Skip to content

Repository files navigation

Digitalised Source-to-Order with Supplier Selection

Team: Aquatis

Module: Digitalisation of Business Processes (DigiBP) — SS 2026 Programme: FHNW · MSc Business Information Systems Supervisors: Andreas Martin · Charuta Pande · Devid Montecchiari


Table of Contents


Project Members

Project Team / Authors

Name Role Email
Talip Ates Make Scenarios & Documentation talip.ates@students.fhnw.ch
Joanne Chimuti-Lobsiger As-IS BPMN & To-Be Documentation joanne.chimutilobsiger@students.fhnw.ch
Harpreet Kaur Forms, Presentation & Documentation harpreet.kaur@students.fhnw.ch
Alexis Marquet Repository, Documentation alexis.marquet@students.fhnw.ch
Lukas Uske To-be BPMN, Make & Documentation lukas.uske@students.fhnw.ch

Supervisors

Name Email
Andreas Martin andreas.martin@fhnw.ch
Charuta Pande charuta.pande@fhnw.ch
Devid Montecchiari devid.montecchiari@fhnw.ch

Abstract & Project Overview

This project digitalises and automates the Source-to-Order (S2O) process of a fictional office-supplies company that sells via an e-commerce platform. The driving scenario: a customer requests an eco-friendly paper product that is not in the existing catalogue and must be sourced from a new supplier.

In the AS-IS process, such sourcing requests are handled manually — triggered by phone calls or emails to the sales team. This leads to missed opportunities, slow response times, and an inconsistent customer experience.

The TO-BE process replaces manual handovers with a structured digital workflow. The customer submits a sourcing request via a web form; the system automatically records it, routes it through procurement, runs an automated supplier evaluation, generates a contract, and produces a purchase order — all orchestrated by Camunda 7 (BPMN/DMN) and Make.com.

The core innovation lies in the automated supplier selection step: incoming supplier responses are scored using a DMN decision table (price × delivery time × compliance × quality × financial stability), and a second DMN compares the resulting scores to recommend the best candidate. A human reviewer retains the final say.

⬆️ Back to Top


Methodology

This project follows a process-driven automation methodology designed to digitalise the Source-to-Order process using low-code tools, APIs, and workflow orchestration. The methodology is practical and implementation-focused, guiding the project from process analysis to end-to-end testing.

1. AS-IS Process Analysis

The first step is to understand the current Source-to-Order process. The existing workflow is analysed to identify how sourcing requests are initiated, processed, and handled by sales, procurement, and master data. Key issues such as manual handovers, missing information, and the risk of missed opportunities are identified.

2. TO-BE Process Design

Based on the AS-IS analysis, a target (TO-BE) process is defined. The new process shifts the trigger from manual customer contact to a structured, digital self-service request. The redesigned flow ensures that all required information is captured upfront and that responsibilities are clearly assigned to each stakeholder.

3. Process Modelling with BPMN

The TO-BE process is modelled using Camunda and BPMN standards. This step defines:

  • Process flow and decision points
  • User tasks for sales, procurement, and master data
  • System interactions and process boundaries The model serves as the central reference for the automated workflow.

4. Data Collection and Stakeholder Interaction

Google Forms is used as the main entry point for customer sourcing requests. The form ensures standardised data collection and acts as the initial communication channel between the customer and internal stakeholders.

5. Integration and Automation

Make is used to automate data movement and system integration through APIs. This includes:

  • Sending form data into the CRM
  • Triggering the Camunda process
  • Updating CRM and ERP data stores APIs are tested and validated using Postman before being integrated into the automated flow.

6. System Simulation (CRM & ERP)

To avoid dependency on real enterprise systems, Google Sheets are used to simulate CRM and ERP systems. These sheets act as a fake database containing materials, suppliers, and sourcing requests, enabling realistic process execution and data handling.

