Buy the frontier, not the known

Most of what boards pay consultants for has just been commoditised. The research, the benchmarking, the synthesis, the first-draft frameworks — a capable leader with a decent AI tool can now produce all of it in an afternoon, for almost nothing. A friend and former colleague of mine, Atif Rafiq, made the argument sharply in a recent piece: now that AI can do the known, the only thing worth buying from a consultant is work on the frontier. I think he's right, and I think it changes how any serious leader should buy outside help.

The arbitrage that's closing

For decades the consulting model rested on a quiet arbitrage. A firm would send a team of capable people to spend weeks learning your business — interviewing your staff, gathering your data, mapping your processes — and then sell that learning back to you as a recommendation. In effect you were paying to have your own knowledge assembled elsewhere and returned in a deck. The uncomfortable truth, more often than anyone admitted, was that the consultants learned more from the client than the client learned from them.

That arbitrage is closing. The work at the centre of it — gathering the known, benchmarking it, reasoning through the commodities, packaging the result — is exactly what a good model now does quickly and cheaply. The price of assembling the known is falling towards zero. Which leaves one honest question for any leader about to spend real money: which part of this problem genuinely needs an outsider, and which part am I just paying someone to learn?

The known and the frontier

The useful distinction Atif draws is between the known and the frontier. Most consequential decisions come with a dozen things you'd need to understand before you could make them. The majority of those are knowable — they sit in your data, your market, your own experience, or in analysis a model can now generate on demand. A few are genuinely uncertain: the calls where no amount of research settles it, where the value is in judgement rather than information.

The known is now cheap to clear. The frontier is where the money should go. Paying frontier prices for known work is the mistake the old model was built on, and it's the one AI now makes obvious.

Build the brief before you buy

There's a practical test in this for any CEO or board, and it costs nothing. Before you take a problem to market, write it down properly. List everything you'd need to know to decide. Mark the items you — or your own team, with the tools now available — can reason through without help. What's left, the genuine unknowns, is the only part worth buying. Take that to market, not the whole problem.

Do this and two things happen. You spend far less, because you're no longer paying a team to assemble what you already half-knew. And you buy better, because you're contracting for judgement on the hard part rather than for effort on the easy part. It's a discipline more leaders will adopt as the cost of the known keeps falling.

What it means for how you buy leadership

Follow the logic and it changes not just what you buy, but who you buy it from. If the scarce thing is judgement on the frontier, then the real asset is transferable judgement — someone who has carried the accountability, made the platform call, redesigned the operating model, navigated the board and lived with the consequences. That doesn't come from a methodology or a brand. It comes from having done the job.

This is the case for operators over firms. A large engagement team gives you capacity to work through the known; an operator who has held the role gives you judgement on the frontier. As the known collapses in price, the second is worth more and the first is worth much less.

How I work

It's why I don't work the way a traditional consultancy does. I embed as a leader, not an adviser at arm's length. I use AI deliberately to clear the known work, so the time and the cost go to the decisions that actually move the business — not to reproducing analysis a model can now generate on its own. And I build capability into the team rather than dependency on mine: when I leave, the knowledge stays, because it was never extracted in the first place.

A traditional engagement leaves you with a recommendation. This one leaves you with a stronger organisation — an owned platform where you were dependent on vendors, a product-led operating model where you had a services mindset, and a leadership team that can hold the line after I've gone. If that's the kind of partner your situation calls for, get in touch — or take a look at how I work.

Credit: this piece builds on Atif Rafiq's argument about consulting and the frontier.