. citeturn21view0turn35view1

Provide feedback

Saved searches

Use saved searches to filter your results more quickly

[apps/web]: Plan #1

AI Assisted 3D and AR Product Visualization for E Commerce

Executive summary

Assumptions and strategic framing

Pain point matrix

Feature priorities and competitive landscape

Competitive audit

Gaps you can exploit

Suggested positioning statements

Delivery, integrations, and go to market

Phase plan

Sprint-level eight-week plan

Seller onboarding flow recommendation

Platform provider integration checklist

Pricing model options

Partner outreach tactics

UX, analytics, compliance, and investor brief

UX copy examples

KPI framework and analytics schema

Privacy, accessibility, and mobile performance release gates

Investor-readiness brief

Metadata

Assignees

Labels

Projects

Milestone

Relationships

Development

Assumption	Default position	Why this is sensible for a hackathon
Target market	SMB Shopify merchants	Shopify provides a large merchant base, native product media handling, and a mature app-extension model, but still leaves enough integration pain for a differentiated app. citeturn33search3turn33search14turn34view0
Pricing model	SaaS + credits	Generation cost is variable; subscription-only models struggle when 3D workloads spike per SKU. Meshy and Tripo both expose credit/pay-as-you-go logic, which matches a credit-based COGS model on your side. citeturn21view0turn9search2
AI provider	Unspecified	Provider abstraction is essential because quality, latency, and cost vary by category and vendor, and no single provider is clearly dominant across all ecommerce product types. citeturn21view0turn9search2
Initial category focus	Furniture/home + accessories	These categories benefit directly from scale, placement, and “view in room” behavior. They also align with Google’s current 3D/AR support for home goods and Amazon’s “View in Your Room” positioning. citeturn26view0turn26view1turn31view0turn32search0
First channels	Shopify PDP + Google Merchant export	This combines owned PDP monetization with a discoverability channel that already supports 3D model links in limited markets/categories. citeturn34view0turn26view0turn26view1

Priority	Who benefits first	Why this is the right sequence
First	Merchant/store owner	If merchants cannot publish quickly and measure ROI quickly, the product dies before shopper delight matters. Shopify merchants still need guidance to activate app blocks and supported themes. citeturn34view0turn34view1turn34view2
Second	Platform/channel teams	Providers care about media quality, policy compliance, payload size, and support burden. Google, Amazon, Shopify, and Trendyol each impose different constraints. citeturn26view0turn26view4turn31view0turn31view4turn31view5
Third	Shopper	A better viewer, AR, and hotspot storytelling increase confidence, but only after asset quality, speed, and compatibility are solved. citeturn15search0turn16search0turn31view0

