Every review of the Semrush AI Visibility Toolkit on page one is published by a company selling a competing tracker. Trakkr, Profound, Mint, EchoWi, Dageno, HoneyB. The verdict was written before the testing was.
We sell no tracker. We run the measurement and the work behind it for clients, so the only question we care about is whether the number on the dashboard can carry the decision you are about to make with it.
Every one of those reviews lands on the same complaint: 25 prompts is not enough data. That complaint is wrong, and getting it wrong is why none of them tell you what the cap actually costs you.
The prompt cap, in the unit that matters
Semrush sells the toolkit in prompts. Coverage is decided in topics, and the entry tier buys exactly one
Convergence research reports thresholds per topic, not per program. At 25 to 40 prompts per topic, a prompt allowance converts into a topic count, and that count is what your reporting can actually cover.
25 prompts = 1 topic
not 25 topics
Split the entry cap across five product lines and you have five prompts each, which covers none of them.
AI Visibility Base · $99/mo per domain
25 prompts/day
Base + Guru or Business SEO plan · $99 plus the SEO plan
50 prompts/day
Semrush One Pro+ · $299/mo
100 prompts/day
Semrush One Advanced · $549/mo
200 prompts/day
Depth is not the constraint. Breadth is.
Are 25 prompts enough for the Semrush AI Visibility Toolkit?
For sample depth, yes. Tracking runs daily, so 25 prompts collect roughly 750 answers per engine per month, well past the 33 to 94 answers where published research says rankings stabilize. For coverage, no. Those 750 answers come from 25 distinct questions, which is one topic's worth of buyer intent.
The cap is a breadth limit dressed as a depth limit. That distinction changes which tier you buy.
Depth and breadth are different budgets, and only one of them is capped
Prompt allowances get compared like storage plans. More is better, less is worse. That framing hides the fact that the number is doing two jobs at once, and only one of them is under pressure.
The daily refresh already solves the sample-size problem most reviews raise
Semrush's published plan page lists the Base tier at $99 a month per domain, billed annually, with 25 custom prompts on daily AI rankings and data updates on daily, weekly and monthly cycles.
Run 25 prompts every day for a month and you have 750 answers per engine. Compare that to a pure-play tracker at 30 answers per prompt per engine per month and the entry tier looks generous rather than thin.
Convergence is measured per topic, and 25 prompts is one topic
Ronald Sielinski's convergence framework for AI visibility measurement, published July 11, 2026, tested 30 platform-topic combinations across Gemini, SearchGPT and Perplexity. Rank stability fired between 33 and 94 collected answers per topic-engine pair. Three combinations never converged at all.
The unit in that sentence is the topic. A program tracking one topic to convergence has done real measurement. A program spreading the same allowance across five product lines has five prompts per line and has measured nothing about any of them.
Twenty-five prompts is not a small sample. It is a complete sample of a small thing.
Splitting the cap is the decision nobody documents
Nothing in the product stops you from putting five prompts each on five topics. Nothing in the reporting tells you that you did something statistically fatal when you did.
The dashboard renders both setups identically. One is a measurement and one is a set of anecdotes with a chart on top.
What a prompt count tells you:
- •How many questions you can enter
- •What the next tier costs
- •Whether you hit the cap this month
What a topic count tells you:
- •How many buying conversations you can report on
- •Which product lines are outside the instrument
- •Whether last quarter's number covered the thing that changed
Every published review answers the first list. The second list is the one that decides whether the subscription works.
5 things the toolkit cannot see, no matter how many prompts you buy
None of these are Semrush defects. Each one follows from measuring a generative system on a fixed schedule and reporting the output as a brand's AI visibility.
Blind spot #1: Which reasoning mode produced each answer
This is the sharpest one, because Semrush published the evidence itself. On June 30, 2026 its research team ran 100 prompts twice through GPT-5.2, once in minimal reasoning and once in high reasoning, across 20 buyer journeys in four categories.