7. Workflow Execution and Monitoring

Camunda Cockpit is used to run, monitor, and manage process instances. User tasks are assigned, executed, and tracked, allowing validation of the workflow logic and stakeholder interactions.

8. End-to-End Testing and Validation

The final step validates the complete Source-to-Order flow:

  • Customer submits a request
  • Data is processed via APIs
  • Tasks are executed by stakeholders
  • CRM and ERP records are updated This confirms that the automated process works as intended and delivers a more reliable and efficient sourcing experience.

⬆️ Back to Top


How to Run the Process

The implemented process combines Camunda, Make.com and Google Forms to demonstrate a digital source-to-order workflow with supplier selection. The process can be tested end-to-end by submitting a new material request and then following the created process instance in Camunda.

1. Start the Process

The process starts with the external Google Form New Material Request.

Open New Material Request Form

After submitting the form, Make.com transfers the submitted request data to Camunda and starts a new process instance under the project tenant 26DIGIBP34.

2. Follow the Process in Camunda

After the process has been started, it can be followed in the DigiBP Camunda Tasklist.

Open DigiBP Camunda Tasklist

After logging in, select the project tenant 26DIGIBP34 and complete the generated user tasks in sequence. The main user tasks are:

  1. Review Customer Request and Catalogue Check
    Sales reviews the request and checks whether the requested material is available in the catalogue.

  2. Alternative Procurement Request
    If the requested material is not available or not feasible, Sales can propose an alternative product.

  3. Feasibility Check and RFQ Contact List
    Procurement checks feasibility and enters the supplier contacts for the RFQ.

  4. Review Supplier Response
    Procurement reviews the submitted supplier quotations.

  5. Review Contract Draft
    Procurement reviews the generated contract draft before finalisation.

3. Submit Supplier Responses

Supplier quotations are submitted through the external Google Form RFQ Response.

Open RFQ Response Form

In the actual process, the RFQ links are sent to suppliers with pre-filled fields such as the process instance ID and supplier information. This allows each submitted supplier response to be matched to the correct Camunda process instance in tenant 26DIGIBP34.

Limitation of Make.com

The Make.com scenarios use polling-based triggers for Google Form responses. This means that Make.com checks regularly whether new responses are available. If no new data is found, the scenario does not continue (interval: every 15 minutes).

Lecturers can access our dashboard on Make.com directly and trigger the scenarios manually as well. 🔗 Open Make.com project dashboard


Repository Structure

/
├── BPMN/                                   BPMN process models
│   ├── AS-IS_S2O.bpmn                      Current manual Source-to-Order process
│   ├── TO-BE_S2O.bpmn                      Conceptual digital TO-BE process
│   └── Weiterentwicklung-Automation_2.bpmn Executable Camunda 7 supplier-selection prototype
├── DMN/                                    DMN decision models
│   ├── Evaluate-supplier-response 1.dmn    Scores individual supplier responses
│   └── DMN-Best-Supplier 1.dmn             Selects the best supplier based on supplier scores
├── Forms/                                  Camunda forms and form documentation
│   ├── customer-request-catalogue 1.form   Sales task: customer request and catalogue check
│   ├── alternative_procurement-request 1.form Sales task: alternative product proposal
│   ├── feasbible_RFQ-list.form             Procurement task: feasibility check and RFQ contact list
│   ├── Review-Supplier 1.form              Procurement task: review supplier responses
│   ├── Review_Contract_Draft 1.form        Procurement task: review generated contract draft
│   ├── Form-Customer-Request-Catalogue.png Screenshot of the customer request Camunda form
│   ├── Form-Alternative-proc-Request.png   Screenshot of the alternative procurement form
│   ├── Form-Feasible-RFQ-List.png          Screenshot of the RFQ contact list form
│   ├── Form-Review-Supplier.png            Screenshot of the supplier review form
│   ├── HTTP-Form-Review-Contract-Draft.png Screenshot of the contract draft review form
│   ├── Google-Form-New-Material-Request.pdf Documentation of the external customer request form
│   └── Google-Form-RFQ-Response.pdf        Documentation of the external supplier response form
├── Make/                                   Make.com integration documentation
│   ├── 1-Trigger_Integration_Google-Forms_HTTP.png Starts the Camunda process from Google Forms
│   ├── 2-Send_RFQ.png                      Sends RFQ emails to suppliers
│   ├── 3-Supplier_Response.png             Receives supplier responses and correlates them to Camunda
│   └── 4-Generate_contract-draft.png       Generates the contract draft via Google Docs template
├── Templates/                              Contract-generation templates
│   └── Template_Supplier_Agreement.pdf     Supplier agreement template used for automated contract draft generation
├── AS-IS Process Description: Supplier Selection Additional AS-IS process description
└── README.md                               Main project documentation

