'My competitor shows up in AI and I don't': how to respond without losing the client
This time the client isn't asking: they're accusing. They forward you a ChatGPT screenshot recommending their direct competitor on the exact question that should be bringing them clients, with a short message along the lines of "And this?". And underneath, unwritten, the real question: "What am I paying you for?".
It's the salt-in-the-wound version of the question about whether they appear in ChatGPT. That one was curiosity; this is a direct threat to the retainer. The good news: if you handle it with method instead of excuses, this is usually the moment the client signs new work, not the moment they leave.
First: lower the temperature without denying the problem
Classic mistake number one: playing it down ("hardly anyone uses that AI stuff yet"). False and dangerous: frequent use of ChatGPT in Spain has gone from 4% to 28% in two years (Funcas, III Survey on AI, 2026). The client knows it, because the screenshot was sent to them by a real person.
Classic mistake number two: promising you'll fix it in two weeks. No one controls what an AI answers, and the client will end up using that promise against you.
The response that works has three parts: validate, contextualise, propose.
Key data
"You're right, and I'm glad you're sending it to me: this is exactly what we need to watch from now on. Before drawing conclusions from a single screenshot, let me check it properly: AI answers change depending on how you ask, from where, and on which day. I'll prepare a full diagnosis for you this week: which questions your competitor shows up in, which ones you show up in, and above all, why."
With that you've bought a week and turned a reproach into an assignment. Now comes the important part: the real diagnosis.
The diagnosis: why AI recommends their competitor
An AI doesn't recommend on a whim, nor for advertising. It recommends what it can read and verify in its sources: websites, reviews, directories, media, forums. When a competitor shows up and your client doesn't, the cause is almost always one of these four:
1. Citable content. The competitor has pages that answer concrete questions ("how much does X cost", "what does Y include", comparisons, guides for their sector) and your client has a pretty home page that asserts nothing citable. AIs cite whoever gives them sentences with data, prices and answers; not whoever has the best slogan.
2. Reviews and directory presence. AIs lean heavily on third-party signals: Google Business Profile, reviews with volume and responses, sector directories, mentions in forums where people ask for recommendations. If the competitor has 240 reviews answered and your client has 18 unanswered, there's half your explanation.
3. Structured data and technical clarity. If the client's website doesn't make clear, in a way a machine can read, what they do, where, for whom and at what price, the AI won't take the risk: it cites whoever does make it clear.
4. Mentions off their own site. Local press, sector associations, specialist blogs, podcasts. The AI triangulates: if three independent sources say the competitor is a benchmark, that carries more weight than anything the client says about themselves.
Check all four and you'll have the answer to "why them and not me" with concrete causes, not algorithmic mysticism. If you want the full checklist of what to test and what to note down, it's in the first AI visibility audit, step by step.
The context the client needs to hear
Before presenting the plan, frame the real size of the problem, because the screenshot has made them believe their competitor "dominates AI". Almost certainly they don't: in the study we ran on 9,865 Spanish SMEs across 30 sectors and 10 cities, 91% appear in only 1 of the 4 main AIs (full data here). Among hotels, only 1 in 4 is visible across several AIs; among garages, 1 in 50.
Translation for the meeting: the competitor most likely shows up in ChatGPT and is invisible in Gemini, Perplexity and Claude. There's no wall to scale; there's a small, recent advantage that can be neutralised. And it's worth doing: traffic arriving from AI answers converts at 14.2% versus 2.8% for classic organic (sector data compiled by roymo.es). Whoever asks an AI arrives with the decision almost made — which is why it stings so much when the answer is someone else's name.
The 90-day plan
Don't promise appearances; promise verifiable work with dates. This structure defends itself in a proposal:
Days 1-30: baseline and sources. Define the 20-40 prompts that matter in their sector (the questions their potential customer asks, in natural language). Measure where they show up and where the competitor shows up across the 4 AIs, with dated screenshots. Fix the urgent stuff: out-of-date data, an incomplete Google profile, a website with no structured data, basic information the AI can't read.
Days 31-60: build what's missing. Citable content that answers the detected questions (prices, comparisons, guides). A reviews campaign with a system, not with pleading. Listings and clean-up in the directories the AIs use as a source in their sector.
Days 61-90: measure and show. Repeat the full measurement, compare against the baseline, document every change in the answers. This is where the dated screenshots from day 1 are worth gold: without a baseline there's no story to tell.
Be honest about timeframes in the proposal: in 90 days there's almost always measurable movement, but you don't sign which week the AI will change its answer, because nobody knows that. You sign the work, the measurement and the report.
From crisis to system
There's still the operational problem: measuring 30 prompts across 4 AIs for the client and their competitor, every week, with screenshots, is hours of manual work no one will pay you for at consultant rates. For that we use Surfeo for agencies: each client in their own workspace, their prompts queried every week across the 4 AIs, the main competitor watched in parallel, and a PDF report for the monthly reporting. The first snapshot — the one you need this week to respond to that screenshot — comes free from the AI visibility test.
And a closing piece of advice: next time, make sure you're the one who sends the screenshot to the client before their brother-in-law does. That's the difference between selling monitoring and asking forgiveness.
Frequently asked questions
What if I run the diagnosis and the competitor really does show up in every AI?
Then you have the best possible commercial argument: the proof that in their sector this is no longer theory. The plan is the same, with an urgency that explains itself. Cases of strong competitors in all 4 AIs are rare (remember the 91% from the study), but they exist, and they're caught the same way: sources, reviews and citable content.
Do I tell the client exactly why their competitor shows up?
Yes, with the concrete causes you've verified: "they have 200 more reviews than you", "they have a page that answers that exact question", "they're in these three directories and you're not". The more concrete the cause, the more credible the plan. What you mustn't do is invent causes you haven't checked.
Could the competitor be paying to show up in ChatGPT?
No. As of today you can't pay to appear in the organic answers of ChatGPT, Gemini, Perplexity or Claude. If they show up, it's because their sources back them. That's good news for you: what's earned with work can be contested with work.
How much do I charge for this 90-day plan?
It depends on the client type, but the market already gives clear references: you've got the ranges and the margin maths in how much to charge for AI visibility services.
Before the meeting, turn up with your own screenshot: run the free visibility test on your client's site and their competitor's, and turn the "and this?" into a signed proposal.
