The stakes test: the higher the risk, the harder they check
The more a decision costs to get wrong, the harder the decision-maker leans on proof over claims. Your buyers sit at the high-stakes end of that scale, and so does the way AI systems treat you.
Lucy Aingworth
AI Visibility Consultant
Published 7 min read
The short answer
The stakes test is the rule that the more a decision costs to get wrong, the harder the decision-maker leans on credibility of source rather than the claim itself. At low stakes, a clear promise can be enough to win. At high stakes, the buyer demands proof and outside validation before acting. AI systems apply the same test: when a recommendation carries weight, they privilege the entity they can verify over the one they have to guess about. Founders and executives sell at the high-stakes end, which is exactly where a claim without proof loses.
Key takeaways
Credibility of source scales with the stakes of the decision. Low stakes forgive an unproven claim. High stakes do not.
Your buyers, founders, CMOs and executives, sit at the high-stakes end, where getting the choice wrong is expensive and visible.
A clear claim with nothing behind it is a Promising Claim. At high stakes it loses to a Recognized Authority, whose claim is backed by proof.
AI systems run the same test, leaning hardest on verifiable entities exactly when the recommendation matters most.
The lever is Credibility, the middle corner of the Recognition Triangle, and it is built, not asserted.
Every buyer runs a version of the same test before they act, whether they notice it or not. It is a single question. If I choose you and I am wrong, how badly does that hurt me. The answer decides how hard they check. When the cost of a bad choice is small, they will take your word for it. When the cost is large, your word is nowhere near enough, and they go looking for proof.
This is the most useful thing to understand about why some businesses win high-value work and others, often just as capable, do not. It is not that the winners made a better claim. It is that at the point where the decision actually mattered, they had the credibility to back the claim and their competitor did not.
What is the stakes test?
The stakes test is simple to state and easy to feel. The higher the consequence of getting a decision wrong, the more weight the decision-maker puts on credibility of source rather than on the claim itself. A confident assertion carries a low-stakes decision. It cannot carry a high-stakes one, because at the high end the buyer is no longer asking whether the claim sounds good. They are asking whether they can afford to be wrong about you.
There is a quiet reason the verified option keeps winning at that end, and it is worth naming. A claim you have to take on faith is work for the buyer. They have to assess it, discount it, and carry the risk of it themselves. A claim that is already backed by proof and outside validation is lower effort to accept and lower risk to stand behind. When the decision matters, people default to the option that is easier to justify later, and the proven one is always easier to justify.
Does the bar really change with the stakes?
It changes more than most businesses plan for, and you can watch it climb across a single career of work. A consumer choosing a beauty product is making a low-stakes decision. If it disappoints, they are out the price of one item and they move on, so a clear, appealing claim is often enough to earn the sale. Move up to a business choosing a software platform or an advisory partner, and the stakes rise sharply. Now the wrong choice costs budget, months, and the credibility of whoever signed off on it. The claim alone stops being enough. The buyer wants evidence, references, a track record they can check.
Keep climbing to an enterprise selecting a supplier for a regulated, multi-region contract, and the bar is at its highest. Here the decision is scrutinised by people who were not in the room, and a mistake is visible to all of them. At that altitude, a business that cannot be independently verified is not a risk anyone will take, however good its pitch.
The claim does not change as you climb. What the buyer requires behind it does.
This is why the same message can win at one level and be quietly ignored at another. It was never really the message being judged. It was whether the reality behind it could bear the weight of the decision resting on it.
What separates a promising claim from a recognised one?
The Recognition Triangle gives this a precise shape. Its three conditions are Clarity, what you stand for; Credibility, what can be verified; and Earned Authority, what others say when you are not in the room. Where a business sits against those three produces a recognisable pattern, and two of those patterns explain almost everything about who wins high-stakes work.
Promising Claim vs Recognized Authority
A Promising Claim has strong Clarity and little else. It states plainly and well what it does, but the Credibility and Earned Authority behind it are thin, so the claim rests on assertion. A Recognized Authority has all three conditions strong and aligned: the clear claim, the verifiable proof, and the outside validation. Both are patterns of the Recognition Triangle. At low stakes they can look identical. At high stakes they are not remotely the same.
A Promising Claim can win small decisions all day, because at low stakes clarity is doing enough of the work. The trouble is that its owner often assumes the same claim will carry the large decisions too, and it will not. The moment the stakes rise, the buyer reaches past the promise for the proof, and if there is none, the work goes to the Recognized Authority whether or not their claim was actually sharper. Frequently it was not. They simply could bear the weight.
Find your pattern
Which of the six patterns is yours right now, and what it costs you.
Everything above described human buyers. The reason it matters more now, not less, is that an AI system recommending a supplier is itself a decision-maker running the stakes test, and running it at scale. When it is asked who to recommend for a low-stakes query, it can afford to be loose. When it is asked who a founder should trust with something consequential, it makes a high-stakes call, and like any careful decision-maker it leans hardest on the entity it can verify.
The verified entity is lower risk for the system to name and lower effort to stand behind, the same logic a human uses, applied by a machine at the moment of recommendation. So the harder the stakes of the query, the harder the model leans toward the business with proof it can read, and away from the one making a claim it cannot check. Your buyers ask the high-stakes questions. That means the machine is applying its most demanding version of the test to you, precisely where you can least afford to be the unverifiable option.
None of this is solved by making a bolder claim. Clarity you already have will not move a high-stakes decision on its own. What moves it is Credibility, the middle corner, built deliberately through proof and outside validation until your claim is one that both a cautious buyer and a cautious machine can afford to bet on. That is the difference between being a Promising Claim and being the answer.
Common questions
What is the stakes test?
The stakes test is the simple rule that the more a decision costs to get wrong, the harder the decision-maker leans on credibility of source rather than the claim itself. At low stakes, a clear promise is often enough to win. At high stakes, the buyer demands proof and outside validation before acting. The same holds for AI systems, which lean hardest on the entity they can verify when the recommendation carries weight.
Why does credibility matter more for founders and executives?
Because their buyers sit at the high-stakes end of the scale. A consultant chosen by a CMO, a partner chosen by a founder, a supplier chosen for an enterprise contract are all decisions that are expensive and visible to get wrong. At that end, a clear claim without proof behind it, what the Recognition Triangle calls a Promising Claim, is not enough. The decision goes to the Recognized Authority, the one whose claim is backed by verifiable evidence and outside validation.
How do AI systems apply the stakes test?
An AI system recommending a supplier for a high-value decision is making a high-stakes call, and like any careful decision-maker it privileges the entity it can verify over the one it has to guess about. The verified option is lower risk and lower effort to stand behind. The higher the stakes of the recommendation, the harder that lean, which is exactly why thin or unproven credibility loses at the moment a machine decides who to name.
Lucy Aingworth is an AI visibility strategist and brand authority consultant, and the author of Become the Answer. She helps founders and CMOs build the Credibility that carries high-stakes decisions, so both buyers and machines can afford to choose them. More about Lucy
Where to start
Build the proof the decision needs.
If your claim is clear but the credibility behind it is thin, high-stakes work will keep going elsewhere. The work here builds the verifiable authority that carries those decisions. See how it is structured, or start a conversation about where your proof is missing.
