AI Product Quality & Risk Workshop

1 day

A practical one-day workshop for teams shipping AI features responsibly.

Duration: 1 day
Format: Online or in person
Audience: Product, QA and delivery teams building AI features
Group size: 6 to 16 people
Delivery: Private team workshop or open cohort

Who it's for

This workshop is for:

Product and QA teams shipping AI-powered features
Delivery leads responsible for release quality
Teams that have shipped an AI feature and want to test it properly
Organisations building internal AI risk and release practices

The problem it solves

AI features fail in ways teams don't expect: confidently wrong answers, inconsistent behaviour, and risks that don't show up in a demo. This workshop builds a practical way to find those problems before your users do.

What this workshop covers

Hallucinations and failure modes
RAG testing
Prompt testing
Red-teaming
Evaluation
Release gates
Responsible AI

Agenda

09:30See how AI failsHallucinations, failure modes and why demos hide risk

10:45Test the retrievalRAG testing and where grounding breaks down

12:00Lunch

13:00Test the promptPrompt testing and a red-teaming exercise

14:30Build the evaluationDesigning an evaluation set that matters

15:30Set the release gateRelease gates and responsible AI checkpoints

16:30Wrap-up

What participants will leave with

You'll leave with:

A practical map of failure modes to test for
A working evaluation set for one AI feature
A red-teaming exercise completed on real output
A release gate checklist your team can reuse
A responsible AI checkpoint for future releases

How we work during the session

This is a working session, not a day of lectures. Expect discussion, exercises, challenge, examples and practical work based on your own product or business context.

What's included

Workshop facilitation
Worksheets and templates
A digital workbook
Practical exercises based on your own context
A post-session summary and action plan

Delivery options

Half day: A focused introduction or leadership session.
1 day: The standard, practical working session.
2 days: A deeper team working session across two days.
Where: Remote, on-site in the UK, or international and US delivery available.

Who delivers it

Collins Obasuyi, Founder & Principal Consultant

Collins works across product strategy, AI, delivery, quality and education, and facilitates from real product and delivery work rather than abstract theory. Sessions are built around your team's actual context: real decisions, real trade-offs and practical next steps rather than generic frameworks.

Customisation

Every private workshop can be adapted around your organisation, product maturity, audience and current challenges.

FAQ

Is it suitable for non-technical teams?

This workshop goes deeper into testing technique than our other AI sessions. It's best suited to product, QA and delivery teams with some hands-on exposure to the AI feature being tested, rather than a fully non-technical audience.

Can this be delivered remotely?

Yes. Every workshop can be delivered remotely, in person, or as a mix, depending on what works for your team.

Can you tailor it to our organisation?

Yes. Every private workshop is adapted around your organisation, product maturity, audience and current challenges.

How many people can attend?

Group sizes typically range from 6 to 16 people, though this can flex depending on your needs.

What preparation is required?

Little to none. We may ask a few questions beforehand so examples and exercises reflect your actual context.

Do participants receive materials?

Yes. Every participant receives templates, a digital workbook and a written action plan afterwards.

Can this be delivered in the US?

Yes. Sessions can be delivered remotely to any timezone, and in-person delivery in the US can be arranged.

Start a conversation

Have a product or AI decision to make?

Useful first calls usually start with one unclear decision, a deadline and a team that needs a practical next move.

Tell us about it