⬆️ Back to Top


AS-IS Process

Process Description

In the current setup, a sourcing request begins when a customer contacts the Sales team directly via phone or email. The Sales representative manually logs the request, checks the product catalogue, and — if the item is not available — informs Procurement via email. Procurement then identifies suppliers, sends RFQs (often as Word documents attached to emails), compiles incoming quotations in Excel, manually computes a comparison matrix, and selects a supplier. Contracts are prepared as Word documents, sent for signature, and archived in shared folders. New suppliers are created manually in the ERP system by Master Data.

Key Limitations of the AS-IS Process

The AS-IS analysis surfaced the following structural weaknesses:

Manual, Channel-Dependent Trigger

  • Requests arrive via phone or email; missed calls and lost emails mean missed opportunities
  • No standardised data capture — Sales reps must chase customers for missing information

Fragmented Tooling

  • CRM, Excel, email, and the ERP system are not connected
  • RFQs and quotations move through inboxes, not a structured repository
  • Supplier comparison is built by hand in Excel — error-prone and not auditable

Slow, Opaque Hand-Offs

  • Each transition between Sales, Procurement, and Master Data is a manual notification
  • No visibility on where a request currently sits
  • Contract negotiation loops rely on email threads with no SLA enforcement

No Decision Support

  • Supplier evaluation depends on the procurement officer's intuition
  • No consistent weighting of price, delivery time, compliance, quality, or financial stability
  • Hard to justify why a particular supplier was chosen

Together, these issues result in long cycle times, lost requests, inconsistent supplier choices, and poor traceability.

AS-IS BPMN Diagram

The full AS-IS model is in BPMN/AS-IS_S2O.bpmn. It uses three swimlanes — Customer · Sales · Procurement · Master Data — and shows manual hand-offs, email-based RFQs, and Excel-based supplier comparison.

Open the file in Camunda Modeler to inspect the model.

Project Goal

The goal is to digitalise and partially automate the Source-to-Order process, with a particular focus on structured supplier selection. The TO-BE system should:

  • Replace the manual trigger with a digital self-service request form
  • Standardise data capture across all stakeholders (Sales, Procurement, Master Data)
  • Automate supplier scoring and selection through DMN decision tables
  • Orchestrate the entire flow via Camunda 7 with clear, traceable task assignments
  • Use Make.com to bridge between Camunda and external services (Google Forms, email, Google Docs)
  • Provide a complete audit trail — every decision, every supplier score, every contract version is logged

⬆️ Back to Top


TO-BE Process

Process Overview

The TO-BE process replaces the email-driven workflow with a digital end-to-end pipeline. A customer submits a structured request, which becomes a Camunda process instance. Procurement performs a feasibility check, an RFQ is sent to multiple suppliers via Make.com, supplier responses are stored back into Camunda, each response is scored by a DMN, the best supplier is selected by a second DMN, and the process continues to contract generation and purchase order.

The TO-BE model is split across two BPMN files:

File Purpose
BPMN/TO-BE_S2O.bpmn Conceptual end-to-end TO-BE model covering the full Source-to-Order flow
BPMN/Weiterentwicklung-Automation_2.bpmn Executable Camunda 7 sub-process implementing the automated supplier-selection block (RFQ → score → best supplier → contract draft)

TO-BE BPMN Diagram

Open both BPMN files in Camunda Modeler. Together they describe the conceptual flow and the deployable automation core.

