AI audit for a B2B distribution SME
A manager convinced AI could help his business, but with no idea where or how. Two days of analysis, a 10-page report, three prioritized opportunities. He now knows exactly where to start.
Representative scenario based on real engagements. Names, figures and contexts have been adapted for confidentiality reasons.
Context
An 18-person B2B technical equipment distributor. The manager follows the news, has tested ChatGPT, has read the articles. He’s convinced there’s something to do with AI in his company. But what, exactly? And in what order?
He’d almost signed with an agency offering “a complete AI strategy” for 25,000 EUR without having looked at a single thing about his business. He decided to find someone who looks first before selling.
What we did
Two days inside the company. No questionnaire, no generic slides. Conversations with the manager, the admin lead, two salespeople, and the warehouse manager. Observation of the tools in use, the real workflows, the daily friction points.
The goal: find the 3 places where AI changes something concrete, not just theoretical.
What we found:
Incoming purchase order processing - 40 to 60 per week, all in different formats (PDF, email, sometimes a photo of a paper form). An admin was spending 2 hours a day re-entering data into the ERP. Use case #1, estimated ROI at 3 months.
Overdue payment follow-ups - handled manually, often forgotten, no consistent tracking. An automated follow-up pipeline with a tone adapted to the client and integrated history. Use case #2.
Searching order history to answer client questions (“did you ever deliver this product in this dimension?”) - currently 10 to 15 minutes per search. An internal assistant on top of the orders database. Use case #3.
The deliverable: an 11-page report. For each use case: problem description, recommended solution, estimated effort, projected ROI, identified risks. Plus a “what we wouldn’t do” section - ideas that seemed promising but didn’t hold up under analysis.
What followed
The manager started with use case #1. We delivered the tool 6 weeks later.
Audit methodology
- Semi-structured interviews, on-site observation
- Analysis of existing data flows
- ROI sizing based on the company’s actual data (not generic industry benchmarks)
- Structured report, no jargon, readable by a non-technical person