Transform product engineering knowledge into AI consultant service
Transform product engineering knowledge into AI consultant service
Many companies start to use RAG for knowledge management, but IPEK aims to go further by transforming systems-engineering methodologies into AI-driven consulting workflows. During my time at IPEK, I built two AI product-management consulting MVPs that automate the product-engineering consulting process. These solutions are already in use within the institute and demonstrate strong commercial potential as a low-cost, scalable offering for small and medium-sized enterprises.
Many companies start to use RAG for knowledge management, but IPEK aims to go further by transforming systems-engineering methodologies into AI-driven consulting workflows. During my time at IPEK, I built two AI product-management consulting MVPs that automate the product-engineering consulting process. These solutions are already in use within the institute and demonstrate strong commercial potential as a low-cost, scalable offering for small and medium-sized enterprises.
Product defination
Product defination
The product functions are based on two academic papers published by IPEK. My responsibility is to distill the key information from them, design workflows that enables AI to conduct a consulting-style conversation, and provide the AI with the institute's database to support analyze the user's situation.
The product functions are based on two academic papers published by IPEK. My responsibility is to distill the key information from them, design workflows that enables AI to conduct a consulting-style conversation, and provide the AI with the institute's database to support analyze the user's situation.
Project challenge
Project challenge
Extreme time pressure
Extreme time pressure
Only 100-hour contract to deliver 2 AI chatbot MVP from requirements through deployment.
Budget constraints
Budget constraints
Extremely limited financial resources requiring smart technical choices to minimize development costs.
Technical debt
Technical debt
One of the PM consultant service has a legacy software "InnoFox" built in 2015 with poor usability, outdated technology stack.
My role
My role
My role
There isn’t an official title for my role in the project, but I am the only person responsible for delivering the MVP, reporting directly to Thomas Völk, a researcher at IPEK. My work covers everything from stakeholder interviews and technical architecture decisions to user validation.
There isn’t an official title for my role in the project, but I am the only person responsible for delivering the MVP, reporting directly to Thomas Völk, a researcher at IPEK. My work covers everything from stakeholder interviews and technical architecture decisions to user validation.
There isn’t an official title for my role in the project, but I am the only person responsible for delivering the MVP, reporting directly to Thomas Völk, a researcher at IPEK. My work covers everything from stakeholder interviews and technical architecture decisions to user validation.
MVP tech stack research
Due to the extreme time pressure and the budget constraints, the most critical decision in this project was determining the technical architecture. I conducted comprehensive research comparing three primary approaches:
MVP tech stack research
Due to the extreme time pressure and the budget constraints, the most critical decision in this project was determining the technical architecture. I conducted comprehensive research comparing three primary approaches:
APPROACH 1
RAG with Python Backend
[User Browser]
↓ HTTPS
[Frontend (Streamlit)]
↓ REST / GraphQL
[Backend API (Python)]
├─► Vector DB (Pinecone)
├─► Ingestion (Langchain)
├─► LLM Provider (OpenAI GPT-4.1 nano)
└─► Auth & Logging (AWS Cognito + CloudWatch)
Full control over system
Highly customizable
˜320 hours development timeline
Monthly cost $10-$50
Significant technical risk
APPROACH 2
n8n Workflow
[User Browser]
↓ HTTPS
[Frontend (Streamlit)]
↓ REST
[n8n Workflow Engine]
├─► Vector DB (Pinecone)
├─► Ingestion (LangChain, Google Drive node, Dropbox etc.) ├─► LLM Provider (OpenAI GPT-4.1 nano)
└─► Auth & Logging (AWS Cognito & winston library)
Visual workflow design interface
Faster than custom backend
˜180 hours development timeline
Monthly cost $29-$54
Potential scalability concerns
SELECTED APPROACH
Custom GPT
[User Browser]
↓ HTTPS
[ChatGPT Interface]
↓ Native Integration
[OpenAI Platform]
├──► Custom GPT Configuration
├──► Knowledge Base (File Upload)
├──► GPT-4 Model (OpenAI Hosted)
└──► No additional infrastructure
Rapid prototyping and deployment
Up to 60 hours development
No cost for institute, use personal account
Ability to validate product-market fit rapidly
APPROACH 1
RAG with Python Backend
[User Browser]
↓ HTTPS
[Frontend (Streamlit)]
↓ REST / GraphQL
[Backend API (Python)]
├─► Vector DB (Pinecone)
├─► Ingestion (Langchain)
├─► LLM Provider (OpenAI GPT-4.1 nano)
└─► Auth & Logging (AWS Cognito + CloudWatch)
Full control over system
Highly customizable
˜320 hours development timeline
Monthly cost $10-$50
Significant technical risk
APPROACH 2
n8n Workflow
[User Browser]
↓ HTTPS
[Frontend (Streamlit)]
↓ REST
[n8n Workflow Engine]
├─► Vector DB (Pinecone)
├─► Ingestion (LangChain, Google Drive node, Dropbox etc.) ├─► LLM Provider (OpenAI GPT-4.1 nano)
└─► Auth & Logging (AWS Cognito & winston library)
Visual workflow design interface
Faster than custom backend
˜180 hours development timeline
Monthly cost $29-$54
Potential scalability concerns
SELECTED APPROACH
Custom GPT
[User Browser]
↓ HTTPS
[ChatGPT Interface]
↓ Native Integration
[OpenAI Platform]
├──► Custom GPT Configuration
├──► Knowledge Base (File Upload)
├──► GPT-4 Model (OpenAI Hosted)
└──► No additional infrastructure
Rapid prototyping and deployment
Up to 60 hours development
No cost for institute, use personal account
Ability to validate product-market fit rapidly
Project outcome
Project outcome
AI consultant 1
AI consultant 1
AI consultant 1
InnoFox coach
InnoFox coach
InnoFox coach
InnoFox Coach is a smart online consultant help product development teams systematically choose the right engineering methods. Rooted in the iPeM (Integrated Product Engineering Model) and guided by the SPALTEN problem-solving framework, it evaluates a team's development goals, current phase, resources, and desired outcomes to recommend fitting, practical methods from a curated database.
AI consultant 2
AI consultant 2
Project type assessor
Project type assessor
This AI advisor diagnoses a team’s development style and project context, then maps it to a best‑fit reference path to reduce inconsistency risks in engineering workflows. It targets variability in process models, documentation discipline, risk management, and tool support that commonly lead to cross‑view inconsistencies in product development.
AI consultant 2
Project type assessor
This AI advisor diagnoses a team’s development style and project context, then maps it to a best‑fit reference path to reduce inconsistency risks in engineering workflows. It targets variability in process models, documentation discipline, risk management, and tool support that commonly lead to cross‑view inconsistencies in product development.
50%
Completed in half the planned time.
50%
Completed in half the planned time.
100%
Budget saving.
100%
Budget saving.
50+
Graduate students used.
50+
Graduate students used.
50%
Completed in half the planned time.
100%
Budget saving.
50+
Graduate students used.
Expert feedack
A Trusted Partner
for Your Emotional Wellbeing
PM experts agreed that InnoFox GPT delivers strong, relevant method recommendations. However, the chat provides more explanations and steps than experts need—which is expected, because the product is intentionally designed for non‑experts in management.
Expert feedack
PM experts agreed that InnoFox GPT delivers strong, relevant method recommendations. However, the chat provides more explanations and steps than experts need—which is expected, because the product is intentionally designed for non‑experts in management.