Step-by-Step Walkthrough

1. Customer Request Submission

A customer fills out the "Submit Request" form on the e-commerce site. The form is implemented as a Google Form that captures: customer details, requested product, quantity, sustainability requirements, and deadline.

📄 Form: Forms/customer-request-catalogue 1.form — used by Sales to review and check the catalogue.

2. Trigger Camunda via Make.com

When the form is submitted, Make.com picks it up via the Google Forms module and sends an HTTP POST to Camunda's REST API:

POST https://digibp.engine.martinlab.science/engine-rest/process-definition/key/supplier_request_process/tenant-id/26DIGIBP34/start

This starts a new process instance with the form data as process variables.

🖼️ See Make/1-Trigger_Integration_Google-Forms_HTTP.png.

3. Catalogue Check (Sales)

A user task is created in the Camunda Tasklist for Sales: "View new request and Check Material Catalogue". The Sales representative confirms whether the requested product exists in the catalogue.

  • Yes → Continue the standard sales process (out of scope).
  • No → Procurement is engaged.

4. Propose Alternative (Optional)

If a close alternative exists, Sales proposes it to the customer via Forms/alternative_procurement-request 1.form. If the customer accepts, the standard process resumes; otherwise the procurement flow is triggered.

5. Procurement: Feasibility Check & RFQ List

Procurement opens Forms/feasbible_RFQ-list.form to:

  • Decide whether an existing supplier can fulfil the request (if so, source directly).
  • Otherwise, compile a list of 3 candidate suppliers to receive an RFQ.

6. Send RFQs via Make.com

A Camunda Send Task posts the RFQ payload to a Make.com webhook. Make.com routes the request to each of the 3 supplier email addresses, generating individual RFQ emails.

🖼️ See Make/2-Send_RFQ.png — webhook → router → 3 supplier email branches.

7. Wait for Supplier Responses (Message Catch Event)

The process pauses on supplierResponseReceived message catch events. Suppliers submit their offers via another Google Form (price, delivery time, compliance level, quality score, financial stability rating). Make.com correlates each response back to the running process instance via Camunda's /engine-rest/message API.

🖼️ See Make/3-Supplier_Response.png.

8. Score Each Supplier Response (DMN — Evaluate supplier response)

Each incoming supplier response triggers a Business Rule Task that invokes the DMN/Evaluate-supplier-response 1.dmn decision. See Decision Automation below for details. The resulting supplierScore is stored back as a process variable.

9. Select Best Supplier (DMN — Select Best Supplier)

Once all expected responses are scored, a second Business Rule Task runs DMN/DMN-Best-Supplier 1.dmn, which compares supplier1Score, supplier2Score, and supplier3Score and returns the winner.

10. Human Review (User Task)

The recommended supplier is presented to a procurement manager via Forms/Review-Supplier 1.form. The reviewer can:

  • Accept the recommendation → continue to contract.
  • Override with justification → procurement can manually document a different decision.

11. Generate Contract Draft via Make.com

Camunda fires a webhook to Make.com, which uses Google Docs to render a contract from a template, populating the supplier name, terms, and price.

🖼️ See Make/4-Generate_contract-draft.png — Webhook → Google Docs (create from template) → HTTP back to Camunda.

12. Contract Review

The generated contract draft is reviewed via Forms/Review_Contract_Draft 1.form. Procurement can check the generated contract draft and decide whether it is acceptable. In the current process this step represents the manual review before finalisation.

13. Process Closure

After the contract draft has been reviewed, the process reaches its final review stage. Due to our company's internal policies, signatures must currently still be provided manually. Therefore, the pre-filled contract can be downloaded and reused.

