How I outsmart AI search at the moment of decision.

What is Decision Engine Optimization (DEO)?

Decision Engine Optimization (DEO) is the practice of optimising your online reputation and authority signals so that AI models recommend your business at the final decision-making stage, when a potential customer uses an AI tool like ChatGPT, Claude, Gemini or Perplexity to compare options and choose who to buy from. Where Answer Engine Optimization (AEO) makes you the answer to a question, and Generative Engine Optimization (GEO) makes you visible across generative results, DEO makes you the chosen option when the AI is asked to compare and decide.

Think of the customer journey in three acts. In the first, the buyer is discovering options, this is where AEO and GEO do their work, getting your name into the shortlist. In the second, the buyer is researching and comparing. In the third act, the eleventh hour, the buyer has two or three quotes in hand and asks an AI a deceptively simple question: which of these is the best value? That third act is the zero moment of truth, and DEO is what wins it.

AEO and GEO get you into the room. DEO gets you the signature.

The eleventh hour: how buyers now make final decisions

The behaviour is already widespread and growing fast. A buyer collects two or three quotes, then instead of agonising alone or asking a friend, they paste all of them into an LLM and ask it to judge. The prompt is rarely about price alone. It sounds like: "I have three quotes for a kitchen extension, which company is the best choice and why?" or "Compare these two proposals and tell me which is more trustworthy."

At that instant, the AI stops being a search engine and becomes a decision engine. It weighs the options, and it renders a verdict in natural language, complete with reasons. If your business is the one it recommends, you have effectively closed the sale before the competitor even knew the comparison was happening. If you are not, you lose, silently, and you will never see it in your analytics.

This is why DEO matters so much. The decision is being made in a place you cannot see, by a system you do not control, using signals you may not have optimised. The good news is that those signals are knowable, and they are the same ones a careful human buyer would weigh, only now they are read by a machine.

How AI determines the best value

When a customer inputs multiple quotes, the AI does not simply pick the lowest number. It evaluates each business against a range of reputation and authority signals, then compares those signals against the submitted quotes to identify the most reliable choice, not merely the cheapest. Four factors dominate.

The lesson is blunt: price is one input among many, and often not the deciding one. A business with strong, machine-readable proof of quality can and does win the comparison against a cheaper rival with a thin reputation. DEO is the discipline of making sure those proofs exist, are public, and are readable by the models.

The leaking bucket problem

James Dooley describes the danger perfectly: ignoring DEO creates a leaking bucket. You pour money into lead generation, ads, SEO, AEO, GEO, and you fill the bucket with prospects. But if your online reputation is not optimised for the final decision-making phase, those prospects leak out at the very last step, when the AI comparison quietly recommends someone else. You paid to fill the bucket, and the hole at the bottom is your unoptimised reputation.

Most businesses spend almost all of their budget on the top of the funnel, getting attention, and almost nothing on the bottom, winning the decision. DEO rebalances that. It is often the cheapest, highest-leverage work available, because the prospects are already there. You are not buying new attention, you are converting attention you already paid for.

How to win AI comparisons: the DEO action plan

Winning the decision engine comes down to making your reputation signals abundant, public and machine-readable. Here is the practical checklist.

AEO vs. GEO vs. DEO

The three work together. AEO and GEO fill the bucket. DEO seals the hole. A brand that masters all three is discovered, surfaced, and chosen, and that is the full journey from question to signature.

Why DEO is emerging right now

Decision Engine Optimization is not a rebrand of old reputation management. It is a response to a genuinely new behaviour that did not exist three years ago: ordinary buyers using general-purpose AI as an impartial adviser at the point of purchase. Two things converged to make this happen. First, LLMs became good enough at reasoning that people trust them to weigh trade-offs, not just fetch facts. Second, pasting a few quotes into a chat window became easier and faster than calling a friend, reading a dozen reviews, or building a spreadsheet.

The result is that a huge amount of decision-making has quietly moved into a private conversation between a buyer and a model. In the old world, you could influence the final decision with a persuasive salesperson, a well-timed follow-up call, or a discount. In the new world, the deciding conversation often happens without you in the room at all, and the only version of your business present is the one the AI can read from the open web. If that version is thin, generic, or unverifiable, you lose the argument you never got to make.

This is why DEO belongs alongside AEO and GEO as a distinct discipline rather than a footnote. Discovery and decision are different problems with different signals. You can be brilliantly discoverable and still lose every comparison, and that is the most expensive failure of all, because you paid to be in the running and then lost at the finish line.

