What are the most common complaints about AI visibility tools? In Alium Research's verified buyer interviews, six issues recur: visibility figures whose method buyers cannot see, coverage limited to the prompts and engines a tool runs, reporting that stops short of telling the team what to do, products still being built, SEO-suite AI modules that trail the dedicated tools, and results nobody can yet tie to revenue. Two concern the tool's data and method, three the product, and one the category, as buyers often say themselves. Many of the buyers here are piloting or in their first year with a tool. No tool in this category observes every answer every engine gives every user: the figures draw on answers collected for chosen prompts, modelled estimates of what people ask, referral visits and, in some products, logs of AI agents visiting a site. Those answer different questions and do not add up to a complete record; buyers complain as much about how little of each method they can check as about the limits. Several of the replacement decisions here move AI-visibility work off an SEO suite, some after the buyer judged the suite's AI features behind and some to consolidate tools, while buyers here on dedicated tools describe re-evaluating at renewal. These are patterns in these interviews, not failure rates, tested performance, or evidence that the suites' modules are weaker everywhere or that any complaint caused a move. Cost has its own page, what buyers actually pay for AI-visibility tools, and the ratings live on what AI-visibility buyers wish they'd known.
The six complaints — graded
Each complaint is grouped by how buyers in these interviews respond: tolerating it, working around it, or reconsidering their contract. These are observed patterns, not numerical severity ratings. The same issue can be manageable for one company and serious enough to prompt a switch for another.
Figures whose method buyers cannot see
Why don't buyers trust AI visibility data? They are questioning two different measurements. A visibility score depends on which prompts the vendor runs, on which engines, how often it collects the answers and how it counts a mention or a citation. An estimated prompt volume comes from a different process: where the vendor gets evidence of what people ask, and how it projects that evidence onto everyone. Each can be wrong in its own way, and what buyers object to is how little of either method they can see. A smart-home products company reads its Scrunch data as directional at best, the third story below. A vitamin and supplement brand on Profound finds “it's hard to validate that information and it's hard to understand if what you're doing is impacting it.” A large retail group rates Profound an eight and still says “I don't really trust the methodology behind the data,” and a women's health brand has made both halves, how answers are found and how volumes are projected, its main test at renewal, the first story below. It reaches the services side too: a large financial-services company working with Bluefish says of a results readout “It felt somewhat surface level their findings,” adding “I'm always a little skeptical of results if there's not clear citations.” Buyers still shopping describe the same unease before they sign: a footwear retailer looking at the category calls prompt tracking “still kind of a black box,” and a global beauty company ruled Semrush out of its AI-visibility shortlist because it does not use clickstream data on what people actually type. Vendors do describe their methods. Profound says its prompt volumes come from licensed, double-opt-in consumer panels of answer-engine users, modelled up to the wider population, and that its enterprise plan tracks up to nine answer engines daily; Scrunch says its topic-level volume estimates, also built from a panel and modelling, are directional rather than exact. A described method is not a validated one, though. Buyers mostly live with it by reading the figures as trends rather than counts. Our recommendation, not a practice buyers describe: ask for the two methods separately. For visibility, ask for the tracked prompt list, the engines and countries, how often answers are collected, sample responses, and how the tool defines a mention, a citation and share of voice. For volume, ask what observed data sits behind the estimate, how it is projected and what uncertainty the vendor reports. Compare any AI referral figures with your own analytics as a separate check, once you know which visits and sources each one counts.
Named for this: Profound · Scrunch · Bluefish · Novi · Semrush
Coverage limited to the prompts and engines a tool runs
What can't AI visibility tools see? A prompt-monitoring report can only describe answers to the prompts it runs, on the engines, markets and languages it covers, in the periods it collects. Other features estimate wider demand or log AI agents' visits to a site, but none of them shows every answer users are given, and the complaints here are mostly about the monitoring side: coverage and breakdowns in the plans buyers run. The prompt-list limit, and buyers who walked away from choosing their own prompts, are covered on the lessons page; the complaints here are about the edges of what a tool reports. A global food company that rates Profound a seven names “Lack of some metrics and ability to compare across agents.” A large retail group's head of SEO says “All these tools are very prompt-based tracking, so I'm just really looking for that ultimate tool that does tracking beyond just prompts.” A consumer-hardware brand on Gumshoe gets a competitive share of mentions and wants it cut by intent: “Good to know. But it would be great for us to know that, hey, you're going up in branded or non-branded ways.” And a luxury fashion brand on Scrunch wants engines its plan does not cover and more freedom to add prompts, the second story below. Engine coverage varies by vendor and by plan, so a gap one buyer reports describes its plan and date, not everything the vendor sells. Profound's trial covers three engines and its enterprise plan up to nine, DeepSeek and Claude among them; Scrunch's site lists ChatGPT, Perplexity, Claude, Gemini and Copilot among the platforms it monitors, and it separately reports which pages AI bots visit through a connection to a site's content delivery network or host, which shows the bots' visits rather than the answers they feed. Buyers mostly work around the gap with a second tool, their own prompt sets or a build of their own; the global food company is building parts of its measurement bespoke. Our recommendation, not a practice buyers describe: list the engines, markets, languages and breakdowns (branded and unbranded, by intent, by product line) you need, and check each against the plan you would actually buy.