Only 25.6% of cited domains overlapped between the two modes on identical prompts. Citation rate rose from 50% to 68%, average citations went from 2.6 to 4.5, Reddit's share fell from 15% to 7%, and government and academic sources went from 1.9% to 8.8%.
Three quarters of the source pool turns over on a variable the daily tracker does not record. A month of readings labeled ChatGPT is a blend of two systems in unknown proportion.
Blind spot #2: Whether a daily arrow corresponds to an event
Our concluded CITE Index study ran 500 unaided buyer prompts through ChatGPT, Gemini and Google AI Mode every night for 63 days between May 19 and July 21, 2026, collecting 90,132 answers across 10 consumer categories.
The category leader flipped on only 18.7% of day pairs. In four of the ten categories the leader never changed once across the entire nine weeks. The full corpus sits in the final report.
A daily refresh produces roughly five times more movement than there are events to explain. The tool supplies the arrows and expects you to supply the threshold, and almost nobody does.
Blind spot #3: The engines it does not track
Semrush's pricing page names mentions from ChatGPT, Google AI, Gemini and Perplexity. Claude, Microsoft Copilot, Grok and DeepSeek are outside the product.
For most consumer brands that is a reasonable trade. For anyone selling to developers, researchers, or enterprise buyers working inside Microsoft 365, the engine embedded in the buyer's working day is the one missing from the report.
Blind spot #4: Whether your SEO data still predicts your AI data
The strongest argument for buying Semrush over a pure-play is that AI visibility sits next to the rankings, backlinks and site audit you already run. That is a genuine workflow advantage and the reason we recommend it to some teams.
It is also the assumption most worth testing rather than inheriting. The two measurements disagree more often than the shared dashboard implies, which we worked through in why Google rankings no longer predict AI citations.
Blind spot #5: What the AI Visibility Score is benchmarked against
Semrush's knowledge base defines the score as how often your brand is mentioned in AI answers compared with the median number of mentions for your top industry competitors, with those competitors identified automatically.
Two things move that score: your mentions, and the automatically selected comparison set. A score that falls because the tool swapped in a more visible competitor is not a visibility decline, and the number alone will not tell you which happened.
An index against an auto-selected peer group is two measurements reported as one.
Find out how many topics your current prompt set actually covers
We audit your tracked prompts against the buying conversations that produce revenue, size the allowance you need per topic, and report what your current tier can honestly prove. First findings inside 14 days.
Book a Discovery CallWhat the Semrush AI Visibility Toolkit actually costs
The $99 headline is accurate and incomplete. Two multipliers sit behind it, and both are priced per unit rather than per plan.
The domain is the unit that scales, and it scales at full price
The Base tier covers one domain. A second domain is another $99 a month, and additional users start at $45 a month, according to Semrush's own pricing page. There is no toolkit trial.
For a single-brand B2B company that is fine. For an agency, a multi-region business, or anyone running separate domains for product and docs, the per-domain model is the line item that decides the answer.
Prompts scale two ways, and the cheap route runs through a subscription you may not want
You can raise the prompt allowance by upgrading the Semrush SEO plan underneath it or by buying prompts directly. Third-party breakdowns report the tracking limit rising to 50 prompts per LLM on Guru and Business plans, an add-on of 50 more prompts at $60 a month, and Semrush One bundles at 50, 100 and 200 daily prompts.
Published figures differ between reviews and change often. Treat the table below as the shape of the decision rather than a quote, and confirm the numbers on a call before they enter a spreadsheet.
| Route to more prompts | Reported cost | Daily prompts | Topics it covers | What you are really buying |
|---|---|---|---|---|
| AI Visibility Base | $99/mo per domain | 25 | 1 | A complete read on one buying conversation |
| Base on a Guru or Business SEO plan | $99 plus the SEO plan | 50 | 1 to 2 | Prompts bundled with tools you may already pay for |
| 50-prompt add-on | $60/mo | +50 | +1 to 2 | The cheapest prompts per dollar in the lineup |
| Semrush One Pro+ | $299/mo | 100 | 2 to 4 | A focused B2B category, reported honestly |
| Semrush One Advanced | $549/mo | 200 | 5 to 8 | Multi-topic reporting plus API access |
| Second domain, any tier | +$99/mo | Separate allowance | Starts again at 1 | The multiplier that catches agencies |
At $60 for 50 prompts, the add-on is better value per topic than any tier upgrade. If the SEO tools in a Semrush One bundle are not already in your stack, buy the add-on and skip the bundle.
