Manifesto

Why Perch backs coaches when everyone else is building them out

A short statement on what AI actually changes and why the coach is the wrong job to automate.

By Dave Cheng · June 2026

The dominant story in AI right now goes something like this. The model gets smarter every quarter. The output gets cheaper every quarter. The pace gets faster every quarter. So whatever a human professional does today, the model will do tomorrow, or the day after, or the year after, but soon. The job of the human shrinks. The job of the model grows. The math is supposed to be obvious, but let's be honest the economics are still not proven.

Apply that story to business coaching and you get the conclusion that almost every AI coaching startup has already drawn. They are building products that aim to clone or replace the coach. Always-on virtual coaches. AI advisors. Chat-based business mentors with the personality of whichever LLM is fashionable that month. Coaching as a service, minus the human.

We think they've got this all upside down.

What actually gets cheaper

Information was the scarce thing for most of the last century. Knowing what your competitor is doing, knowing what strategy is working in your industry, knowing what the data says about your pipeline. These required research, time, expertise. A business advisor / coach earned their fee partly by bringing the information and knowledge that their client didn't have.

That advantage is gone. Information is now free, infinite and less than ten seconds away from any browser. A founder who spends ten minutes with a decent AI tool can get a market scan that would have cost £20,000 in consulting fees five years ago.

So the question is, what isn't getting cheaper?

What stays expensive

Three things.

Judgement. The act of looking at a set of options, weighing them against each other and balancing the founder's risk tolerance, their values, their wider life, to ultimately conclude, this one. AI can lay out the trade-offs beautifully. It cannot, in any meaningful sense, choose.

Focus. The act of saying no to the other twelve good ideas this week and keeping the founder on the two that matter. AI is a generator. It produces more options, not fewer. The work of saying "park that, come back to it in March, do this one first" is the work of someone who knows the founder, knows the industry and is the human accountability partner for the call.

Accountability. Someone who shows up next week and asks what happened. Not a notification. A person who sat across from you, looking into the whites of your eyes when you said you would do that one thing to help your business grow.

Those three are the entire reason a coach exists. They are also the three things AI is structurally bad at and not because the model is small or the training is incomplete. AI is bad at them because they are not knowledge problems. They are emotional problems that stall human decision making.

What Perch actually does

Perch braids two things that almost no one else is putting together.

A strategic AI that runs the analyst work overnight according to your business strategy. It scouts sales, marketing, finance, competition, in fact anything you choose. It surfaces what changed, what is working, what is at risk. It writes the brief that a Chief of Staff at a £200m company would write by hand. It does this every day, automatically, for every coach-and-client pair in the network.

A vetted global community of coaches who use that brief as their starting point, not their substitute. The coach walks into the weekly meeting having already read the analyst's work. They challenge it. They reframe it. They hold the founder to the decisions they made last week.

The coach gets sharper because they walk in with better data. The founder gets sharper because they walk in with a thinking partner who has the same data and has been thinking about it differently. The decisions get made in the room, human to human.

What this means for coaches

You become harder to replace, not easier. The AI does not threaten your job. It makes you faster, more rigorous and visibly more valuable to the kind of client who used to think "I might just use ChatGPT for this." When that client compares you to a chatbot, the gap is bigger, not smaller than it was a year ago.

We are recruiting the founding cohort now, on those terms. Annual membership. Per-client licences. One hundred percent of your fees stay yours. Perch never sells AI direct to your client. Your relationship is your relationship.

What this means for founders

You stop having to choose between intelligence and accountability. Most founders today are using AI tools privately, some successfully achieve short-term efficiency gains and feeling slightly guilty that they aren't acting on more of what comes out. Perch gives you the same level of intelligence, integrated into a relationship with someone who reads it, challenges it and holds you to the parts that actually matter. The thinking is no longer optional. The accountability is no longer absent.

Why now

The window is small. Coaches are deciding right now, in 2026, whether to adapt to AI or whether to wait it out. In our opinion those who adapt early own the next decade. Those who wait, won't.

Perch was built to help founders refocus on the importance of emotional decision making in their businesses. By leveraging AI and insisting it is braided together with professional human judgement ensures that you make the most valuable decisions. 

We think that's the better bet.

The next decade belongs to the ones who wish to re-write the rules. We're backing the people doing the reading.

Dave Cheng is co-founder of Perch, a UK business coaching platform that braids a vetted coach community with proprietary strategic AI. Perch is opening to its founding cohort of coaches now.