Challenges Addressed by the TO-BE Process

Challenge (AS-IS) Solution (TO-BE)
Manual phone/email trigger; lost requests Digital self-service form → automatic process start
Excel-based supplier comparison; inconsistent decisions DMN decision tables produce auditable, reproducible scores
Fragmented tooling (CRM, Excel, email, ERP) Camunda orchestrates; Make.com integrates external services
No visibility on request status Camunda Cockpit provides real-time process instance tracking
Email-based RFQ handling Automated RFQ dispatch and message catch events; timer escalations are planned as future improvement
No audit trail Every decision, score, and form submission stored as variables

Users and Stakeholders

Stakeholder AS-IS Role TO-BE Role
Customer Calls/emails Sales Submits structured request via web form
Sales Manually logs request, contacts Procurement Reviews catalogue match, proposes alternatives in Tasklist
Procurement Manages RFQs by email, builds Excel matrix Runs feasibility check, reviews DMN-recommended supplier
Master Data Manually creates supplier in ERP Reviews vendor data, approves automated creation request
Supplier (external) Replies via email Submits quotation via form; responds to contract via form

⬆️ Back to Top


Decision Automation (DMN)

Two DMN decision tables drive the automated supplier-selection logic. They are invoked from BPMN Business Rule Tasks in the executable Weiterentwicklung-Automation_2.bpmn.

1. Evaluate Supplier Response

File: DMN/Evaluate-supplier-response 1.dmn Decision ID: EvaluateSupplierResponse Hit Policy: FIRST

This decision scores one supplier response on a multi-criteria basis.

Inputs

Variable Type Description
lastPrice number Quoted unit price
lastDeliveryTime number Promised delivery time in days
lastCompliance string "FULL", "PARTIAL", or "NONE" — compliance with T&C
lastQualityScore number Quality score (provided alongside the response)
lastFinancialStability number Financial-stability rating

Output

Variable Type Description
supplierScore number Composite score: pricePts + deliveryPts + qualityScore + financialStability + compliancePts

Scoring Logic (summary)

  • Compliance gate:
    • NONE → score is 0 (instant disqualifier — supplier is dropped)
    • PARTIAL → compliance contribution = 2
    • FULL → compliance contribution = 5
  • Price band: ≤1000 → 5, 1000–2000 → 4, 2000–3000 → 3, >3000 → 2
  • Delivery band: ≤5 days → 5, 5–10 → 4, 10–15 → 3, >15 → 2
  • Quality & financial stability are added directly from inputs.

The full table contains 33 rules (1 NONE + 16 FULL × 16 PARTIAL).

2. Select Best Supplier

File: DMN/DMN-Best-Supplier 1.dmn Decision ID: SelectBestSupplier Hit Policy: FIRST

This decision takes the scores of up to three suppliers and returns the winning supplier name.

Inputs

Variable Type Description
expectedResponses integer Number of supplier responses received (1–3)
supplier1Score number Score from DMN #1 for supplier 1
supplier2Score number Score from DMN #1 for supplier 2
supplier3Score number Score from DMN #1 for supplier 3

Output

Variable Type Description
bestSupplier string "supplier1", "supplier2", or "supplier3"

Decision Logic

The table compares scores pairwise across the responses received. Ties resolve in favour of the lower-indexed supplier (consistent with the FIRST hit policy).

⬆️ Back to Top


Technologies Used 🛠

Component Tool / Technology Purpose
Workflow engine Camunda Platform 7 (BPMN + DMN) Orchestrates the entire process; runs DMN decisions
Process modelling Camunda Modeler Designing BPMN and DMN files
Integration / iSaaS Make.com Connects Google Forms, Gmail, Google Docs to Camunda
Customer-facing entry Google Forms Customer request and supplier response submission
Document generation Google Docs (via Make) Contract drafts from templates
Notifications Gmail (via Make) RFQ emails to suppliers
Simulated CRM/ERP Google Sheets Lightweight data store for materials, suppliers, requests
API testing Postman Validating Camunda REST endpoints
User interfaces Camunda Forms (.form) Human-task interfaces in the Camunda Tasklist

⬆️ Back to Top


Workflow Orchestration

Camunda BPMN Engine

The executable BPMN model (Weiterentwicklung-Automation_2.bpmn) runs on the FHNW Camunda instance.