ID	Domain	Pain point	Symptoms	Business impact	Root causes	How your product addresses it	Evidence
P1	Merchant / Store Owner	3D asset creation is too slow and too expensive	Merchants rely on photo shoots, agencies, or manual modeling; long turnaround before publish	Launch delays, higher content cost, missed campaigns; Shopify’s own partner workflow expects multiple quality photos and dimensions in mm, which is a high-friction input burden	Manual 3D production, lack of standardized inputs, no fast self-serve workflow	Image-to-3D generation from existing SKU media, optional dimension capture, and provider abstraction with human QA fallback	citeturn35view0turn35view1turn16search2
P2	Merchant / Store Owner	Unclear ROI makes adoption difficult	“Looks cool” but merchant cannot tie it to conversion or AOV	Budget objections; deprioritized initiative	Missing instrumentation and weak before/after reporting	Built-in experiment mode and ROI dashboard for interaction rate, ATC uplift, conversion uplift, and per-SKU impact	Rebecca Minkoff saw 44% higher add-to-cart after 3D interaction, 27% higher orders after 3D interaction, and 65% higher orders after AR interaction. citeturn16search0
P3	Merchant / Store Owner	Theme integration is still operationally messy	Merchant installs app but widget is not visible, misplaced, or unsupported by theme structure	Support tickets, churn during onboarding, poor install-to-publish conversion	App blocks are not present by default after install; themes must support `@app` blocks and JSON templates; statically rendered sections do not support app blocks	Guided Shopify onboarding, theme compatibility check, deep links, fallback script embed, and automatic “safe location” recommendations	citeturn34view0turn34view1
P4	Merchant / Store Owner	Variant-heavy catalogs break media logic	Color/material/size variants do not map cleanly to media states	Rework, shopper confusion, poor merchandising	Shopify variant media still supports only images; 3D models/videos cannot be used as variant media	Variant-aware asset mapping, grouped media states, material/color toggles inside viewer, and “best available variant” fallback	citeturn34view2turn34view4
P5	Merchant / Store Owner	Channel mismatches multiply asset ops	One asset works on Shopify but not on Google, Amazon, or marketplace surfaces	Extra operational work, duplicated assets, asset waste	Different file, product-type, policy, and surface constraints per channel	Channel capability matrix and export rules: Shopify-native, Google `virtual_model_link`, Amazon 3D upload path, Trendyol image-first fallback	citeturn26view0turn26view1turn31view0turn31view4turn31view5
P6	Merchant / Store Owner	Generated assets often fail quality review	Wrong scale, poor materials, distorted logos, bad UVs, blurry textures	Revisions consume time; lower trust in AI output	Weak mesh/material standards, lack of commerce-focused QA, no real-world scale checks	Human-in-the-loop QA workbench with scale, texture, material, and logo checks based on commerce asset standards	citeturn35view1turn35view0turn3search0
P7	Platform Providers	Platforms bear media-policy and quality enforcement costs	Feed disapprovals, listing errors, broken rich results, unsafe or misleading media	Support burden, lower marketplace quality, lower buyer trust	Low-quality images, conflicting data, missing attributes, unsafe URLs, unsupported categories	Export validation, preflight linting, metadata preservation, and safe-host checks before publish	Google says inaccurate or missing product data can cause disapprovals or display issues; safe-search standards also apply to `virtual_model_link`. citeturn26view4turn1search16
P8	Platform Providers	App/channel integration heterogeneity increases support load	Merchants need different instructions for Shopify themes, Amazon listing APIs, Trendyol uploads	Higher integration cost and longer time to value	No common surface for 3D/AR commerce export	Adapter layer per channel plus a single merchant-facing publishing workflow	Shopify app blocks/theme support, Amazon Listings Items API/media submissions, and Trendyol Partner API all differ materially. citeturn34view0turn31view4turn18search4turn18search3
P9	Platform Providers	Marketplace surfaces are limited or category-specific	3D/AR works in some geographies or product types only	Coverage gaps reduce merchant confidence	Channel programs are staged and constrained	Capability-aware publishing; when unsupported, show enhanced 2D/lifestyle/hotspot output instead of a failed 3D promise	Google limits 3D/AR experiences to specific countries and primarily home goods/shoes; Amazon limits “View in Your Room” and other AR surfaces by type; Trendyol public docs remain image-centric. citeturn26view0turn26view1turn32search0turn31view5
P10	Shopper / Customer	Shoppers cannot judge size and context from photos alone	Questions about fit, scale, room placement, material detail	Wrong rejection of relevant products, hesitation, lower conversion	Static images lack “in-scale” context and spatial understanding	AR placement, dimension overlays, scale-calibrated viewer, and environment shots generated from the same asset	Baymard found 42% of users try to infer size from images, and poor “in scale” imagery causes misinterpretation and abandonment. citeturn15search0turn15search3
P11	Shopper / Customer	Shoppers distrust AI-enhanced media	“Will it really look like this?” skepticism	Lower conversion and potential complaints or regulatory exposure	Lack of disclosure, over-polished visuals, mismatch between render and delivered product	Trust microcopy, AI-assisted media disclosure, reference-photo view, and hotspot callouts labeled as illustrative vs exact	FTC requires advertising claims to be truthful and not misleading; Google requires AI-generated image metadata preservation in Merchant surfaces. citeturn4search2turn4search10turn26view2
P12	Shopper / Customer	Mobile/device support is inconsistent	AR button appears but fails, opens native app unexpectedly, or is unavailable	Frustration, bounce, lower engagement	ARCore device dependence, iOS/Android viewer differences, mixed in-browser vs native flows	Device detection, ar-capability checks, native Quick Look/Scene Viewer launch where appropriate, and graceful fallback to interactive 3D/poster	Apple Quick Look uses USDZ in Safari and other built-in apps; Google AR requires ARCore support; `<model-viewer>` supports WebXR/Scene Viewer/Quick Look modes. citeturn36view0turn2search1turn2search5turn25search2turn25search4
P13	Product Team / Operations	Asset governance becomes chaotic at catalog scale	Multiple versions, uncertain source of truth, no clean publish/unpublish history	Operational waste, bad exports, duplicated generation costs	No asset registry, no QA states, no versioning	Central asset registry with status, channel exports, and approval workflow	VNTANA’s product messaging exists largely because 3D workflows become unmanageable without purpose-built orchestration. citeturn19search7turn19search15
P14	Product Team / Operations	Feed and site content drift over time	Storefront, Google, marketplace, and asset metadata stop matching	Listing errors, customer confusion, policy risk	Manual field edits and siloed systems	Canonical product-to-asset mapping with export jobs and reconciliation checks	Google Merchant explicitly warns that conflicting data between feed and website, missing variant attributes, and low-quality images cause issues. citeturn26view4
P15	Product Team / Operations	Teams lack instrumentation across the 3D funnel	Can measure page views but not 3D load success, hotspots, AR launch, or fallback rate	Impossible to improve the product with confidence	Standard ecommerce analytics stops at product/cart/checkout	Viewer event schema mapped to GA4, Shopify pixels, and custom app events	Shopify exposes standard customer events like `product_viewed`, `product_added_to_cart`, and `checkout_started`; custom app events must be published separately. GA4 supports ecommerce plus custom events. citeturn7search0turn7search1turn7search8turn7search9turn7search24
P16	Legal / Privacy	Consent, tracking, and AI-content handling can become non-compliant	Pixel data collected without proper notice; AI images stripped of metadata	Regulatory risk and merchant distrust	Weak privacy design and careless asset transformation pipeline	Consent-aware analytics, minimal data collection, retention controls, DPA, and metadata-preserving media pipeline	GDPR applies to personal data including online identifiers; CCPA requires notice at or before collection; Google requires AI-generated image metadata tags to be preserved. citeturn30search7turn4search12turn4search1turn4search9turn26view2
P17	Legal / Privacy	Ownership and authorship of AI-generated 3D outputs is unclear	Merchant asks: “Do I own this asset? Can I reuse it? Can I copyright it?”	Contracting friction, enterprise security review delays	AI output law and vendor contracts remain nuanced	Provider-agnostic contract layer, provenance logs, and explicit merchant terms that distinguish AI output from human-authored edits	The U.S. Copyright Office states generative AI outputs are protectable only where sufficient human authorship exists; prompts alone are generally insufficient. citeturn17search0turn17search1turn17search7
P18	Performance / Infrastructure	3D payloads can break mobile performance	Slow LCP, janky interaction, GPU overload, failed low-end device experience	Lower SEO resilience, lower conversion, more abandonment	Large GLB textures/meshes, no lazy load, weak compression, over-eager hydration	Poster-first loading, click-to-load option, KTX2/Draco/Meshopt where beneficial, dynamic scaling, CDN edge delivery, strict asset budgets	Google’s good thresholds remain LCP ≤2.5s, INP <200ms, CLS <0.1; `<model-viewer>` recommends poster/lazy loading and warns that DRACO’s decoder is large enough to be a net loss on small files. citeturn5search1turn5search2turn36view2turn3search1turn3search2
P19	Performance / Infrastructure	Generation pipelines are asynchronous and bursty	Jobs queue up; cost and latency spike during bulk uploads	Merchant frustration and runaway COGS	Third-party generation APIs, rate limits, per-call pricing, unpredictable SKU bursts	Queue-based architecture, retry policy, budget caps, and progressive publish states	Meshy prices generation per call/credit; Tripo uses pay-as-you-go credits; Cloudflare Queues are purpose-built to buffer asynchronous work. citeturn21view0turn9search2turn23search4turn23search0
P20	Business / Go to Market	Buyer education is still required	Merchant understands photos and video, not “why 3D now?”	Longer sales cycle, especially above SMB	Market still spans enterprise 3D DAM, configurators, and AI novelty tools	Position as conversion-confidence infrastructure, not “3D for 3D’s sake”	Shopify frames 3D commerce around better confidence/conversion, not novelty; average product-page UX is still mediocre, leaving room for measurable improvement. citeturn16search1turn16search6turn15search10
P21	Business / Go to Market	Platform-provider partnerships require proof, not just a demo	Agencies/platform teams care about support cost, compatibility, and merchant outcomes	Hard to win app-store traction or referrals without hard data	Few startups bring provider-facing analytics and compliance reporting	Add provider-ready reporting: activation, asset pass rate, performance delta, merchant ROI snapshots	This is an inference from the reviewed Shopify/Amazon/Google integration surfaces and their operational constraints. citeturn34view0turn26view4turn31view4
P22	Hackathon constraints	You cannot solve perfect image-to-3D quality in eight weeks	Inconsistent results across reflective, transparent, deformable, or highly detailed products	Demo risk, overpromising risk	Model generation quality varies by category and provider; asset cleanup still matters	Limit MVP to furniture/home/accessories, add manual QA, and publish “AI-assisted” rather than “fully automatic”	Shopify and Khronos guidance both emphasize real-world scale, clean UVs, optimized textures, and model review; these are not reliably auto-solved in all categories today. citeturn35view1turn3search0turn3search4
P23	Hackathon constraints	Lack of real merchant data makes prioritization noisy	Team builds impressive demo features with low operational value	Lower odds of post-hackathon traction	No pilot stores or limited SKU corpus	Ship with 2–3 pilot merchants and a repeatable measurement plan	This is an execution inference, but it follows directly from the need to prove merchant ROI and channel compatibility early. citeturn16search0turn26view0turn34view0

