AI Visibility Consultant: What You Get for $1,500 vs $99
A consultant sells three things. Only one of them is measurement, and measurement is the part software already does for a tenth of the price.
Deciding what to measure: which prompts your buyers actually type. This is judgement, and it is where consultants earn their fee.
Measuring it: tracking mentions and citations across assistants. Profound from $99/mo, Searchable from $125/mo, Cognizo from $499/mo.
Fixing it: earning mentions on the third-party sites assistants draw from. WISLR $3,500/mo, Orange Marketing $6,950/mo.
The gap between a tool and a human here is roughly 7x to 70x. Both are real prices. They are not the same purchase.
Most pages about hiring an AI visibility consultant skip the only question that matters at the start: whether you need a person at all. The honest answer depends on which of the three jobs above is actually your bottleneck, and for a lot of companies it is not the one they think.
This page separates the three, prices each from what providers publish on their own sites, and says plainly where a consultant is worth the money.
Disclosure: MentionAgent is our own product. It is outreach software, not a consultancy, and it does the third job on that list rather than the first two. We publish our own measurement of how often assistants mention brands, which is the number this whole category is built on.
Across 4,598 assistant answers covering 1,088 distinct queries, we measured a brand mention rate of 4.6% overall, and it was near-identical across engines: ChatGPT 5.0%, Gemini 4.6%, Claude 4.7%.
Citation rate diverged sharply where mention rate did not: Gemini cited a source in 98.9% of answers and Claude in 98.6%, but ChatGPT in only 73.6%. A quarter of ChatGPT answers cite nothing at all.
Why this matters before you hire anyone: a 4.6% baseline is normal, not a crisis, and any consultant showing you a low starting number is showing you the industry norm.
Caveat we keep attached: these scans ran on the cheaper model tiers, not flagship models. Full method in the AI citation study.
There is no shortlist to target. We pulled every source those answers cited, refreshed on 26 August 2026: 45,567 citations across 12,880 distinct domains, from 4,250 answers carrying at least one source.
The AI citation study analyses an earlier, smaller cut of the same corpus.
The distribution is extraordinarily flat. The ten most-cited domains account for just 8.5% of all citations. The top hundred account for 23.3%.
7,554 domains, or 58.6% of every domain we saw cited, were cited exactly once.
Why this matters when hiring: there is no list of twenty sites to get on. A consultant promising to place you on "the sites AI trusts" is describing something our data says does not exist as a short list, which is exactly why choosing your prompts is the job that carries the value.
The three jobs, and which one is actually your problem
Job one is deciding what to measure. Nobody can track every possible prompt, so someone has to choose the fifty or so questions that represent real buying intent in your category. Get this wrong and everything downstream is noise. This is genuinely hard, it is specific to your business, and it is the single most defensible reason to pay a human.
Job two is measurement. Running those prompts across assistants on a schedule, recording who gets mentioned and cited, and charting the change. This is software. It is well served by software. Paying consulting rates for it is paying a person to read you a dashboard.
Job three is fixing it. This is where most of the money should go and where most of the confusion sits. Assistants answer from sources they already trust, so becoming citable means getting mentioned on those third-party sites. That is earned media and content work pointed at a new target. Our GEO guide covers the mechanism in full.
If your bottleneck is job two, buy a tool. If it is job one or three, a tool will hand you a chart and change nothing.
What the market actually charges
Unusually for a young service category, a lot of providers here publish real numbers. We checked each of these on the provider's own site in August 2026.
| Provider | What it is | Published price |
|---|---|---|
| Profound | Platform | Starter $99/mo, Growth $399/mo, Enterprise custom |
| Searchable | Platform | $125 to $400/mo, 14-day trial |
| Cognizo | Platform, plus done-for-you | Platform $499/mo, Done-for-You $899/mo, Enterprise custom |
| WISLR | Consultancy, Atlanta | Audit $1,500 one-time, consulting $3,500/mo |
| Orange Marketing | Agency, Newport Beach | Audit $6,950, implementation $6,950/mo on a 3-month programme |
| Visibly AI | Platform, EU | None stated on the page |
Prices are what each provider published on its own site on 26 August 2026 and will move. "None stated" means no number was visible, not that it is cheap.
Read the shape rather than the individual numbers. Software clusters at $99 to $499 a month. Human service clusters at $3,500 to $6,950 a month. The order-of-magnitude gap is the labour, and the question you are really answering is whether you have someone internally to do that labour.
How to judge a consultant before you sign
1. Ask them to name your prompts before they quote. A consultant who cannot describe the twenty questions your buyers ask, in your language, has not done the one job you cannot do with software. This is the fastest filter available and it costs you one email.