Configuration

  • Engine REST base URL: https://digibp.engine.martinlab.science/engine-rest
  • Camunda Tasklist: Open DigiBP Camunda Tasklist
  • Tenant ID: 26DIGIBP34
  • Process Key: supplier_request_process
  • Authentication: FHNW DigiBP Camunda credentials are required

Key BPMN constructs used

Construct Where it is used
User Task view new request and Check Material Catalogue, Review supplier response, Review Selected Supplier and contract draft
Service Task (Send) Send RFQs to all suppliers (→ Make webhook); Generate contract draft (→ Make webhook)
Business Rule Task Evaluate supplier response (DMN #1); Select best supplier (DMN #2)
Message Catch Event supplierResponseReceived, contractDraftGenerated
Exclusive Gateway (XOR) material in catalogue?, client agrees?, Can an existing supplier fulfill the request?, All responses received?

For a detailed overview of the process variables used across Camunda, Make.com, Google Forms and DMN decision tables, see the Data Dictionary.

Starting a Process Instance

Process instances are started by Make.com (Google Forms trigger) via:

POST https://digibp.engine.martinlab.science/engine-rest/process-definition/key/supplier_request_process/tenant-id/26DIGIBP34/start

Content-Type: application/json

{
  "variables": {
    "customerName":    { "value": "...", "type": "String"  },
    "requestedItem":   { "value": "...", "type": "String"  },
    "quantity":        { "value": 100,   "type": "Integer" },
    "sustainability":  { "value": "...", "type": "String"  }
  }
}

Make.com Integration

Make.com scenarios bridge Camunda and external services. There are four scenarios:

Note: The Make.com scenario links are shared for documentation and review purposes. They provide a read-only view of the automation logic. The executable scenarios remain in the project Make.com workspace.

Scenario 1 — Trigger Process from Google Form

🖼️ Make/1-Trigger_Integration_Google-Forms_HTTP.png 🔍 View read-only Make scenario

Google Forms (Watch Responses)HTTP POST /engine-rest/process-definition/key/{key}/start

Scenario 2 — Send RFQs to Suppliers

🖼️ Make/2-Send_RFQ.png 🔍 View read-only Make scenario: Send RFQ

Custom Webhook (from Camunda)RouterGmail × 3 (one branch per supplier email).

Scenario 3 — Receive Supplier Responses

🖼️ Make/3-Supplier_Response.png 🔍 View read-only Make scenario: Supplier Response

Google Forms (Watch Responses)HTTP POST /engine-rest/message to correlate supplierResponseReceived to the running instance.

Scenario 4 — Generate Contract Draft

🖼️ Make/4-Generate_contract-draft.png 🔍 View read-only Make scenario: Generate Contract Draft

Custom WebhookGoogle Docs (Create from Template)HTTP POST /engine-rest/message to correlate contractDraftGenerated.

