Your agency says the SEO is “doing great.” Rankings are up on the report. Traffic is down, leads from organic are down, and when you ask an AI assistant about your category, it recommends three companies, and none of them is you.
That's the new shape of search. Buyers ask a question and get an answer, often without clicking anything. The technicians are still optimizing for the list of blue links. I'm a digital futurist who started with words, and I'll say it plainly: answer engine optimization is mostly a writing job. The machines quote sentences. You have to write the sentence worth quoting.
Why answer engine optimization matters: what's happening to the click
The numbers on this keep moving, and the providers dispute some of them. Here's what the better studies say:
- Zero-click is the majority. One clickstream study found roughly 68% of US searches on the largest search engine ended without a click to any website in the first four months of 2026 [1], up from about 60% in 2024.
- AI summaries take clicks directly. A randomized field experiment, where some people saw AI summaries at the top of results and others didn't [2], found those summaries cut outbound clicks to websites by about 38% on the searches where they appeared. Zero-click searches in that group rose from about 54% to 72%.
- AI assistants are their own front door. A growing share of buyers now start with an assistant and never see a results page at all.
So the old deal (rank high, earn the click, convert on the page) still exists, but it's shrinking. The new deal is: be the source the answer leans on, and be named when it does.
What the technician sees, and what the strategist sees
The technician sees a technical problem with a technical checklist. Add structured data, fix crawl access, publish an FAQ block, track mentions in AI answers. Every item is worth doing.
The strategist asks what the machine will actually say about you, and whether you've written anything it can say. Most company websites are full of sentences no answer engine would ever quote: “We're a leading provider of innovative solutions tailored to your unique needs.” That sentence is invisible to a machine looking for an answer, because it doesn't answer anything.
Compare a sentence written for an imaginary agency: “A 12-week program for multi-location dental groups that cuts the cost of each new patient, with call tracking set up in week one.” Specific, checkable and quotable. A machine can lift that into an answer about dental marketing. It can't lift a slogan.
The technician makes you readable. The strategist makes you worth reading aloud.
The Marine Corps trained me as a Combat Correspondent, and the first rule of news writing is the inverted pyramid: put the most important fact first, because editors cut from the bottom. Answer engines are editors that cut from the bottom, at the speed of light, on every page you've ever published.
The machine quotesyour first clearsentence. Make sureyou wrote one.
The mechanism: Citation Gravity
Citation Gravity is the pull a sentence has on an answer engine. Four forces create it:
- Directness. The sentence answers a question someone actually asks, in the first line of the section, without warm-up.
- Specificity. It names who it's for, what it does and a concrete detail: a number, a place, a timeframe, a method.
- First-hand weight. It contains something only you could know: your experience, your data, your process. Generic advice is available everywhere. First-hand detail isn't.
- Agreement. Other places on the web (reviews, directories, articles, your own profiles) describe you the same way. Machines trust what they see repeated consistently.
Notice that three of the four forces are about words. Only one, agreement, depends on distribution, and even that is a words problem: if your profiles, your site and your reviews all describe you differently, the machine can't tell what you are.
Writing pages with Citation Gravity is part of the prompt and page work we do inside the iFix Lab.
The math (illustrative): losing the click, winning the name
These are modeled numbers for illustration, drawn from no real account.
Before. A company gets 200,000 organic visits a month from non-branded searches. It converts 1.5% to leads, so 3,000 leads, and closes 10%: 300 customers a month from organic.
The squeeze. Say AI summaries and assistants cut non-branded clicks by 35%. Visits fall to 130,000. At the same rates, leads fall to 1,950 and customers to 195. That's 105 customers a month gone, and the ranking report hasn't changed.
The words fix. The company rewrites its top 40 pages so each section opens with a direct, specific, first-hand answer, and it aligns how it's described across its profiles. Say it now gets named in AI answers often enough that branded searches rise by 20,000 visits a month. Branded visitors already know the name and convert at, say, 4%: 800 more leads. At a 15% close rate on those warmer leads, that's 120 customers.
The net. More customers than before the squeeze, from fewer clicks. The clicks that remained were worth more because the answer had already done some of the selling.
Your categories and conversion rates will differ. Model your own exposure before you trust mine.
More customers than before the squeeze, from fewer clicks.
What to write first
If you're starting from a site full of slogans, here's the order I'd follow:
- Your category answer. One paragraph that says exactly what you are, who it's for and what makes it different, in plain words. Put it on the homepage and the About page.
