I have said the same thing to clients for 25 years. "Every click on a hyperlink constitutes someone looking for the answer to a question. So our only real job then is to know who those people are, to know what they are asking, and to answer them better and with less effort than anyone else."
For most of those 25 years, that was a hard sell. It sounds obvious when you say it out loud (I would often even acknowledge that in advance, with: "This is a bit of a head-slapper, but..."). People would nod, agree completely, and then go right back to arguing about the color of a button or which image to use. The idea was easy to accept and easy to ignore, because nothing forced anyone to act on it. A site could be vague and self-absorbed and still limp along on brand searches and referrals.
That grace period is over. There is an acronym for the idea now, AEO, for Answer Engine Optimization, and the whole industry is suddenly discovering that the web was always about answering questions.
What changed is not the truth. What changed is that something now sits between your buyer and your website, and it reads. When a buyer asks an AI a question about your category, the model does not admire your design or count your backlinks. It looks for whoever answered that question most clearly and repeats them by name. If your site never plainly answered the question, there is nothing to repeat, and someone else gets named. For the first time in 25 years, being vague has an immediate, measurable cost.
We have the receipts on the other approach, because we were doing this before it had a name.
Years ago we took on Eldorado Natural Spring Water. Before we designed a single page, we mapped what their customers were actually trying to do and organized everything around those questions instead of around the company's org chart. Then we changed the site every month based on what real people did on it. Revenues and conversion rates climbed, and they started breaking their own sales records... every year. Nobody called it AEO. It was the obvious way to build something that works.
We did the same thing at a much larger scale for a city government website. When we took it on, it had 16,000 pages and 252 things to click on the homepage. A resident looking for one answer had almost no chance of finding it. So we did the slow, unglamorous work first. We identified who actually used the site and the specific questions they arrived with, then mapped all 16,000 pages against those questions to find what was answered, what was buried four clicks deep, and what was missing entirely. A great deal was missing. In fact, 40% of the questions people came to the site to ask were not answered anywhere on it. Not buried. Absent. We rebuilt the whole thing around the questions, not the org chart.
That is the entire discipline, and it has not changed. Find out who your audience is. Learn the real questions they ask, in their words and not yours. Score how well your site answers each one. Build what is missing, and make all of it easy to reach. It was the job in 1990, it was the job when we rebuilt that city's site, and it is the job now that an AI is grading the same work in real time.
The tools have finally caught up to the principle. That is good news, because it makes the conversation easier to have. When a prospect can open ChatGPT, ask their own buying question, and watch a competitor get named instead of them, they understand in about ten seconds what we have spent 25 years trying to explain.
So if you want to know where you stand, run that test yourself. Ask an AI the question your best buyer asks right before they choose a company like yours, and see whose names come back. If you are not in the answer, that is not a mystery and it is not permanent. It means the questions have not been mapped and the answers have not been written clearly enough to be quoted. That is fixable, and it is exactly the work we have been doing since before the web had a name for it.