End-to-End Flow of the Executable Prototype

1. Google Form submission: New Material Request
   ↓ Make scenario 1
2. Camunda process instance starts under tenant 26DIGIBP34
   ↓
3. Sales user task: Review Customer Request and Catalogue Check
   ├─ material in catalogue → standard sales process / end of prototype path
   └─ material not in catalogue → continue with procurement
   ↓
4. Optional Sales user task: Alternative Procurement Request
   ├─ alternative accepted → standard sales process / end of prototype path
   └─ no suitable alternative or procurement required → continue
   ↓
5. Procurement user task: Feasibility Check and RFQ Contact List
   ├─ existing supplier can fulfil request → source through existing supplier / end of prototype path
   └─ external RFQ required → continue
   ↓
6. Camunda sends RFQ payload to Make scenario 2
   ↓
7. Make.com sends RFQ emails to the entered supplier contacts
   ↓
8. Suppliers submit quotations through the RFQ Response Google Form
   ↓ Make scenario 3
9. Make.com correlates supplierResponseReceived messages to the running Camunda instance
   ↓
10. Business Rule Task: Evaluate each supplier response with DMN "EvaluateSupplierResponse"
   ↓
11. Gateway: all expected supplier responses received?
   ├─ no → wait for further supplier responses
   └─ yes → continue
   ↓
12. Business Rule Task: Select best supplier with DMN "SelectBestSupplier"
   ↓
13. Procurement user task: Review Supplier Recommendation
   ↓
14. Camunda sends contract-generation request to Make scenario 4
   ↓
15. Make.com generates a contract draft in Google Docs
   ↓
16. Make.com correlates contractDraftGenerated message to Camunda
   ↓
17. Procurement user task: Review Contract Draft
   ↓
18. End of executable prototype. The user can download the contract draft and continue manually with the process

⬆️ Back to Top


Digital User Interfaces (Forms)

All human interactions occur through Camunda Forms (.form) rendered in the Camunda Tasklist:

File Screenshot Used by Purpose
Forms/customer-request-catalogue 1.form Screenshot Sales Review customer request + record catalogue check
Forms/alternative_procurement-request 1.form Screenshot Sales Propose an alternative product; prepare procurement
Forms/feasbible_RFQ-list.form Screenshot Procurement Feasibility check + capture RFQ contact list (3 suppliers)
Forms/Review-Supplier 1.form Screenshot Procurement Review individual supplier responses
Forms/Review_Contract_Draft 1.form Screenshot Procurement Review the generated contract draft

The customer-facing request and the supplier-facing quotation submission are both implemented as Google Forms outside Camunda, integrated via Make.com.

The following Google Forms are used for interactions outside the Camunda Tasklist. They are integrated into the workflow through Make.com.

Google Form Link Screenshot Used by Purpose
New Material Request Open form Screenshot Customer / Requester Starting form for submitting a new material procurement request
RFQ Response Open form Screenshot Supplier Supplier-facing quotation submission form with pre-filled fields for process instance and supplier data

⬆️ Back to Top


Limitations & Future Improvements

This section captures the current limitations of the prototype and the future improvements we would prioritise for a production-grade Source-to-Order solution. The two are presented together because each limitation directly motivates a specific next step.

1. Supplier database is still manually maintained

One limitation of the current workflow is that supplier identification and RFQ contact-list creation are still performed manually. While this keeps the demonstrator understandable and controllable, it limits scalability and prevents the process from automatically leveraging supplier master data, historical performance, certifications, or external supplier databases.

Improvement. Connect the workflow to a supplier master-data system or supplier database. This would allow the process to automatically suggest suitable suppliers based on product category, previous orders, delivery performance, certifications, location, compliance status, or historical quality ratings — making the workflow more data-driven and less dependent on manual input.

2. Supplier evaluation uses a simplified scoring model

The current supplier evaluation is based on a simplified scoring model. The DMN decision evaluates price, delivery time, quality score, financial stability, and compliance. This is useful for demonstrating decision automation because the logic is transparent and easy to understand. However, the model is relatively static and does not yet reflect the full complexity of real supplier selection. In practice, supplier selection is a multi-criteria decision-making problem and different procurement cases may require different weightings.

Improvement. Move from a static rule-based scoring to case-specific weighting profiles (e.g. "cost-driven", "quality-driven", "sustainability-driven"). The DMN could be split into a weighting decision plus a scoring decision, and the weights could be parameterised at process start.

3. Webhook-Based Make.com Triggers

The current Make.com scenarios are based on polling triggers. This means that Make.com checks regularly whether new Google Form responses or data entries are available. Although this works for the demonstrator, it is not fully event-driven and may create unnecessary scenario executions.