Feature	Maps to pain IDs	Why it matters	Complexity	Est. effort	MVP or post-MVP
F1 Merchant onboarding wizard with catalog import	P1, P3, P20, P23	Reduces install-to-first-value time	Medium	1.5 pw	MVP
F2 Provider-agnostic image-to-3D generation adapter	P1, P19, P22	Avoids lock-in and lets you swap quality/cost profiles	High	2.0 pw	MVP
F3 Human QA workbench with approve/reject reasons	P5, P6, P13, P22	Prevents low-trust AI output from reaching shoppers	Medium	1.5 pw	MVP
F4 Shopify app block + fallback script embed	P3, P8, P21	Gives the fastest path to deploy on most merchant themes	Medium	1.5 pw	MVP
F5 Asset optimization pipeline with poster-first loading	P18, P19	Protects mobile UX and budget	Medium	1.0 pw	MVP
F6 3D viewer with hotspots, dimensions, and AR launch	P10, P11, P12	Core shopper value proposition	Medium	1.5 pw	MVP
F7 Device/channel capability detection and fallback UX	P5, P9, P12, P18	Prevents broken AR promises	Medium	0.75 pw	MVP
F8 Analytics schema + experiment mode + ROI dashboard	P2, P15, P20, P23	Converts “cool demo” into measurable product value	Medium	1.5 pw	MVP
F9 Google Merchant export for `virtual_model_link`	P5, P7, P9	Strong seller/provider differentiator for supported categories	Medium	1.0 pw	MVP
F10 Variant-aware asset mapping and material toggles	P4, P14	Important for real catalog utility beyond single-SKU demos	High	1.5 pw	Post-MVP
F11 Marketplace adapters for Amazon and Trendyol	P5, P8, P9, P21	Needed for provider-first GTM beyond Shopify	High	2.0 pw	Post-MVP
F12 Compliance layer for consent, AI metadata, provenance logs	P16, P17	Needed for trust and enterprise readiness	Medium	1.0 pw	MVP-lite, full post-MVP
F13 Bulk ingestion, queue controls, retries, and budget guardrails	P19, P22	Required once real usage appears	Medium	1.0 pw	Post-MVP
F14 Merchant benchmark insights and category templates	P2, P20	Strong long-term moat if you can compare “3D-ready PDP quality” across stores	High	2.0 pw	Post-MVP

