Global Manufacturer

From Internal Chatbot to Enterprise AI Platform

A leading Danish manufacturer with 12,000+ employees across 40+ countries wanted to give every employee an AI assistant that actually understands their business -- one that can search company documents, find emails, check calendars, analyse data, and research the web, all through a single conversation.

Enterprise AI Manufacturing Microsoft 365

Enterprise AI Platform Architecture — multi-cloud (Azure + Google ADK), MCP tool integration, pgvector private data, authenticated SharePoint access

16 Weeks Idea to Live Platform

12,000+ Employees Served

8+ AI Models Best Model for Each Task

40+ Countries Global Organisation

The Client

A major Danish manufacturing company with 12,000+ employees across 40+ countries, producing critical materials for fire safety, energy efficiency, and sustainable construction worldwide.

Their digital innovation team had already built an internal AI chat tool for employees. But it was limited to basic conversation -- it couldn't access company data, understand uploaded documents, or integrate with the Microsoft 365 tools employees use every day. They needed to turn it into something that would genuinely change how people work.

The Challenge

A Chatbot That Couldn't Access the Business

No access to company knowledge -- couldn't search internal documents, policies, or historical data
Disconnected from daily tools -- no integration with SharePoint, Outlook email, or calendars
One-size-fits-all -- the same generic assistant for every employee, regardless of their role or needs
Stuck on one AI model -- no way to use different models for different tasks, or take advantage of new ones as they emerge

The gap between "AI demo" and "AI that transforms how people work" is vast. The manufacturer needed an experienced AI architect to bridge it.

What Employees Can Now Do

Every employee now has an intelligent assistant that understands their role, connects to their corporate systems, and uses the best AI model for whatever they need.

Search Company Knowledge

Upload documents -- PDFs, Word files, spreadsheets, images -- and ask questions in plain language. The assistant finds relevant information across everything it has access to, with sources you can verify.

Access Microsoft 365

Search SharePoint files and pages, find emails, and check calendar availability -- all from within a single chat conversation. No switching between apps to find what you need.

Analyse Data On the Fly

Upload a spreadsheet and ask for analysis. The assistant runs calculations, creates charts, and generates reports you can download -- turning hours of manual work into a conversation.

Choose the Right AI for the Job

Switch between AI models from OpenAI, Anthropic, and Google depending on the task. Need creative writing? Deep analysis? Image understanding? Pick the best model for the moment.

Personalised Per User

Each employee gets an assistant configured for their role, with access to the tools and data sources they need

Web Research

The assistant can search the internet to supplement corporate knowledge with external information

Transparent Reasoning

See what the assistant is doing in real time -- which tools it's using, what it's searching, and why

Enterprise Security

Full corporate authentication -- employees only access data they're authorised to see

The Outcome

From concept to live platform in 16 weeks -- now serving employees across the global organisation.

16 Weeks

From Prototype to Production

A complete enterprise AI platform -- designed, built, tested, and deployed. Comprehensive documentation and handover sessions ensured the internal team could maintain and extend the platform independently.

12,000+ Employees

Platform available across the global organisation

5 Specialised Capabilities

Document search, data analysis, web research, email, and calendar

3 AI Providers

OpenAI, Anthropic, and Google models available to every user

Full Handover

55 design documents, deployment runbooks, and training sessions for the internal team

Under the Hood

For the technically curious -- the key architectural decisions that made this possible.

AI Architecture

Google Agent Development Kit (ADK) Multi-model routing via LiteLLM Model Context Protocol (MCP) Vector search (PostgreSQL + pgvector)

AI Models Integrated

Azure OpenAI (GPT-5, GPT-4o) Anthropic Claude (Sonnet 4.5) Google Gemini (2.5 Pro, Flash)

Platform

Python / FastAPI Vue 3 / TypeScript Azure Web Apps Microsoft Graph API

Ready to Turn Your AI Prototype into a Platform?

Most organisations have an AI experiment. Few have turned it into something that genuinely changes how employees work. We can help you make that leap.

Schedule a Consultation View More Case Studies