Your next shortlist is being drafted by an AI assistant

Ask ChatGPT to name vendors for a specific water treatment problem and it will answer with confidence. It will list four or five companies, describe what each does, and sound authoritative doing it. Whether your company appears in that answer is already affecting your pipeline, because the person asking is often a utility engineer or a consultant doing early research for a project you'll never hear about otherwise.

This is the case for generative engine optimization, or GEO. It overlaps with SEO, but the signals that get you cited by an AI assistant differ from the ones that rank you on Google, and most water companies have invested in neither.

How buyers use these tools

The early research phase of a water purchase has always been invisible to vendors. An engineer scoping a project reads, asks colleagues, and builds a mental shortlist months before any RFP. What has changed is where that reading happens. AI assistants compress a day of research into a ten-minute conversation, and the answers they give carry the tone of a neutral expert.

The practical effect is that the shortlist forms earlier and with fewer touchpoints. If the assistant names your competitors and omits you, you're losing deals you never knew existed.

The questions are more specific than most vendors expect. An operations manager asks something like "continuous chlorine analyzers that handle high-manganese groundwater" or "vendors for acoustic leak detection on cast iron mains." A consulting engineer building an alternatives analysis asks for "established suppliers of MABR systems in North America" and gets a tidy list with descriptions, which is uncomfortably close to the format of the memo they were about to write. When a name from that answer lands in the alternatives table of a preliminary design report, it tends to survive into the RFP.

Nobody involved thinks of this as vendor discovery. It reads as research, it happens months before procurement, and no analytics tool you own will show it happening. The first observable trace is a branded search or a direct visit, long after the shortlist formed.

What AI answers reward

Three things show up again and again in the companies these tools cite.

First, entity clarity. The model needs to know your company exists as a distinct thing: what you make, who you serve, how you relate to the rest of the industry. Structured data, a consistent description across your site, Wikidata and directory presence, and third-party mentions all feed this.

The test is whether a model can write one accurate sentence about you. Pick that sentence yourself, "X builds Y for Z," and use it everywhere: site footer, About page, LinkedIn company page, directory listings, conference bios. Add Organization schema to the site, claim or create the Wikidata and Crunchbase entries, and make sure the About page states plain facts (founded, headquarters, what you make, who buys it) instead of mission language. Models can't cite a vibe.

Second, content that answers questions directly. AI assistants synthesize from pages that state problems and answers plainly. A page titled with the question a buyer asks, answering it in the first paragraph, gets cited. A brochure page about your "innovative solutions" doesn't.

In practice this means pages built around the questions themselves. "How much does SCADA integration cost for a small utility" as the heading, a direct answer in the first hundred words, and the nuance below it. FAQ blocks help because they're pre-extracted question-and-answer pairs. Comparison content helps because assistants lean on ranked, structured source material. The brochure section of your site can stay; it just won't be what gets you cited.

Third, independent corroboration. Models weight claims that appear in more than one place. Trade press coverage, conference talks, podcast appearances, and industry newsletters all function as citations. This is where years of executive thought leadership quietly compound.

Mentions count even without links, which makes this industry's small trade press unusually valuable. A WaterWorld feature, a WEF technical session writeup, two podcast appearances, and a mention in a well-read newsletter give a model four independent places that agree you exist and agree on what you do. That agreement is what turns you from a website into an entity worth naming.

How this differs from SEO

Classic SEO competes for position on a results page; GEO competes for inclusion in a synthesized answer. That changes the work. Rankings reward domain authority and links accumulated over years. Citations reward clear, extractable statements and corroboration, which is why a small company with sharp, specific content can get named alongside incumbents that outspend it a hundred to one.

There's also no paid lane. Google sells the top of its results page, and most water companies waste money there anyway, but nobody can buy their way into ChatGPT's answer about membrane suppliers. The answer runs on evidence, and for once that favors companies whose main asset is expertise.

Where to start

Run the test yourself. Ask ChatGPT, Perplexity, and Google's AI mode the questions your buyers would ask, phrased the way an operator or an engineer would phrase them. Note who gets named. Do it monthly.

Score it simply: for each question, you're either named, named and described accurately, or absent. Track the tally in a spreadsheet each month. The number moves slowly, which is normal; treat it like any other leading indicator over a long sales cycle and judge it in quarters.

Then work the gap. Fix the entity basics on your own site first, since they are fully in your control. Build content around the questions where you should be the answer. Earn the third-party mentions. The compounding is slow, which is exactly why the water companies that start now will be hard to displace from those answers in two years.

On a calendar, that's entity fixes in month one (days of work, fully in your control), the first two question pages in month two, and the earned-mention push (a trade-press pitch, a podcast, a conference abstract) from month three onward. Almost none of your competitors are doing this deliberately yet.

---

Adam Tank is the founder of Full Flow Marketing, a marketing agency built for the water industry. Search and AI visibility is one of the three services we run. Ask for a free pipeline diagnostic and we'll show you who the assistants name in your category today.