Price the tool per topic covered. Everyone else prices it per prompt, which is how a $99 plan turns into a $400 one three months in.
Sizing a Semrush prompt budget
The diagnostic half is done. Here is the sequence we run with clients before they commit to a tier, Semrush included.
Step 1: Name the buying conversations before you count prompts
List the distinct decisions your buyers make where an AI answer could name you. Category selection, build versus buy, vendor shortlist, integration fit, pricing sanity check. Each one is a topic.
Most B2B companies find three to five. That number, times 25 to 40, is your real prompt requirement, and it is usually larger than the tier they were about to buy.
Step 2: Rank the topics and fund them one at a time
You will not fund all five at once, and you should not. Pick the topic closest to revenue, spend the whole 25-prompt cap on it, and report that topic properly for a quarter.
One converged topic beats five unconverged ones. Our guide to selecting prompts for LLM tracking covers how to source and filter the list inside a topic.
Step 3: Establish the noise band before you set any threshold
Run the frozen set daily for four to six weeks with no content or off-page changes. Record the spread. That spread is your category's noise band, and our data says it is wider than the weekly arrows suggest.
Anything inside the band is not news. Setting the threshold after you see the number is explanation, not measurement. We worked the full arithmetic in how many prompts are enough.
Step 4: Log the cited domains, not only whether you appeared
Across our 90,132 answers, reddit.com drew 14,698 citations and appeared in 13.6% of all answers. Four of the twelve most-cited domains were brand-owned sites.
The cited-source panel is the off-page target map. A tracker used only for a mention rate is throwing away the half of the output you can act on this month.
Step 5: Check the score against a manual run once a quarter
Open all four tracked engines and run your ten highest-intent prompts by hand. Record where you appear and which competitors appear instead.
Two hours of manual work will tell you whether the automatically selected peer group in your score still matches the companies you actually lose deals to. No dashboard checks its own benchmark.
When the Semrush AI Visibility Toolkit is the right buy
We recommend it regularly. Fit is more useful to you than a verdict.
| Situation | Verdict | Why |
|---|---|---|
| One domain, one core buying conversation, Semrush already in the stack | Strong fit | Daily refresh on a single topic clears the convergence band, and the SEO data sits beside it at no extra integration cost. |
| SEO team taking on AI visibility without new headcount | Strong fit | The workflow is familiar, which is the difference between a tool used weekly and a tab nobody opens. |
| Three or more product lines to report on separately | Price the add-ons first | Each line needs its own 25 to 40 prompts. Budget from the topic count, not the headline tier. |
| Agency or multi-region brand | Check the domain math | Every domain restarts the allowance at full price, which is where per-domain pricing stops competing. |
| You sell to developers or enterprise Microsoft buyers | Poor fit alone | Claude, Copilot and Grok are outside the product, and for these audiences that is the surface that matters. |
| Nobody has been named as the weekly owner | Wrong purchase entirely | The measurement was never the bottleneck. The follow-through is. |
If you are still shortlisting, our AI visibility platform buyer's guide covers the six jobs any serious platform should do. For the pure-play comparison, what Profound AI actually measures works through sampling depth and what Peec AI actually tracks works through engine coverage. Semrush's constraint is neither. It is topics.
FAQ
What is the Semrush AI Visibility Toolkit?
