Realistic AI visibility goals by sector: what to expect at 3, 6 and 12 months
"So when will I show up in ChatGPT?" is the question that decides proposals, and the temptation to answer it with a figure and a date is enormous. Resist it: in AI visibility —GEO (Generative Engine Optimization): measuring and improving how a brand appears in the answers of AI search engines like ChatGPT, Gemini, Perplexity or Claude— promises with a date are the shortest path to the cancellation conversation.
What you can —and should— give the client is a framework: what usually happens at 3, at 6 and at 12 months, and why their specific sector will move faster or slower. Because the most important fact in this whole article is this: the starting point is nothing alike between sectors, and a goal that ignores the starting point isn't a goal, it's a wish.
The fact: between sectors there's a 10-fold difference
In our study of 9,865 Spanish SMEs across 30 sectors in 10 cities (full study) two figures came out that should be in every one of your proposals:
- 91% of SMEs appear in only 1 of the 4 AIs. Starting from almost zero is the norm, not the exception.
- But "almost zero" varies enormously: 1 in 4 hotels appears across several AIs, against 1 in 50 garages. Same country, same AIs, an order-of-magnitude difference.
Why? Because AIs compose their answers from sources, and the density of sources is radically different by sector. A hotel has hundreds of public signals: booking platforms, thousands of reviews, comparisons, guides, travel press. A car garage, with luck, has a half-filled Google listing. The AI doesn't discriminate by sector: it simply can't recommend something it knows nothing about.
From that come three starting situations that change the goal:
Source-dense sectors (hospitality, restaurants, leisure, travel): almost everyone "exists" for the AIs. The fight isn't appearing, it's position, description and accuracy — and it's a fight against many.
Intermediate sectors (professional services, clinics, training, specialised retail): partial sources, uneven competition. There are real gaps, especially in the specific and the local.
Desert sectors (garages, trades, industrial, niche B2B): almost nobody appears. It's the worst starting picture and the best strategic news: whoever builds sources first finds a field with no one in it. Here the first moves show sooner, precisely because there's no one to lose against.
What to expect at 3 months: foundations and first signals
Work for the period: measured baseline, false facts corrected in the sources, listings and directories up to date, first reviews and content under way.
What's reasonable at 90 days: visible improvements in the brand prompts (what the AI says when asked directly about the client: better description, correct facts) and, for local businesses, first movements in some local prompt. The generic discovery prompts —"best X in Y"— will barely move, and that's normal in any sector. The classic mistake of quarter one is judging the service by the prompts that take longest to move; when defining the list it helps to be clear about what each type of prompt measures.
In desert sectors, there are sometimes early surprises: with so few sources in play, a well-built listing and a few dozen reviews can be enough to enter some local answer. Don't promise it; celebrate it if it arrives.
What to expect at 6 months: the first appearances that matter
Work for the period: content that answers the target prompts, third-party mentions, authority under construction.
What's reasonable at the half-year mark: appearances in some of the problem and long-tail prompts ("who fixes oil boilers in the Madrid sierra?" before "best heating company in Madrid"), presence in more than one AI for the prompts where you were already showing, and a picture against competitors that starts to move. In dense sectors, the progress will show more in how the client is described and alongside whom they appear than in the raw count.
Six months is also the honest threshold: if there's absolutely no signal in any AI, it's not time for another normal report — it's time for a diagnosis, and that's what what to do when a client isn't improving is for.
What to expect at 12 months: consolidation and defence
What's reasonable at the year mark: stable presence in a significant share of the target prompts, several AIs, correct facts, and a position against competitors you can show in a chart with pride. In desert sectors, a year of serious work can mean being the default answer for their niche and their area. In dense sectors, it means having earned a recurring place in the conversation — which is already being where the 37.9% of the Spanish population who used generative AI in the last quarter of 2025 (INE) can find them.
And a new goal appears that's worth announcing from the start: defending. AI answers are volatile and the position won has to be watched and sustained. The year-two service isn't the year-one service, and that's not a failure of the plan: it's the nature of the channel.
How the goal is set with a specific client
- Measure their real starting point before promising anything — an initial audit, not an intuition.
- Compare it with their sector, not with "businesses" in general: a hotel's goal and a garage's goal can't look alike.
- Agree phased goals with reviews at 3 and 6 months, committing to verifiable work and not results — the difference between what's committed and what's only reported is developed in what to promise (and what not) about AI visibility.
- Write it down. The timeframe agreed in the proposal is what turns month four into a review instead of a crisis.
Frequently asked questions
In SEO I give traffic estimates. Why not here?
Because there's no equivalent (yet) to search volume per keyword, and because AI answers vary between sessions and weeks. You can estimate direction and order of magnitude ("in your sector, with your starting point, what's reasonable at 6 months is X kind of progress"); you can't estimate "you'll appear in 14 prompts in March". Whoever promises that has either never measured, or is counting on no one checking.
Does a desert sector guarantee quick results?
No: it guarantees little competition, which isn't the same. If the sector is a source desert, you have to build them almost from scratch (reviews, mentions, content), and that takes months of work even with no one across the table. The advantage of the desert is that each piece built pays off more, not that the work is less.
Do these timeframes apply to a brand-new brand?
Add time to them. A brand with no established website, no reviews and no history starts behind the 91%: the first half-year is pure construction. To win as an agency, a better client is one that already has substance (clients, possible reviews, a story to tell) and poor visibility — there the work shows sooner.
What do I show the client while the big results aren't here yet?
Work done and intermediate signals, every month, with data: sources corrected, content published, evolution per prompt and per AI even if modest. A good monthly report sustains trust during the construction phase; a bad one burns it even when things are going well.
To set goals by sector you need the starting point measured, not assumed. With Surfeo for agencies you have each client's baseline across the 4 AIs and their weekly evolution against that baseline; and if you want a client's (or a prospect's) starting point right now, run it through the free AI visibility test.
