For the first time in about fifteen years, a firm outside tech can take the leading search position in its category without outspending the incumbent. That sentence would have been irresponsible in 2019. It is simply accurate now. AI assistants are becoming the front door for evaluations in legal, healthcare services, industrials, financial services, and professional services broadly, and the answer layer they draw on carries almost none of the incumbency that made traditional search a trench war. The board reset, and most of your competitors have not noticed. That is the opportunity, and it has a clock on it, so we will be specific about what taking it requires.
A decade of trench warfare
Recall what the last decade of search looked like in most non-tech categories. The firm with the oldest domain, the deepest backlink pile, and the largest paid budget held the high ground, and everyone else paid agencies to shell it. Rankings crawled. Cost per click ratcheted upward annually like rent. Challengers who tried to buy their way in learned that an auction rewards whoever can lose money longest, which is usually the incumbent. So marketing settled into maintenance: defend the brand terms, publish the blog, renew the retainer, hold position.
Incumbency won because those ranking systems were built on accumulation: age, links, spend, history. If you were behind in 2015, you were probably still behind in 2024, and not for lack of effort or fees. Plenty of good firms spent a decade and a great deal of money confirming this.
Why your category is behind, and why that is good news
Non-tech categories are structurally behind on modern GTM, and the reason is uncomfortable: most agencies serving them sell activity, not instrumentation. Deliverables, not baselines. Eight posts a month, a ranking report with green arrows, a social calendar. Ask what changed in the business and you get a traffic chart. The retainer model rewards motion, and motion is what gets sold. None of this makes the people involved bad at their jobs. It means the category's marketing muscle never got conditioned, because nothing ever demanded conditioning.
Tech companies grew a different muscle, and not because their marketers are smarter. Their boards demanded instrumentation. Raise venture money and someone eventually asks which experiments worked, what payback looks like, and whether the metric means the same thing it meant last quarter. That pressure produced a discipline, and the discipline is portable. You can hire tech-grade GTM the way you would hire tech-grade finance, and in a category where nobody else has it, it compounds fast.
The low sophistication of your competitive set is exactly why the window is so open. In software, a dozen vendors per category are already contesting AI visibility. In legal, healthcare services, and industrials, the answer layer is mostly unclaimed ground. Models hunting for credible sources in your category are finding thin material. The first firm to hand them dense, structured, verifiable evidence tends to get cited, and citations compound the way links once did, only faster and with far less deference to domain age.
What tech-grade means, concretely
Four things, none of them exotic.
Versioned measurement. A baseline taken before the work begins, metrics with written definitions, a change log when a definition moves. If your agency cannot state your citation share last quarter under the same definition it uses today, you have decoration, not measurement. In the answer layer this means tracking the questions buyers actually ask, which sources the assistants cite, and where you sit in that set, on a schedule, against a peer group you named in advance.
Pre-registered experiments. Decide in writing what success looks like before the campaign runs. This one habit eliminates most of the self-congratulation that passes for reporting, which is why almost nobody volunteers for it. It is the center of an evidence-first operating style: claims are cheap, predictions are expensive, and only expensive things earn the trust of a skeptical partner group.
Entity discipline. Machines have to be able to work out what your firm is. One canonical name, consistent facts on every profile and directory, structured data on the site, a set of claims about practices, people, and credentials that never contradicts itself anywhere on the web. Tedious, unglamorous, and the most common gap we find in non-tech firms, because no one has ever owned it.
Citation-ready proof. Assistants cite sources that make specific, attributable, verifiable statements. "A full-service firm with a client-first approach" is invisible to a model, and to everyone else. Named methodologies, published thinking, specific expertise attached to specific people, ungated. Most established firms are sitting on twenty years of proof locked in PDFs, pitch decks, and partners' heads. The work is excavation and publication, not invention.
Underneath all four sits a cadence: experiments reviewed on a fixed rhythm, assumptions repriced quarterly, one senior owner who can be embarrassed if the numbers are wrong. Discipline is mostly a calendar with consequences.
A worked example from legal
A national legal firm, anonymized here, ran exactly this play. Nothing exotic: entity cleanup, an attorneys page rebuilt as a primary asset rather than a directory, proof published in citable form, measurement versioned from a baseline and reviewed on a schedule.
Their attorneys page became the most-cited page in its space out of 1,900+ tracked pages. Sentiment toward the firm in AI answers runs at roughly 90 percent positive, the best in their peer set. Over the same period, new client acquisition rose 72 percent, their strongest six-month stretch on record. We are careful with causal language, and legal buyers have many inputs. But when the most-cited page in your space is yours, and the machines describing you are getting the story right, you would expect the phones to behave differently. Theirs did.
Notice what was not required: a decade of link building, a doubled media budget, a viral anything. The answer layer rewarded structure, proof, and consistency, which are things a disciplined firm can produce in quarters rather than decades. That is the whole argument in one example. The game your category spent ten years losing has been replaced by a game nobody in your category is playing yet.
The window is open because it is early, not because it is easy
Let us be plain about the physics. This advantage exists because the answer layer is young and your category is quiet. Both conditions expire. Models keep retraining. Competitors eventually notice their absence from answers, usually when a client mentions it at exactly the wrong moment. And the early mover's citations become part of the corpus that the next round of answers draws on, which is why ground taken first is the easiest ground to hold. In tech the window is already crowded. In your category it is likely still open, and a firm that moves this year can occupy territory it failed to take in ten years of trench warfare. The same effort in 2028 will buy much less.
How to start
Start with a baseline, not a campaign. What do the assistants say about your category today, who gets cited, where are you, and what do the answers get wrong about you. Fix the entity layer next, because everything else builds on machines knowing what you are. Then publish proof, ungated and specific, beginning with the questions clients actually ask in the first year of an engagement. Measure quarterly against unchanged definitions. Put one senior name on the program, inside the firm, with a standing slot in the partner meeting. And defer the tool shopping until the fundamentals hold; software is where undisciplined programs go to hide.
The firms that win this will not be the ones that worked hardest at the old game. They will be the ones that noticed the new one first, and brought professional discipline to a category that has never seen it.
If you are curious what the assistants say about your firm right now, the free self-audit takes a few minutes and tends to be clarifying.