The Semrush AI Visibility Toolkit is an add-on to Semrush that tracks how often a brand is mentioned and cited in AI answers. It contains six reports: Visibility Overview, Competitor Research, Prompt Research, Brand Performance, Prompt Tracking, and AI Search Site Audit. Prompt Tracking runs your set daily across ChatGPT, Google AI Mode, Google AI Overviews, Gemini and Perplexity, with coverage in over 220 countries and territories.
How much does the Semrush AI Visibility Toolkit cost?
Semrush publishes the Base tier at $99 a month per domain billed annually, covering one domain, 25 daily tracked prompts, and 300 reports a day. Additional users start at $45 a month and each extra domain is another $99. Third-party breakdowns report a 50-prompt add-on at $60 a month and Semrush One bundles at $199, $299 and $549 with 50, 100 and 200 daily prompts. There is no toolkit trial, and published figures vary, so confirm on a call.
Semrush vs Profound: which should you buy?
They constrain different things. Semrush caps breadth at 25 prompts on the entry tier but refreshes daily, so its risk is a topic you cannot see. Profound sells more engines and a prompt-volume dataset built on real user queries, but its self-serve tiers deliver about 30 answers per prompt per engine per month, so its risk is thin depth. Buy Semrush if you run one core topic and already live in the platform. Buy Profound if engine breadth or real prompt-demand data decides your reporting.
Which AI engines does Semrush AI visibility track?
Semrush names ChatGPT, Google AI Mode, Google AI Overviews, Gemini and Perplexity across its pricing and knowledge base pages. Claude, Microsoft Copilot, Grok and DeepSeek are not covered. Check which assistants your buyers actually use before treating that list as complete, because engines differ sharply in what they cite: our study found Google AI Mode cited a source in 97.4% of answers, ChatGPT in 92.5%, and Gemini in 79.1%.
What are the best AI visibility tools?
The named field includes Semrush AI Visibility, Profound, Peec AI, Scrunch AI, Otterly, Evertune and Ahrefs Brand Radar. Compare them on answers per topic-engine pair rather than on engine count or headline price, because that ratio decides whether any number they report can be defended. Then compare on whether anyone on your team will open the dashboard weekly, which decides everything else.
The bottom line
Semrush built a competent instrument and priced it in the wrong unit. Prompts are what the plan page sells. Topics are what your reporting is made of, and the conversion rate between them is 25 to 40 to one.
Run the conversion before you pick a tier. Count the buying conversations you need to report on, multiply by 25, and compare that against the allowance. If the answer is one topic and you have five, the honest move is to fund one properly rather than five badly.
Then go do the work the dashboard points at. Nothing in the software writes the answer block, fixes the passage the model could not extract, or earns the third-party mention that puts you in the source pool. That gap is why a managed GEO agency exists next to the tools, and an AI visibility audit will show you where your gap sits before you sign for a year of anything.
Buy the right allowance, then fix what the tracker finds
Cite Solutions maps your buying conversations to a sized prompt set, establishes your category noise band, and runs the content and off-page work that moves cited-source share.
Book a Discovery CallContinue the brief
Otterly AI: Is It Watching Your Market?
Otterly AI monitors 50+ countries, wider than any tracker at its price. Country tracking covers four of seven engines, and each market costs you prompts.
What Does Peec AI Actually Track?
Peec AI includes three AI engines on every self-serve tier, out of six on offer. Here is what the three you drop would have told you, and what they cost.
What Is an AI Visibility Platform? (2026 Guide)
An AI visibility platform tracks whether ChatGPT, Gemini, and Perplexity cite your brand. Here is what one does, who needs it, and how to pick.
Framework
Learn the CITE framework behind our GEO and AEO work
See how Comprehend, Influence, Track, and Evolve turn AI visibility into an operating system.
Services
Explore our managed GEO services and AEO execution model
Audit, prompt discovery, content execution, and ongoing monitoring tied to AI search outcomes.
Audit
Start with an AI visibility audit before execution
Understand prompt coverage, recommendation gaps, source mix, and where competitors are winning.