Consulting Firm: Internal Chatbot for 200+ Consultants

Context

Client name anonymized for confidentiality reasons.

This French consulting firm (strategy, organization, digital) employs over 200 consultants across 4 offices. Their intellectual capital relies on years of accumulated experience in:

• Project deliverables (presentations, reports, studies)
• Internal methodologies and frameworks
• Lessons learned and post-mortems
• Commercial proposals and references

The problem: each consultant was reinventing the wheel. New hires took 6+ months to know all available resources. Existing deliverables were only reused in 15% of cases.

The solution

We deployed an internal RAG chatbot accessible via their collaboration space (Teams/Slack), allowing each consultant to:

Instantly find relevant resources

"Show me examples of organizational diagnostics in the banking sector"

"What methodology should I use for a digital transformation project?"

Benefit from collective experience

"What are the classic pitfalls on a cloud migration project?"

"Who on the team has worked on ESG topics?"

Accelerate deliverable production

The system automatically suggests templates and reusable sections based on the current project context.

Technical architecture

• Automatic ingestion: Daily scan of shared drives and project spaces
• Classification: Automatic categorization (sector, mission type, client)
• Permissions: Respect for existing access rights (confidential, restricted, public)
• Integration: Slash commands in Teams and Slack

Results

After 4 months of deployment:

The increase in searches (+220%) shows massive tool adoption. Consultants now actively search the database rather than recreating.

Testimonial

"It's become our consultants' number one reflex. Before starting a project, they query the chatbot to see what already exists. We're finally capitalizing on our collective experience."

— Knowledge Management Director

Tech stack

• Python, FastAPI
• Qdrant (vector database)
• OpenAI GPT-4o
• Microsoft Teams / Slack SDK
• Docker

---

Project completed in 8 weeks. Contact us for a similar project.