Layer	Recommended choice	Pros	Cons	Cost / commercial note
Canonical 3D format	glTF/GLB	Royalty-free, efficient for runtime delivery, interoperable across web engines; effectively the default web-commerce target	Needs optimization discipline for mobile	Use GLB as canonical master; generate USDZ for Apple surfaces. citeturn2search3turn2search6turn3search9
Apple AR format	USDZ	Native Quick Look support in Safari, Messages, Mail, Notes, iPhone/iPad/Vision Pro	Apple-specific; not ideal as canonical authoring format	Necessary for iOS-native AR surface. citeturn36view0
Viewer	`<model-viewer>`	Web-friendly, AR-capable, supports WebXR/Scene Viewer/Quick Look patterns, hotspots/annotations, poster/lazy-load techniques	Requires careful compression and fallback logic	Best fit for hackathon speed and embeddability. citeturn25search1turn25search2turn36view2
Compression / optimization	glTF-Compressor, KTX2, Meshopt or Draco	Official Khronos ecosystem, geometry + texture compression	Decoder overhead can outweigh file-size savings on small assets; needs tuning	Use per-asset heuristics, not blanket compression. citeturn3search1turn3search2turn36view2
Generation provider	Meshy primary, Tripo alternative	Fast API integration, credit-based economics, good for hackathon velocity	Output quality varies by product type; cleanup may still be needed	Meshy exposes per-call credit pricing; Tripo uses pay-as-you-go credits. citeturn21view0turn9search2
Edge compute / orchestration	Cloudflare Workers + Queues	Good async pattern, low-friction edge execution, no egress charge on R2, request-based pricing	More low-level than full-stack managed app platforms	Workers Paid plan starts at $5/month; Queues bill by operations. citeturn23search1turn23search0turn23search4
Object storage / CDN	Cloudflare R2	S3-compatible, zero egress, low storage cost	Need your own metadata/index logic unless paired with DB	Published pricing includes $0.015/GB-month standard storage and no egress fees. citeturn21view2
Dashboard hosting	Vercel for merchant/admin UI	Extremely fast FE iteration, CI/CD, spend controls	Serverless compute overages can climb if backend logic is heavy	Pro starts at $20/month plus usage. citeturn21view3
App backend alternative	Supabase	Auth, Postgres, storage, edge functions, S3-compatible storage, fast setup	More centralized than edge-first, can feel heavier for simple widgets	Pro starts at $25/month; edge function invocations are $2 per 1M above quota. citeturn22search0turn22search1turn22search6turn22search3
Product analytics	PostHog + GA4 + Shopify Web Pixels	PostHog for product loops, GA4 for commerce reporting, Shopify events for storefront truth	Requires event governance to avoid duplication	PostHog has a generous free tier incl. 1M analytics events; GA4 ecommerce and Shopify standard events are well documented. citeturn21view4turn7search1turn7search0
Platform integrations	Shopify first; Google Merchant second; Amazon/Trendyol next	Maximizes speed to value and control	Marketplace surfaces are more constrained and slower to operationalize	Aligns with the strongest SMB motion. citeturn34view0turn26view0turn31view4turn31view5