- Your top ten buyer questions. The real ones, from sales calls and support tickets. Each gets its own section with the answer in the first sentence.
- Your first-hand proof. Process details, specific outcomes you can document, how you actually do the work. This is the information gain that generic pages can't copy.
- Your consistent description. The same core sentence on every profile and directory you control. Agreement is how machines decide what you are.
This is also where a lot of buyers' frustration with “supposedly doing SEO” comes from. They paid for rankings and got a report. What they needed was pages that say something.
The pages machines skip
Most sites are full of pages answer engines pass over. Here's what they tend to have in common:
- Throat-clearing intros. Three paragraphs about the industry before the page says anything. The machine looking for an answer doesn't wait.
- Slogans in place of descriptions. “Your partner in growth” tells a machine nothing about who you serve or what you do.
- Gated substance. The real answer sits behind a form, so the public page is all promise and no proof.
- Hedged everything. “Results may vary depending on many factors” with no specifics. Machines quote sources that commit to something they can back.
- Contradictions across pages. One page says you serve small businesses, another says enterprise, a third says everyone.
None of these is a technical failure. Every one is a writing choice. That's why so many owners feel their agency is supposedly doing SEO while nothing changes: the technical work got done, and the words never got rewritten. Fix the writing on your top twenty pages and you'll often learn more about your AI visibility in a month than a year of ranking reports taught you.
Where the common advice breaks between $1M and $10M a month
At $1M a month, the advice is to “add FAQ schema and write for AI.” For a small site with a narrow offer, a few well-written answer pages can get you named quickly.
Between $3M and $10M, the site has hundreds or thousands of pages written by different teams over many years, and they contradict each other. Product pages say one thing, the blog says another, the sales deck a third. Machines see that disagreement and hedge, which often means naming someone else whose story is cleaner.
A machine can'trecommend a brandthat describesitself fivedifferent ways.
What changes once you're already past $10M a month
- One canonical description of the company. Approved, short, specific, used everywhere. It's a brand asset and a search asset now.
- Content gets audited for answers, not keywords. Each important page should answer a buyer question in its first lines. If it doesn't, rewrite or retire it.
- First-hand content gets protected. Your proprietary data, field experience and process detail are what machines can't find elsewhere. Publish more of it, carefully.
- Track what the machines say. Ask the major assistants your top buyer questions monthly and log whether you're named, how you're described and who's named instead.
- SEO and brand sit together. At this scale, the people writing the words and the people tuning the pages have to share a plan.
The decision framework I'd use
Five checks on any page you want quoted:
- The question test. What exact buyer question does this page answer?
- The first-line test. Is the answer in the first sentence of the section?
- The specificity test. Does it contain a detail a competitor couldn't paste onto their own page?
- The agreement test. Does it match how you're described everywhere else?
- The read-aloud test. Would an assistant sound smart quoting it?
Questions senior buyers ask
Is SEO dead?
No. Rankings and clean technical setup still matter, and plenty of searches still end in clicks. What's changed is that being quoted in the answer is now a separate prize, and it's won mostly with words.
Can we just add structured data and be done?
Structured data helps machines understand your pages. It doesn't give them anything worth saying. Do both.
How do we know if AI answers mention us?
Ask them. Run your top buyer questions through the main assistants each month and record the results. It's crude and it works.
Should we write content for AI or for people?
For people, in a way machines can parse. Direct answers, specific details and clear structure serve both.
Will this traffic convert?
Visitors who arrive after an AI answer often convert better, because the answer pre-sold them. Expect fewer clicks worth more each.
How long does it take?
Rewriting your most important pages can take weeks. Getting consistently named can take months, because machines need to see the same description repeated across sources.
Before your next SEO report
Ask an AI assistant the question your best buyer would ask about your category. Read the answer. If you're not in it, read your own homepage and find the sentence it should have quoted.
If you can't find one, write it today. Working with the machines instead of against them is what the iFix Lab is built for. The LLM Ops Vault holds 35 prompts sorted by job, including ten built for LLM operations, so you can audit what the assistants say about you and rewrite your pages before the next quarter's traffic report. Membership is $97 a month, cancel any time, and month one comes with a 14-day refund. Join the iFix Lab.
Sources
- Search Engine Land, Google zero-click searches reach 68% in early 2026: Study, 2026.
- Search Engine Journal, Study Confirms Google AI Overviews Cut Organic Clicks 38%, 2026.