Format: Full-day workshop (6 hours)
Target audience: CTO/IT manager + selected key people from the business
Background
A service company with 80 employees had invested in AI licenses for six months without seeing results. The problem wasn't the tools—it was that no one used them seriously.
The CEO put it like this: "We've bought cars but haven't built any roads."
What We Did
Part 1: Inventory of current state (90 min)
We mapped: - Which AI tools were available - Who actually used them (and for what) - Where data lived (and who had access)
This pattern was familiar. Tools existed, but they were isolated. The AI couldn't "see" the company's data, CRM, project tools, or documentation, without manual copy-paste.
Part 2: "The MCP mindset" (120 min)
We introduced the concept from Shopify: making internal data available to AI through structured connections (MCPs—Model Context Protocols).
Practical exercise: Participants identified their five most important data sources and prioritised which to connect first.
The priority list: 1. CRM (customer history, pipeline) 2. Project tools (status, time reports) 3. Document archive (quotes, contracts) 4. Email (communication history) 5. Finance system (invoices, budget)
Part 3: Security and governance (90 min)
The most critical part. AI with access to company data requires clear rules: - Who may use what? - How is customer data handled? - Where are AI interactions stored?
We built a simple governance matrix with the IT manager and legal lead.
Part 4: Pilot plan (90 min)
Instead of a large rollout we defined a 30-day pilot: - One team (sales, 6 people) - One data source (CRM) - One use case (preparation for customer meetings)
Results and decisions
The workshop resulted in:
1. Technical roadmap: Prioritised list of integrations with timeline 2. Governance document: First version of AI policy for the company 3. Pilot setup: Concrete plan to test with limited risk 4. Owner: Appointed "AI coordinator" to drive the work forward
Participant reflections
> "We thought the problem was that people didn't want to use AI. The problem was that AI had nothing to work with." >, CTO
> "The governance discussion was unexpectedly important. We hadn't thought of half of the questions." >, CFO
The workshop is suitable for
- Companies that have bought AI tools but don't see adoption
- IT departments that need a structured plan
- Leadership teams that want to understand what "AI infrastructure" actually means
Prerequisites
- Access to IT responsible for the full day
- Overview of current system landscape
- Mandate to make decisions about pilot projects
Next steps
We help you move from tools to infrastructure. Contact us for an initial review of your current situation.