Competitor / tool	Core motion	AI image→3D	Viewer / AR	Configurator	Channel / commerce ops	Likely weakness you can exploit
Shopify native 3D support	Native product media on Shopify	No native generation pipeline	Yes, via Shopify product media and compatible themes	Limited	Good for Shopify-only media handling	No end-to-end generation, onboarding guidance, QA, or ROI analytics. citeturn34view2turn34view3turn20search6
Amazon 3D & AR	Marketplace-native 3D/AR surfaces	Upload/create within Amazon flow	Yes, but mostly app-based customer surfaces	Try-on niche	Strong marketplace surface where supported	Not an owned-store solution; product-type constraints; weak multi-channel story for merchants. citeturn31view0turn32search0
Zakeke	3D customization and configuration	AI creative layer, but generation is not its main moat	Yes	Strong	Integrates with major commerce platforms	Best for custom/configurable products, less focused on fast 2D-catalog-to-3D onboarding for SMB merchants. citeturn10search0turn10search4turn10search18
Threekit	Enterprise visualization + configuration + guided selling	Not primarily AI image→3D self-serve	Yes	Very strong	Deep enterprise workflows	Powerful, but enterprise-heavy and services-oriented for SMB hackathon positioning. citeturn10search1turn10search5turn10search19
Emersya	Interactive 3D/AR authoring platform	Not core	Yes	Strong	Rich storytelling and no-code authoring	Strong platform, but your opening is faster AI-assisted asset creation plus merchant ROI instrumentation. citeturn10search3turn10search7
VNTANA	3D orchestration / optimization / DAM	Not core generation moat	Yes	Moderate	Strong transformation/governance	Excellent enterprise orchestration; opportunity is SMB simplicity and self-serve workflow. citeturn19search7turn19search12turn19search14
Meshy	AI 3D generation platform	Strong	Limited commerce workflow	No	API-first creation	Great raw generation, weak commerce-native publishing/governance/analytics. citeturn21view0turn9search11
Tripo	AI 3D generation platform	Strong	Limited commerce workflow	No	API/pay-go generation	Similar gap: generation is strong, commerce operations are not the main product. citeturn9search2turn9search6turn9search10

Gap	Why it matters
Self-serve SMB onboarding from existing catalog images	Most enterprise vendors assume existing 3D/CAD maturity or services budgets; raw AI vendors assume creator workflows, not merchant workflows. citeturn10search1turn19search7turn21view0
Channel compliance + fallback publishing	Merchants do not want to think about GLB vs USDZ, app block support, Google country limits, or marketplace media constraints. citeturn34view0turn26view0turn31view0turn31view5
Commerce-specific QA and asset governance	“Looks plausible” is not good enough for product detail pages; scale, material truthfulness, and logo fidelity matter. citeturn35view1turn3search0
ROI analytics baked into the product	Every merchant asks the same question: “Did this increase add-to-cart, conversion, or reduce confusion?” Many competitors mention outcomes; fewer operationalize them in the day-to-day product. citeturn16search0turn10search1turn19search7
Provider-first reporting	Agencies, platform teams, and merchant-success teams need fewer tickets, faster activation, and fewer broken embeds. Your product can become “safe to recommend.” citeturn34view0turn31view4