Improvement. A future improvement would be to replace polling-based triggers with webhook-based triggers. With webhooks, Make.com would only start when a new event actually occurs, for example when a new material request is submitted or when a supplier sends an RFQ response.

4. Missing timeout and escalation for supplier responses

The current workflow assumes that all contacted suppliers submit a response. The process counts received responses and compares them with the expected number. However, if one supplier does not answer, the process may remain open and wait indefinitely. In practice, suppliers may answer late, forget to respond, or decline to participate.

Improvement. Add timer events and escalation paths to the BPMN model. The workflow could send an automatic reminder after a defined number of days. If no response is received after the final deadline, the process could continue with the available responses or escalate to a procurement employee — making the workflow robust against stuck process instances.

5. Risk of duplicate or incorrect supplier responses

The current response handling counts incoming supplier responses and stores values based on supplier number. This works for the demonstrator but does not fully prevent duplicate responses: if a supplier submits the form twice, the response counter increases twice. Incorrect supplier numbers or process instance IDs could also lead to wrong mappings.

Improvement. Introduce unique response tokens per supplier. Each RFQ email would contain a supplier-specific link with a unique token. When a supplier submits a response, the workflow checks whether this supplier has already responded — and either rejects the duplicate or updates the existing record without re-incrementing the counter. This improves data consistency and reduces process errors.

6. Limited error handling for external service integration

The workflow depends on external services such as Make.com for sending RFQ emails and generating contract drafts. This demonstrates service integration well, but the current BPMN has limited error handling for technical failures — a Make.com webhook could be unavailable, a payload malformed, or a response incomplete.

Improvement. Model technical error handling directly in BPMN. Service tasks would include retry mechanisms, boundary error events, and fallback paths. If an external service fails after several retries, the workflow could create a manual correction task for procurement — making the process production-ready.

7. No negotiation loop included

The current workflow evaluates supplier responses and selects the best supplier. It does not include a structured negotiation loop. In real procurement, the first offer is rarely the final offer — teams typically negotiate price, delivery time, payment terms, warranty, or technical specifications before deciding.

Improvement. Add a negotiation sub-process before the final supplier selection. After reviewing responses, procurement could decide whether negotiation is required. If yes, the workflow would send negotiation requests to selected suppliers, wait for updated offers, and re-evaluate them through the DMN — aligning the process with realistic procurement practice.

8. Implicit tie-breaking in best-supplier selection

The current best-supplier selection may include implicit tie-breaking. If two or more suppliers achieve the same score, the FIRST hit policy of the DMN automatically picks the first matching supplier. This works technically but is not transparent: a procurement employee may not understand why one supplier was preferred when both scored identically.

Improvement. Define explicit tie-breaking rules in the DMN. For example: if scores are equal, compare compliance level first, then delivery time, then price. If there is still no clear winner, route to a manual review task — making the selection logic transparent and easy to justify.

9. Limited process monitoring and reporting

The current workflow focuses on executing the procurement process. It does not yet provide advanced monitoring or reporting. Procurement managers cannot easily see how many RFQs are open, which suppliers have not responded, which cases are delayed, or how long each step takes.

Improvement. Add monitoring dashboards showing open process instances, average response times, supplier participation rates, decision outcomes, and bottlenecks. This helps procurement managers control the process and identify opportunities for continuous improvement.

10. Digital signature integration is missing

The contract signing step is not yet fully digitalised. The workflow generates and reviews a contract draft, but the final signature is handled manually outside the process. This creates a media break — the contract may need to be downloaded, printed, signed, scanned, and exchanged by email.

Improvement. Integrate a digital signature solution such as DocuSign or Adobe Acrobat Sign. After the draft is approved, the workflow would automatically send the contract for digital signature, store the signed contract, and continue or close the case without leaving the system.

⬆️ Back to Top

About

No description, website, or topics provided.

Resources

Stars

0 stars

Watchers

0 watching

Forks

Releases

Packages

Contributors