How Long Does It Take to Go From Invisible in ChatGPT to the Brand It Recommends?

Honest answer: the first measurable movement usually shows up in weeks, and durable, consistent recommendation takes months. One of our customers, Bosten Shoes, went from unmentioned to the number-one recommended leather shoe brand in El Salvador and logged its first ChatGPT-attributed sale in under 30 days. That is the fast end, and it happened in a market where almost nobody was publishing content shaped like an answer.

A crowded B2B SaaS category with five funded rivals already writing comparison posts is a different fight. We are not going to hand you a confident month count for that, because we do not have a dataset of hundreds of timelines and neither does anyone else selling you one. What we can tell you is which variable actually moves your clock: not how hard you work, but how much the engines have already read about your competitors across sources you do not control. If you have domain authority and press mentions, plan in quarters. If you are starting from a thin site nobody has written about, plan for longer and start sooner.

Here is the full story behind the fastest case we can put a name to, and what the sequence looked like.

What did "invisible" actually look like for Bosten Shoes?

Rodrigo, CEO and Founder of Bosten Shoes, did the thing every founder eventually does. He opened ChatGPT and typed the question his buyers type: the best leather shoe brand in El Salvador. It answered confidently. It named competitors. It did not name him once.

That is the first half of the wound. The second half is quieter and costs more. When a customer does arrive from an AI assistant, the visit tends to land in analytics looking like direct traffic with no clean source. So you are invisible where buyers are asking, and blind to the traffic you do earn. You cannot tell your board whether AI search is a channel or a rumor.

The stakes are not theoretical. AI search referrals grew 527 percent year over year, 71 percent of B2B buyers now use AI chatbots, and that traffic converts at 4.4 times the rate of ordinary search traffic. A small number of sessions you cannot see is worth more than a large number you can.

What does the first month actually involve?

This is the sequence Bosten ran. It is the same sequence every account runs, which is the point: nothing here was bespoke heroics.

Week one: measure, do not guess. Setup is three steps and takes about five minutes. Connect the site, connect analytics, set up the agents. Clark scans the site, runs a short interview to learn the business and the voice, and monitoring makes its first pass across ChatGPT, Claude, Gemini, and Perplexity. What comes back is not a score. We do not invent those. It is a list: the exact buyer questions, which companies got named in each answer, and where your brand was absent. That baseline is the first real deliverable, and it exists on day one, before a single post is published.

Weeks one to three: publish into the gaps, not into the dark. Every question where the engines named someone else becomes a priority topic. Clark writes answer-first posts against those exact questions, in the brand's voice, grounded in the company's own material, and publishes them straight to the connected CMS. The structural choice that matters most is the opening paragraph: it answers the question completely and makes sense lifted out of context. Retrieval layers quote clean passages. They do not stitch together a post that spends 200 words warming up.

Weeks three to four: the answer moved, and the money followed. Monitoring re-asks each tracked question daily on a rotation, so a change is visible as it happens rather than discovered a quarter later. Bosten went from unmentioned to the top recommendation. Then the part most tools cannot show you: a sale arrived from ChatGPT and landed attributed to revenue, with the engine and the post named, instead of disappearing into direct traffic. Under 30 days from switch-on. We have seen the same under-30-day pattern in funded B2B SaaS, though a shoe brand in a quiet market and a SaaS company in a loud one are not the same race.

Why was this case fast, and why might yours be slower?

Check each of these against your own situation.

• Low published competition. In that category and market, almost nobody was writing content shaped like an answer. Engines had thin material to pull from, so new, well-structured pages had room.
• Questions came from evidence, not a keyword tool. The posts targeted measured gaps, the real questions returning competitor names, so nothing was spent on topics buyers never ask.
• Attribution was wired on day one. The sale was not faster because of attribution. It was provable because of it. Plenty of brands earn an AI-driven sale in month one and never know it happened.

What slows you down is an established competitor set the engines have read about repeatedly across review sites, roundups, and trade coverage. Models lean on brands they have encountered across independent sources, and that corroboration is not something you publish your way out of in two weeks. It is a reason to start now, not a reason to wait.

What is the reusable takeaway?

Stop measuring AEO in "did we rank." Measure it in dated events you can point at: the day you first appear in an answer, the day you hold the top recommendation on a tracked question, and the day a dollar traces back to a specific post and a specific engine. Bosten hit all three inside a month. Your first one is likely weeks away. The third depends on your sales cycle as much as on the engines, which is why we would rather you watch the dates than a score.

If you want your own baseline, the free visibility check takes about a minute and shows which questions name your competitors and not you. That is one snapshot. Clark runs it every day, and ties what changes back to revenue.