AI Intern – After-Sales Assistant System

During my internship at Ahooga, I developed an AI-powered after-sales assistant designed to support dealers and internal after-sales users with technical questions. The project focused on making company knowledge easier to access by connecting product documentation, processed video content, and structured support information to an intelligent retrieval-based assistant.

The system was built using FastAPI, Azure OpenAI, Azure AI Search, Azure Blob Storage, Python, Docker, and Odoo integration. Instead of generating generic AI answers, the assistant first retrieves relevant company knowledge and then generates context-aware responses based on that information.

Due to company confidentiality, a public live demo is not available.

View internship documents

Duration

Feb 2026 – May 2026

Tools

FastAPI, Azure OpenAI, Azure AI Search, Azure Blob Storage, Python, Docker, Odoo API, GPT-5

The Goal:

The goal of this internship project was to improve Ahooga’s after-sales support process by making technical knowledge easier to find and use. The assistant was created to help dealers and internal after-sales users receive faster, more consistent, and more reliable answers to technical questions.

The project focused on transforming scattered technical documents, videos, and support information into a structured AI-accessible knowledge system.

The Challenge:

One of the main challenges was building an AI solution that could be useful in a real company environment while still remaining reliable and controlled. The assistant needed to understand different bike configurations, retrieve the correct information, and avoid giving generic or unsupported answers.

Another challenge was preparing unstructured support data, such as Odoo tickets, so that it could later be validated and used safely without exposing irrelevant or sensitive information.

Challenges of the Ahooga internship project

The Architecture:

I designed a modular AI system for after-sales support, focused on delivering reliable, product-specific answers rather than generic chatbot responses. The architecture separated the user interface, backend logic, retrieval layer, AI generation layer, storage, and internal API services.

To improve reliability, the system used a Retrieval-Augmented Generation workflow. This means the assistant did not simply generate answers freely. It first searched the structured knowledge base and then generated responses using the retrieved company information as context.

The system also included an internal API layer and the Arthur interface. The Arthur interface is an internal management screen where after-sales employees can add, correct, or approve specific answers without changing the backend code. These parts made the solution easier to maintain and gave the company a foundation to continue improving the assistant after the internship.

Frontend Interface

↓

FastAPI Backend Layer

↓

Azure OpenAI

Azure AI Search

↑

Knowledge Ingestion Pipeline

↑

Ticket Cleaning / Validation Layer

↑

Odoo API / Company Data

Final Outcome:

By the end of the internship, the project had evolved into a functional dealer-focused AI assistant for the MAX bike platform. The assistant used structured retrieval, metadata filtering, and GPT-5-based response generation to answer technical questions more accurately and consistently.

The internship also delivered a structured knowledge base, processed video content, an Odoo ticket preprocessing workflow, the Arthur interface for response control, and a centralized internal API layer. These elements created a solid foundation that Ahooga can continue building on after the internship.

Internship Documents:

The mandatory internship documents are available below. These documents provide a more detailed overview of the project planning, implementation, technical realization, and personal reflection.

Project Plan

Initial planning, scope, objectives, and project approach.

Download PDF

Realization Document

Full technical explanation of the internship project and implementation.

Download PDF

Reflection

Personal reflection on the internship experience, growth, and learning process.

Download PDF

Learning Now, Leading Soon

Get in touch

Feel free to contact me if you would like to know more about my work, internship project, or portfolio.

+32 484 20 90 20

paschalfrancisifewulu@gmail.com