Reference format
Understand enterprise AI under the hood, without becoming a technician.
A short, concrete, live-projected session that shows what happens when an application calls a large language model. No ChatGPT tutorial. No code to write. A durable reading grid for evaluating tools, vendors, and internal AI projects.
Duration
2 to 2.5 hours
Format
In person, code projected and explained
Audience
Executives, managers, entrepreneurs
Prerequisite
No technical prerequisite
Why this session
The offers change. The building blocks stay the same.
Assistants, copilots, low-code platforms, and agents change names, interfaces, and commercial promises quickly. Under the hood, they often rely on the same building blocks: messages, context, model call, external memory, tools, and orchestration. Seeing those blocks once helps leaders ask better questions and brief technical teams or vendors more clearly.
What becomes visible
A model call
OpenAI, Anthropic, Mistral, or Google: the interfaces differ, but the basic technical gesture is comparable.
Context makes the difference
The quality of an answer depends less on model magic than on the instruction, the data supplied, and the requested format.
Agents are orchestration
An agent is not a mysterious autonomous entity. It is a decision loop around a model, tools, and stopping rules.
What participants leave with
What participants leave with
01
A stable mental model of what happens when an AI tool calls a language model.
02
The basic vocabulary: token, context, embedding, RAG, tool call, agent.
03
A practical reading grid to separate the real mechanism from vendor marketing.
04
A one-page specification template to brief a technical team or external provider.
insideAi.formatsLabel
Executive session
2 to 2.5 hours
The reference format for a chamber of commerce, professional association, or executive committee that wants to give members or teams a shared foundation.
- · Commented demonstration of calls to major model providers
- · Explanation of key concepts: tokens, context, memory, RAG, tools, agents
- · One-page specification template
- · Vendor question checklist
For: Executives, managers, entrepreneurs, professional associations
Half-day deep dive
3.5 to 4 hours
For teams that want to move from understanding to first operational framing. The session adds a workshop on two business processes and the first governance decisions.
- · Full path: context, stuffing, RAG, tool calling, agents
- · Framing workshop on two real use cases
- · Initial list of risks, dependencies, and decisions to clarify
- · Extended specification template and vendor checklist
For: Executive committees, business units, innovation teams
Format references
First delivered to a private executive circle in Casablanca on May 20, 2026. The same format is scheduled for the Belgian-Luxembourg Chamber of Commerce in Morocco on June 4, 2026. Built for non-technical executives and managers who want to understand the mechanism before deciding.
Why this format
This session does not replace an AI strategy or a full operating model. It creates the shared foundation that is often missing before selecting tools, launching a project, or briefing a vendor.