Most of the Market Still Hasn't Tried AI, So Don't Write for the Bubble

Most of the market has not tried what you tried last Tuesday. If your bottom-of-funnel pages are written as if they have, you are talking to a corner of the market and calling it the buyer.

An operator I interviewed put the scale of that corner in plain terms. People who have tried it in a chat window, then tried it inside an IDE, and then carry something like a Claude Code subscription are in a vanishingly small corner. It is easy to forget that, because you live in bubbles, and even more so in a place like Austin or Silicon Valley. Not everyone is deep into this.

That is a market fact. It is also a content fact. The page that wins a buying search, and the page an answer engine will repeat, is the one that meets the reader who is not in the bubble.

The bubble writes pages the buyer cannot use

Inside the bubble, the interesting argument is which tool, which workflow, which new release. The homepage starts there. The comparison page starts there. The "how it works" page starts there. Everyone in the draft meeting already has opinions about the tools, so the page sounds normal to the people who approved it.

It does not sound normal to the buyer. Their week is a backlog, a compliance worry, a report nobody trusts, a team that tried a chat box once and went back to the old process. They are not choosing between coding assistants. They are deciding whether this category is even for them.

When the page assumes they are already deep in it, three things happen.

They do not recognize the situation, so they leave. The search they typed, which was about the job, does not match a page about the tools, so you do not rank for it. An answer engine, asked the question in their words, repeats whoever described the job. Often that is not the company with the better product. It is the company that did not write for the bubble.

Write the page for the person who is not deep in this

Start from the reader who has not lived your last six months. Maybe they opened a chat window. Maybe a colleague pasted an answer into a doc. Maybe they have not done even that, because the last time they asked, someone told them the data could not leave the building. That reader is the market. The corner that already has a coding-assistant subscription is not.

Say the job before you say the tool. What is stuck. Who feels it. What they tried. What they are afraid will go wrong. Then say what you do about that job, including the step that is not magic, and where you are the wrong fit.

This is the same discipline as naming the problem in the buyer's words. If the only words on the page are the ones your team uses with each other, you have not named it yet. Buyers who do not know your category exists are already patching the problem some other way. Meet them in that search, not in a recap of what is new inside the bubble.

The searches that matter here are still a short list: alternatives, versus, best-for, switching, integrations, and the pricing questions they actually ask. That list is the BoFu Map. A term like "best AI coding stack for founders in Austin" is a bubble query. A term like "how a clinic keeps records questions in-house" is a buying query. Put the second one on the URL. Let the first one go.

Specifics are what still get trusted when everyone assumes a model wrote the draft. A specific from outside the bubble is a workflow, a constraint, a person who has to explain the purchase upstairs. A specific from inside the bubble is a product name the reader does not use. One of those earns the next paragraph. The other earns a bounce.

What changes in the draft

You do not need a new strategy. You need a harder edit on the pages that sit next to a decision.

• First screen. The situation, in their words. No release names. No assumption they have tried an IDE.
• The change. What is different on a Tuesday after this is in place, including what they still do by hand.
• The limit. Where this is the wrong fit. A buyer who is not deep in the tools needs that more, not less, because they cannot tell the boundary themselves.
• One sentence an outsider can repeat. If it only makes sense to someone with a coding-assistant subscription, it will not get cited and a champion cannot forward it.

Read that sentence to someone who does the job and does not read AI news. If they ask what the words mean, the page is still for the bubble. Replace the words. Do not add a glossary underneath a headline they already skipped.

The operators who remember the bubble are not dumbing the work down. They are refusing to confuse their own peer group with the market. The peer group already knows them. The market is who the page is for.

What to do this week

Open your three pages closest to a decision: a comparison, an alternatives page, and one best-for page. Highlight every sentence that assumes the reader already uses a chat tool daily, lives in an IDE, or knows what a coding assistant subscription is.

If the first screen is mostly highlight, rewrite the opening as the job they are stuck on. Keep one concrete limit. Cut the tool tour, or move it below the answer.

Then ask one customer who is not in your city and not in your group chat to read the first screen and say, in one sentence, who the page is for. If they describe your peer group instead of their own team, it is still a bubble page.

Common mistakes

• Writing for the room you had lunch in. Austin and Silicon Valley are real markets. They are not the market.
• Treating "has tried a chat window" as the advanced segment. For a lot of buyers, that is the whole experience. Meet them there.
• Hiding the constraint. The reader who is not deep in this needs the limit on the page. Insider confidence reads as a skip.

See which buyer searches you're writing past

The searches that close are the ones typed by people who are not in the bubble. The 15-minute traffic check shows five of those searches, where you're missing, and whether the AIs are naming you.

See the traffic you're missing. BoFu, Baby builds bottom-of-funnel content engines for B2B companies: 15 money pages and 15 supporting posts in 90 days, built to rank on Google and get you named by ChatGPT. See what you're missing.