What a generative AI session usually is, and what this one is
The standard version explains how the models work, shows what they can do, and closes on prompting. It is useful once, it dates within a year, and an audience of professionals has usually seen most of it already.
This session takes the tools as read and asks the question that survives the next release. Where does the value actually land, what does it consume on the way, and what does the organisation still have to be capable of checking. Those answers come from measurement rather than from a product roadmap, so they keep.
The gains are real and they are not evenly distributed
Across 5,172 customer-support agents and three million chats, resolutions per hour rose 15 per cent on average. The average hides the finding: the least experienced gained about 30 per cent, rising to 36 in the lowest skill quintile, and the most skilled saw no significant gain at all. The evidence, graded.
So a room being told generative AI will make everyone more productive is being told something the best study of it does not support. It substitutes for expertise people do not have. Where the expertise is already there, it adds much less, and what it adds is unpredictable at the individual level.
Capability is the cost nobody prices
Nineteen endoscopists averaging 27.6 years of experience lost six percentage points of detection in procedures performed WITHOUT the tool, within months of their centres adopting it. That is measured deskilling in people who had spent decades acquiring the skill. Capability debt, with dates.
The same mechanism runs through the pipeline. The first draft, the routine analysis and the ordinary case are the tasks a junior was given because doing them badly and being corrected is how the judgement gets built. Automate the reps and the capability stops forming, quietly, for a cohort at a time. The missing rungs.
Adding a human to the loop is not the answer either
The obvious response is to let the model produce and the person check. A preregistered meta-analysis of 106 experimental studies covering 370 effect sizes found human and AI combinations performed significantly worse on average than the better of human or AI alone, at a Hedges’ g of -0.23, with the losses concentrated in decision-making and the gains in content creation.
That is not an argument for removing the person. It is an argument for specifying the pairing: which decision, made by whom, on what grounds, checked how. Why the loop is not a safeguard.
What the session covers
Where the gains land. Who in your organisation gets more from these tools, who gets less, and why the assumption runs the wrong way round.
What they consume. The capability being spent to buy today’s speed, and how to tell whether it is being spent in your setting.
What has to stay checkable. Which outputs need a person who could have produced them unaided, and what that implies for how work is assigned.
The pipeline. What happens to the people who were supposed to become senior when the junior work is the work that automates first.
The room is then placed on the Drift versus Design Matrix and asked to argue about where it sits.
Who this is for, and who should book somebody else
Right for an organisation two or three years into generative AI, where adoption is no longer the interesting question and somebody senior has started wondering what the work will look like in five years.
Wrong if you want the tools demonstrated, the models explained, prompting taught, or a vendor comparison. Those are real needs and there are people who do them well. This session assumes the audience has already used the tools and wants the argument about what they are doing to the organisation.
It is a keynote rather than a training programme. Nothing is deployed and no tool is recommended. The advisory work behind it is at AI adviser to CEOs, boards and leadership teams.
The three keynotes, and which fits this room
Keynotes run 30 to 90 minutes, in person or virtual. A 30-minute main-stage version of each talk is available for conferences; the full argument needs 40 or more. The sector examples change for the room; the argument does not. All three, with the showreel, are at keynotes, and the programme copy is at the speaker pack.
Drift versus Design
Drift versus Design: why most organisations hand their judgement to AI without deciding to, and how to design your way through instead.
Most organisations are adopting AI by drift: a thousand reasonable decisions that add up to judgement nobody chose to give away. Rahim Hirji, author of SuperSkills, shows leaders which mode they are running, where AI sharpens judgement and where it weakens it, and hands them the controls.
We Are Superheroes
We Are Superheroes: AI is the suit. The human decides.
AI makes everyone faster, stronger and more capable. It does not decide; people do. Through a family story across four generations and three continents, Rahim Hirji hands the audience the seven human skills that grow more valuable as the tools spread. They arrive thinking AI is the story and leave knowing they are.
WTH (What the Human)
WTH (What the Human): a live test of a board's own judgement.
Business is being rewired as AI arrives, and human judgement is leaving with it. In three acts, Rahim Hirji shows a board what AI is changing in how organisations run, then ends with a live test of the board's own judgement, in the room and in real time. Nobody forgets the result.