Phase	Purpose	Milestones	Deliverables	Success metrics	Main risks	Mitigation
Discover and constrain	Lock the wedge and merchant story	3 pilot merchants, 20–50 SKUs, one supported theme family, one category template	Problem briefs, sample dataset, success criteria, event schema v1	3 merchants agree to pilot; one real SKU ready end-to-end	Scope sprawl	Freeze category + channel list early
Build the production-shaped MVP	Deliver the shortest path to publish	Image ingest, generation job, QA, viewer, Shopify publish, analytics	Working Shopify app/widget, asset registry, viewer, metrics dashboard	<30 min from install to first live model for demo merchant, excluding human QA	Generation quality inconsistency	Human approval gate; use best-effort category filters
Add channel credibility	Show provider-first readiness	Google export, performance budget, compliance basics	`virtual_model_link` export, AI-content handling, CWV audit	Asset publish succeeds on Shopify; Google-ready feed object produced	Channel-specific edge cases	Make exports explicit and capability-gated
Pilot and investor readiness	Turn demo into believable business	Case-study data, onboarding docs, GTM assets, pitch narrative	Pilot report, deck, KPI baseline, pricing tests	1–3 pilots, measurable interaction and conversion proxy data	No real uplift data yet	Use interaction/ATC leading metrics if purchase volume is low

Week	Goal	Key tasks	Owners
Week one	Frame the wedge	Pilot merchant interviews, SKU collection, architecture, quality rubric, success metrics	PM, Design, Growth, ML/3D
Week two	Build ingestion and generation path	Product import, image upload, provider adapter abstraction, generation queue, DB schema	BE, ML/3D, PM
Week three	Build merchant review flow	QA workbench, approve/reject states, asset versioning, thumbnail/poster generation	BE, FE, Design
Week four	Build the storefront experience	`<model-viewer>` integration, hotspots, dimensions overlay, AR entry, mobile fallback UX	FE, Design, ML/3D
Week five	Shopify-first publishing	Theme app extension, app block placement, fallback embed, product mapping, install guidance	FE, BE, PM
Week six	Measure and harden	PostHog + GA4 + Shopify events, experiment mode, performance budgets, compression pipeline	FE, BE, Growth
Week seven	Add channel credibility	Google export, compliance checks, AI disclosure UX, Amazon/Trendyol adapter stub or export plan	BE, PM, Growth
Week eight	Pilot, polish, pitch	Pilot deployment, bug bash, case study draft, pricing page, investor memo, demo script	PM, FE, BE, Design, Growth

Step	Merchant experience	Product requirement
Install	“Connect your store and select products to 3D-enable.”	Shopify app install; theme support check; fallback embed available if app blocks are unsupported or inconvenient. citeturn34view0turn34view1
Select products	Merchant chooses 1–20 SKUs	Pull existing Shopify media and variant data
Generate	“We’ll turn your existing images into a draft 3D asset.”	Queue job, provider selection, budget estimate
Review	Merchant approves geometry/material/scale	QA panel with yes/no checklist based on Shopify/Khronos best practices
Publish	“Add to product page”	App block insertion instructions or auto-safe embed install
Measure	“See interaction, AR launch, add-to-cart lift”	Dashboard with 7-day and 28-day windows

Platform	Checklist	Critical constraints
Shopify	Theme app extension, app block support, product→asset mapping, Web Pixel event plan, 3D model/file storage policy, supported theme check	App blocks are not added by default; theme must support `@app`; 3D/video not valid as variant media; product media limits and plan-specific upload/storage limits apply. citeturn34view0turn34view1turn34view2
Google Merchant	Produce hosted public model URL, populate `virtual_model_link`, keep AI metadata intact, validate feed/site consistency	Current 3D/AR support is limited by geography and category; AR in Shopping ads remains constrained; AI-generated image metadata must be preserved. citeturn26view0turn26view1turn26view2turn26view4
Amazon	Use Seller Central 3D upload/media flow or listing workflows, handle product-type restrictions, ensure mobile-app expectation is clear	Listings Items API is 1x1 oriented for many tasks; product type support varies; customer 3D/AR surfaces are primarily app-based and category-limited. citeturn31view0turn31view4turn32search0
Trendyol	Use Partner API auth, HTTPS image URLs, barcode/product mapping, image-size compliance; treat 3D as off-platform/PDP enhancement until first-party support is confirmed	Public docs reviewed are image/listing-centric: max 8 images, HTTPS URLs, 1200×1800 at 96 dpi; no equivalent public first-party 3D/AR seller workflow was identified in the reviewed docs. This last point is an inference from the documentation reviewed. citeturn18search3turn18search4turn31view5