A worked example: three quotes, one AI verdict

Imagine a homeowner planning a loft conversion. They gather three quotes: Company A at the lowest price, Company B in the middle, and Company C, the most expensive. In the old world, price alone would tempt them toward Company A. Instead, they paste all three into ChatGPT and ask: "Which of these companies should I choose for a loft conversion, and why?"

The model does not have the builders' private files. It has only what it can retrieve about each business from the open web. Here is what typically decides the verdict:

Company A has the lowest quote but a bare website, no case studies, a handful of reviews scattered on their own page, and no visible accreditations. To the AI, it is cheap but unproven, a risk.

Company B has a decent site and some reviews, but nothing that stands out. The AI treats it as a safe, unremarkable middle option.

Company C, the most expensive, has dedicated project pages with photos, dozens of independent reviews on third-party platforms, clearly stated accreditations, "trading since 2011," and mentions on external directories. The AI reads all of this and concludes that Company C offers the best value despite the higher price, because the evidence of reliability reduces the buyer's risk.

The most expensive quote won, because it had done the DEO work the others ignored.

This is not a hypothetical edge case. It is increasingly the default path a considered purchase takes. The business that invests in machine-readable proof of quality changes the very basis of the comparison, from "who is cheapest" to "who is safest to trust," and that is a comparison a strong reputation wins every time.

Where DEO matters most

DEO delivers the biggest returns in considered purchases, the ones where a buyer collects multiple quotes and the decision carries real risk or cost. Construction and trades, where a homeowner compares builders, are a textbook case, which is why accreditations like CHAS and years-trading signals matter so much there. Professional and financial services, where FCA and equivalent registrations establish legitimacy, are another. Medical and healthcare providers, where trust is paramount and buyers compare clinics, live and die by public reputation signals. Engineering firms, agencies, and any high-value B2B service where proposals are compared side by side all sit squarely in DEO territory.

The common thread is choice under uncertainty. Wherever a buyer faces several credible options and cannot easily judge quality themselves, they now outsource that judgement to an AI, and the AI rewards the business that has made its quality legible. Low-consideration, impulse, or pure-commodity purchases benefit less, because there the comparison really is mostly about price and availability. But for everything that involves trust, DEO is decisive.

Common DEO mistakes to avoid

The failures are consistent, and most are quietly self-inflicted. The most common is keeping proof private: awards, accreditations and impressive projects that live in a PDF, a brochure, or an image the model cannot read, rather than as visible on-page text. If an LLM cannot read it, it does not exist for the comparison. The second is relying only on reviews hosted on your own website, which the AI heavily discounts as self-reported; the trust comes from independent, third-party platforms. The third is vagueness, claims like "market-leading" or "trusted by many" with no verifiable specifics, when the AI is looking for concrete numbers, dates, and named credentials.

And, as always, the temptation to fake it. Manufactured reviews, invented accreditations, or fabricated case studies are not just unethical, they are increasingly detectable and they poison the very trust signal you are trying to build. The whole point of DEO is that it rewards businesses that genuinely deserve to win. The work is to make real quality visible, not to manufacture the appearance of it.

How to measure DEO

You cannot see the private conversations where AI recommends or rejects you, but you can test and infer. The simplest method is to run the comparison yourself: paste your own quote alongside realistic competitor quotes into ChatGPT, Claude, Gemini and Perplexity, and ask which is the best value and why. Read the reasoning. It will tell you exactly which signals the AI found, and which it wished it could find. Repeat this monthly and watch the verdict shift as you strengthen your reputation footprint.

Beyond that, watch the downstream effects: rising conversion rates on quoted work, more prospects arriving already convinced, and buyers referencing details about your business you never told them, a sign the AI surfaced your proof during their decision. DEO success shows up not as a vanity metric but as a bucket that finally stops leaking.

The bottom line

Search marketing has always chased the moment of attention. Decision Engine Optimization chases the moment of choice, and in an era where buyers hand that choice to an AI, it may be the most valuable moment of all. AEO and GEO earn you the shortlist. DEO wins you the decision. Ignore it and you fund a leaking bucket; master it and you convert the attention you already paid for into signed work, even against cheaper rivals.

The businesses that win the next decade will not simply be discoverable by AI. They will be the ones AI actively recommends when it is asked to decide. That is the work. Make your quality real, make it public, make it readable, and become the option the decision engine chooses.

Frequently asked questions