Your best people just got twice as fast. Congratulations. You have a problem.
Ross Barnes has seen this movie from the projection booth. Twenty-five years in tech and advertising, ending as Global CTO of a WPP agency he helped scale from 15 people to more than 1,500 and from $80M to over $1B in revenue. Toyota. Lexus. EA Sports. British Airways. He sat on a global board and drove agile transformation across every function of a company growing faster than its own org chart could redraw itself.
Then he left to build Galahad, and the tagline he chose tells you exactly what he learned on the way out: build the machine, protect the human.
Here is the trap he watches companies walk into. A gaming client's engineering team was flying. Claude Enterprise across the org, shipping 10X features every week, genuinely at the frontier. And every one of those features needed legal sign-off, because gaming is a regulated business. Legal was not using AI. So the engineering team was not shipping faster. It was building a snowball and rolling it downhill into a team that had no way to catch it.
That is the insight Ross has built a consultancy around. AI adoption is not man versus machine. It is your fastest team versus your slowest team. And if you only fund the fast ones, you have not created leverage. You have relocated your bottleneck to someone who never asked for it.
His answer is almost aggressively unfashionable: raise the floor, not the roof. Get everyone to a baseline of capability before you go build a superhuman pocket of one department. The org that wins is not the one with the best AI users. It is the one with the fewest people who cannot keep up.
Which is why his enablement sessions do not touch a tool for the first two sessions. In any leadership cohort, someone has been living in Claude for two years and someone has never opened it, and you cannot predict which is which from a job title. So you level set. Then, and only then, you ask the question almost nobody asks first: what do you actually want this to accomplish? Handing a team a seat and saying "crack on" does not produce adoption. It produces a room full of people quietly wondering whether they are training their own replacement.
The conversation gets genuinely contentious in one place, and it is the best part of the episode. Ryan builds agents with personalities, accents, backstories, the full treatment, partly because a charming agent converts better. Ross builds agents with tone too, but draws a hard line at personification, and his reason is not squeamishness. It is accountability. The moment an agent feels human, you start grading it like a colleague, extending it the benefit of the doubt you would extend a person, and quietly outsourcing your own judgment to something that has none. It is a machine. Treat the inputs and outputs exactly like that. You are accountable for what you ask it and you are accountable for checking what comes back.
Ryan pushes on the word "coercive." Ross pushes back, and the exchange surfaces the real question underneath consumer AI: where is the line between charming and manipulative, and who is responsible for drawing it? Ross's position is that the guardrails are the job. No overclaim. No false validation. Build them in before the agent meets a customer, not after.
Galahad Group: https://galahadgroup.co.uk
Ross Barnes on LinkedIn: https://www.linkedin.com/in/rossbarnes
ImNotOK: https://www.imnotok.uk
Host and Show
https://www.linkedin.com/in/estesryan/
#1 AI Founder Newsletter! -
https://aiforfounders.co
Build your audience for life! -
https://inboxalchemy.co/
If you're not AI native; you're not getting the job! -
https://ainativestudent.com/
Get 35% off any supplement subscription with Momentous! -
https://crrnt.app/MOME/8RDrnXDd
Use code RYAN30 to save $30 on your AI men's fashion stylist!
https://taelor.style/
Your podcast's autonomous AI sponsorship agent! - https://gethowdi.com/