Model	Best fit	Pros	Cons	Recommendation
Flat SaaS only	Low-volume merchants	Easy to explain	Mismatched to variable generation cost	Not ideal as default
Credits only	Creator or API users	Matches COGS	Can feel punitive for merchants	Good for API-only add-on
SaaS + credits	SMB and agencies	Predictable base + variable usage	Slightly more complex	Default recommendation
White-glove setup + SaaS	Agencies / mid-market	Faster ACV growth	Services-heavy, less scalable	Use selectively post-MVP

Target partner	Why they matter	Outreach angle
Shopify agencies / partners	Control store setup and theme placement	“Reduce custom theme work and show measurable PDP uplift.”
Product photography studios / 3D freelancers	Already serve the same merchants	“Your service becomes faster with AI-assisted first draft + QA workflow.”
Feed / PIM / DAM consultants	Own catalog and channel pain	“Add 3D/AR readiness as a measurable channel export layer.”
Marketplace integrators in Türkiye	Useful for Trendyol/marketplace expansion	“Offer image-first fallbacks now, marketplace-rich media later.”
Merchant success teams at platforms	Provider-first wedge	“Fewer broken embeds, better media quality, documented adoption metrics.”

KPI group	Core KPI	Why it matters
Merchant activation	Install→first published model time	Measures real setup friction
Merchant adoption	% of installed stores publishing at least one model	Measures actual value, not vanity installs
Shopper engagement	3D interaction rate, hotspot open rate, AR launch rate	Measures whether the experience is discoverable and useful
Commercial outcome	Add-to-cart uplift, PDP conversion uplift, assisted revenue	Ties the product to spend justification
Operations	Approval rate, revision rate, generation latency, cost per approved SKU	Protects margins and team bandwidth
Performance	3D-enabled PDP CWV pass rate	Prevents the feature from harming the store

Event	Trigger	Key properties	GA4 mapping	Shopify mapping
`product_viewed`	Shopper visits PDP	product_id, variant_id, category	Standard GA4 ecommerce context	Native Shopify standard event `product_viewed` citeturn7search0turn7search1
`app_3d_widget_loaded`	Viewer successfully initializes	product_id, asset_id, load_ms, fallback_state	Custom event	Custom app pixel event; prefix with app name per Shopify guidance citeturn7search9turn7search24
`app_3d_interaction_started`	First rotate/zoom interaction	product_id, device_type, dwell_ms	Custom event	Custom Shopify pixel event
`app_hotspot_opened`	Shopper opens hotspot	hotspot_id, product_id	Custom event	Custom Shopify pixel event
`app_ar_launched`	Shopper taps AR	platform, ar_mode, product_id	Custom event	Custom Shopify pixel event
`product_added_to_cart`	Add to cart from PDP	product_id, variant_id, quantity	GA4 `add_to_cart`	Native Shopify `product_added_to_cart` citeturn7search2turn7search1
`checkout_started`	Shopper starts checkout	cart_value, item_count	GA4 `begin_checkout`	Native Shopify `checkout_started` citeturn7search8turn7search1
`purchase`	Order complete	revenue, items, model_interacted_before_purchase	GA4 `purchase`	Native purchase pipeline / order reconciliation citeturn7search10turn7search1
`app_asset_generation_started`	Merchant starts generation	provider, sku_count, estimated_cost	Internal ops event	N/A
`app_asset_generation_completed`	Job done	provider, latency_s, quality_score	Internal ops event	N/A
`app_publish_clicked`	Merchant publishes asset	channel, product_id, asset_id	Internal ops event	N/A
`app_quality_rejected`	QA rejects asset	reject_reason, provider, category	Internal ops event	N/A

Dashboard	Primary widgets	Suggested charts
Merchant activation	installs, published stores, time to first publish	Funnel: install → select products → generate → approve → publish
Content ops	generation volume, approval rate, rejection reasons, cost per approved SKU	Stacked bar by rejection reason; heatmap by provider × category
Shopper engagement	viewer loads, interaction rate, hotspot usage, AR launches	PDP engagement timeline; device split donut; AR launch rate by product
Commercial impact	ATC uplift, conversion uplift, assisted revenue, AOV delta	A/B uplift chart; cohort line chart by product category
Performance	LCP/INP/CLS for 3D vs control PDPs	Box plot by device class; origin-level CWV pass rate
Channel exports	Shopify published, Google exported, marketplace-ready backlog	Status board by channel