2. Get the baseline captured first, in writing. The mention rate on a fixed prompt list, dated, before any work starts. Without it, month-three reporting has nothing to compare against and every number is flattering by default.
3. Ask what proportion of the retainer is earned media. If the diagnosis is that you lack third-party mentions and the plan is all on-site schema and FAQ blocks, the plan does not match the diagnosis. Ask directly how many placements a month the fee buys.
4. Treat citation guarantees as you would press guarantees. Nobody places anything inside a model. Influence is real, control is not, and a promise of specific citations is a promise nobody can keep.
5. Check whether you are buying the dashboard twice. Several consultancies resell a platform you could license directly at $99 to $499. That can be perfectly fair if the fee covers interpretation and action, but you should know which part of the invoice is the licence.
When a consultant is clearly worth it
You are in a category where the prompt set is not obvious. Regulated products, technical B2B, or anything where buyers describe the problem in language that does not match your product name. Choosing what to track is most of the work here.
You have nobody who will act on the findings. A report that lands in an inbox and produces no outreach, no content and no placements is a pure cost. If there is no internal owner, buying the doing as well as the measuring is rational.
You need it defensible internally. A named outside expert with a documented method carries weight in a budget conversation that a dashboard screenshot does not. That is a real reason, and it is worth saying out loud rather than pretending the analysis is otherwise unobtainable.
When a consultant is the wrong purchase
Your prompt list is obvious and short. If ten questions cover your category and you can write them in five minutes, you have done job one yourself. Buy the tool.
You have not measured anything yet. Get the baseline before you buy strategy, because the baseline sometimes says the problem is smaller than assumed. You can start with our free AI mention checker and our list of prompts to test AI brand visibility.
The diagnosis is going to be "you need more mentions". For most companies it will be, because that is the mechanism. If you already know that, the money is better spent on getting the mentions than on being told to get them.
If the answer is more mentions
The uncomfortable part of this category is that the fix is rarely exotic. Assistants cite sources they already trust, so the work is getting written about on the sites in your niche that those systems draw from. That is outreach, and it is unglamorous.
MentionAgent is $99/mo flat and does that job. The agent finds relevant posts in your niche, looks up the author's verified email, writes a pitch tied to the specific article, and sends it once you approve it in Telegram. Follow-ups run until someone replies. No credits, no per-contact fees, no minimum term.
What it does not do is job one. It will not decide which prompts matter in your category, and it will not tell you your strategy. If that is your gap, a consultant at $3,500 a month may be exactly right and we would rather say so than pretend otherwise.
How to pick
Decision tree
- Never measured your mention rate? → do that first, free, before spending anything.
- Know your prompts, need the data, have someone to act? → a platform at $99 to $499/mo.
- Do not know which prompts matter in your category? → a consultant, and judge them on that question alone.
- Know the answer is "more third-party mentions" and have nobody to get them? → done-for-you, or MentionAgent at $99/mo.
- Need a named outside expert for an internal budget case? → a consultant, and say that is the reason.
- Being offered guaranteed citations? → walk, or reprice it as advertising.
The fix is mentions. That part is $99/mo.
MentionAgent finds the blogs assistants already cite in your niche, looks up the contacts, writes the pitches, sends them and follows up. Flat rate, whatever the volume. No credits, no minimum term.
Start Getting Mentioned On AutopilotFrequently asked questions
What does an AI visibility consultant actually do?
Three things: decide which prompts to measure, measure them, and fix what the measurement shows. The middle one is software. The first and third are what you are paying a person for, and the third is usually earned media rather than anything on your own site.
How much does an AI visibility consultant cost?
WISLR publishes $1,500 for a one-time audit and $3,500 per month for consulting. Orange Marketing publishes $6,950 for an audit and $6,950 per month on a three-month programme. Platforms sit far below: Profound from $99/mo, Searchable from $125/mo, Cognizo from $499/mo. Both are real prices for different products.
Do I need a consultant or is a tool enough?
A tool is enough if you know which prompts matter, have someone to act on findings, and just cannot see the data. A consultant earns the fee when you do not know what to measure, or when the fix needs third-party coverage and nobody internally will go and get it. A dashboard nobody acts on is the common waste here.
Can anyone guarantee I will be cited by ChatGPT?
No. Nobody places anything inside a model. Assistants draw on third-party sources, so citation is earned by being mentioned where those systems already look. Influenceable, not controllable. Treat a citation guarantee like a guarantee of specific press coverage.
What should I measure before hiring anyone?
Your mention rate on a fixed prompt list, captured and dated before work starts. Our own measurement across 4,598 answers put the overall brand mention rate at 4.6%, so a low starting number is the norm. Without a baseline you cannot separate improvement from noise.