Talks

AI in business, without the buzzwords (or the nap).

An AI talk should do more than run through abstract slides, buzzwords and impossible promises. I make generative AI and its implications understandable, engaging and practical! Have a specific need? Get in touch for a tailored format.

01

Rules & risks

AI governance: set clear rules without blocking adoption

The AI Act, CNIL, sensitive data, SaaS tools, human validation and accountability: how to set useful rules without making AI off-limits.

Between people who upload confidential company data to an unknown AI tool launched yesterday 'just to try it' and those who have never used an LLM for fear the world might collapse, there is room for a sensible approach. Active AI governance provides the compass.

This talk connects compliance, security and everyday business use, helping teams benefit from AI productivity gains while managing the risks.

  • Audience

    • executives
    • CIOs
    • CISOs
    • DPOs
    • legal
    • HR and business teams

  • Benefit

    learn to set clear, practical usage rules proportionate to actual risks

  • Approach

    compliance connected to technical practice

02

AI in production

Your AI proofs of concept will not reach production (and that is normal)

Why generative AI projects stall after the demo, and how to create the conditions for scaling up.

AI prototypes can look like a dream come true in a boardroom demo. Production can quickly become a nightmare of costs, data security and real-world performance...

This talk provides a practical framework: data, evaluation, security, workflows, costs, governance, LLMOps and team adoption.

  • Audience

    • executives
    • CIOs
    • innovation
    • data and business teams

  • Benefit

    know what to scale, what to stop and what to scope before investing

  • Format

    30–45-minute keynote, 60-minute talk + Q&A, 2-hour masterclass

03

AI agents

So what actually is an AI agent?

Understand what the word 'agent' really means: an enhanced assistant, an automated workflow, an autonomous system or a marketing rebrand?

They will book your tickets, answer your customers, run your workflows, negotiate with suppliers and perhaps even make the coffee: AI agents will do it all! Before handing them the office keys, how about checking what we actually mean?

This talk explains the differences between assistants, automation, agents and autonomous systems, then examines the issues that matter: data access, permissions to act, supervision and safeguards.

  • Audience

    • CIOs
    • CTOs
    • business leaders
    • software vendors and innovation teams

  • Level

    adaptable, from an executive overview to architecture and risks

  • Benefit

    understand what an AI agent can actually do, what you should be cautious about delegating and how to spot agentwashing

04

Engineering teams

From vibe coding to AI-assisted development

Copilot, Codex, Cursor, Claude Code and coding agents: how to develop faster while protecting architecture, tests, security and understanding of the code.

AI promises faster developers, tickets resolved in minutes and thousands of lines generated before the first coffee. Until everything breaks and nobody is quite sure how it works any more.

This talk explains how to integrate AI into team practice: delegation, code review, testing, documentation, security, technical debt, CI/CD and the role of tech leads in staying in control while benefiting from faster development.

  • Audience

    • CTOs
    • VPs of Engineering
    • tech leads
    • developers
    • IT consultancies and software vendors

  • Benefit

    speed up development while preserving the engineering standards that keep a product maintainable

  • Format

    45–60-minute talk, 2-hour event masterclass

Also available as team training.

Next step

Let's discuss AI in your organisation.

A short briefing is enough to identify the right format, level and audience.

Arrange a briefing→See client references→