Area	Requirement	Concrete implementation check	Release gate
Privacy	Data minimization	Do not collect more than needed for analytics and QA; separate merchant/admin data from shopper event data	Pass only if event catalog and retention policy are documented. GDPR’s data framework and online identifiers are relevant here. citeturn30search7turn4search12
Privacy	Notice and consent	Respect consent before pixels/session replay; provide notice at or before collection where applicable	Pass only if consent management and event suppression are tested. citeturn4search1turn4search9turn7search6
AI content	Metadata preservation	Preserve IPTC `DigitalSourceType` tags for AI-generated commerce imagery; do not strip tags during optimization	Pass only if media pipeline preserves metadata for applicable image outputs. citeturn26view2
Truthfulness	Do not imply exactness where approximate	Mark AI-assisted visuals clearly; do not overclaim material/fit/size precision	Pass only if trust microcopy and QA thresholds are enabled. FTC truth-in-advertising applies. citeturn4search2turn4search10
IP / authorship	Asset provenance logging	Record provider, prompt/input images, human edits, publish approver	Pass only if provenance exists for every published asset. citeturn17search0turn17search7
Accessibility	Provide non-3D equivalent information	Keep product images, dimensions, and text specs available even if viewer fails or is skipped	Pass only if PDP remains usable without 3D
Accessibility	Keyboard and pointer alternatives	If rotation or hotspot access depends on dragging, provide button or single-pointer alternatives	Pass only if viewer controls and hotspot navigation work without drag. WCAG 2.5.7 requires a non-drag alternative where dragging is non-essential. citeturn27search1turn27search8
Accessibility	Touch target sizing	Hotspots and control targets at least 24×24 CSS px or valid spacing exception	Pass only if mobile hotspot targets meet WCAG 2.5.8. citeturn28search0turn28search1
Accessibility	Focus visibility	Hotspots/buttons visibly focused and not obscured	Pass only if keyboard QA on mobile/desktop is clean. citeturn4search3turn27search11
Performance	Asset budgets	Default target GLB around 4 MB when possible; avoid routine publish above 15 MB for Shopify-first workflow unless aggressively optimized	Pass only if asset budget report is clean. citeturn35view0turn34view3
Performance	Lazy load and poster	Do not auto-hydrate heavy 3D before user intent on weak/mobile contexts	Pass only if poster-first path exists. citeturn36view2
Performance	CWV thresholds	Monitor LCP, INP, CLS on 3D-enabled PDPs	Pass only if origin/page remains within acceptable thresholds in real-user monitoring. citeturn5search1turn5search4

Layer	Assumption logic	Illustrative figure
TAM	Commerce merchants who could benefit from confidence-centric 3D/AR product media across owned stores and major channels. Shopify alone says millions of businesses use the platform in 175+ countries, so the relevant merchant universe is clearly in the millions even before marketplaces are counted. citeturn33search3turn33search14	2.0M merchants (illustrative subset assumption)
SAM	SMB Shopify merchants in furniture/home/accessories, with enough image quality and SKU density to justify 3D/AR, initially concentrated in Shopify-friendly and Google-supported markets	60k merchants
SOM	Merchants reachable within three years through Shopify app distribution, agencies, and pilot-led referrals	1.5k merchants

Revenue stream	Comment
Core SaaS subscription	Access to onboarding, publishing, analytics, QA workflow
Usage credits	Charged on generation/export volume; aligns to underlying provider cost curve
Agency / partner seats	Multi-store workflows, centralized billing, white-label reporting
Professional services	Optional onboarding, catalog cleanup, template setup, and asset QA for larger accounts

Horizon	Product focus	Commercial focus
Year one	Shopify-first workflow, furniture/home templates, QA workbench, ROI analytics, Google export	Pilot merchants, Shopify agencies, early app-store motion
Year two	Variant logic, Amazon/marketplace adapters, bulk ops, stronger governance, richer experimentation	Mid-market brands, feed/PIM partners, agency channel expansion
Year three	Multi-modal input pipeline, stronger automated QA, deeper DAM/PIM integrations, category-specific models	Platform/provider partnerships, enterprise upsell, API revenue

[apps/web]: Plan #1

Description

AI Assisted 3D and AR Product Visualization for E Commerce

Executive summary

Assumptions and strategic framing

Pain point matrix

Feature priorities and competitive landscape

Competitive audit

Gaps you can exploit

Suggested positioning statements

Delivery, integrations, and go to market

Phase plan

Sprint-level eight-week plan

Seller onboarding flow recommendation

Platform provider integration checklist

Pricing model options

Partner outreach tactics

UX, analytics, compliance, and investor brief

UX copy examples

KPI framework and analytics schema

Privacy, accessibility, and mobile performance release gates

Investor-readiness brief

Metadata

Metadata

Assignees

Labels

Projects

Milestone

Relationships

Development

Issue actions