Named for this: Profound · Scrunch · Gumshoe · Yext
Reporting without direction
Do AI visibility tools tell you what to do? Several buyers say theirs stops at the report. The pattern itself is covered on the lessons page; what the complaints add is how it lands tool by tool. A watch company that rates Profound an eight says “it's really been more educational. The action items haven't been overly over the top.” An online services company rates Bluefish a five because “there are a bunch of other platforms moving at a faster pace,” and wants a vendor that can “guide, not just provide a solution, but potentially guide strategy as well.” Where a tool does recommend, buyers judge the recommendations: an insurance company piloting Adobe LLM Optimizer says “their tool provides recommendations, but even their own tool indicates that the thing that they recommend has no value,” and decided that was not worth implementing. Buyers still shopping now write it into the brief; a global consumer-goods company evaluating platforms says “We want clear, robust outputs that enable us to take action.” Vendors are moving toward it: Profound and Bluefish both sell agents or optimization suites that turn findings into tasks, and Bluefish offers a full-service option alongside its software. Buyers mostly work around the gap with an agency, a content partner or their own team. Our recommendation, not a practice buyers describe: during the trial, take three findings the tool surfaces and ask who on your side would act on each, how, and by when.
Named for this: Scrunch · Bluefish · Adobe LLM Optimizer · Profound · Semrush
A product still being built
Are AI visibility tools mature? Their users describe products that ship fast and are not finished. A skincare brand trialling Scrunch rates it a five so far: “My experience with it so far has been a little confusing in terms of onboarding, and it's been a bit buggy,” and “It definitely feels like it's still being built.” It keeps trialling because it wants the insight, and expects to keep evaluating. A hardware manufacturer rates Profound a seven because “it feels a little more complicated than it needs to be,” and “I would have liked something a little bit more intuitive.” A large technology company that also rates it a seven says “I don't think their agentic solutions are working as well as we'd like them to be yet,” and a consumer subscription company found connecting its content delivery network to Profound's platform through its secure infrastructure “has been a challenge.” Some of the limits are plan limits rather than defects: Profound's pricing page lists API access on its enterprise plan and not its trial, which a smaller customer on a lower tier found when it wanted its data inside its own AI assistant; and a luxury fashion brand finds the page allowance on Scrunch's page-optimization product, which serves a lightweight version of a site to AI agents, too small for its catalog. The content features add their own check: a global food company using Profound's content generation says “you have to ensure there are no hallucinations in the content.” Buyers mostly tolerate this as the cost of a young category, and several expect to re-evaluate as the tools change. Our recommendation, not a practice buyers describe: confirm in writing which features, engines, prompt counts, page limits and API access come with the plan you are buying, rather than with the demo.
Named for this: Profound · Scrunch · AirOps
SEO-suite AI modules that trail the specialists
Are SEO suites' AI-visibility features good enough to replace a dedicated tool? Some buyers who compared them say the modules in their plans did not meet their needs, while praising the suites for SEO. A life-sciences tools company rates Semrush an eight and calls it “a great platform, still very relevant,” but says “I don't see a lot of new innovations around the AI capabilities.” A heritage fashion brand shopping for a dedicated tool says “If SEMrush were strong enough, we probably wouldn't bother.” An outdoor apparel brand relies on its agency's AI-search dashboard, which is “kind of like filling the gap of what Semrush or one of those SEO tools can't provide right now.” The same complaint reaches other suites: a large retail group wants its enterprise SEO platform to “catch up on tools like Peak AI and Profound,” and a company that runs marketing for a network of local businesses says the Ahrefs product it uses “does not have any visibility into those AI tools,” and finds its Yext listings tool lacking granular detail on where a location ranks in AI answers. These are judgments about particular capabilities on particular dates, not proof that Semrush, Ahrefs, Yext or any other suite cannot now measure enough. The suites now sell AI-visibility features: Semrush's AI Visibility Toolkit tracks prompts on ChatGPT and Google's AI Mode and is sold on its own or bundled in Semrush One, Ahrefs' Brand Radar tracks how brands appear in AI Overviews, AI Mode, ChatGPT, Perplexity, Gemini and Copilot, and Yext sells Scout. Further examples, and how buyers weigh a suite's module against Profound, are on the lessons page and Profound vs. Semrush. Our recommendation, not a practice buyers describe: run the suite's AI module and a dedicated tool on the same prompt set for the same weeks before deciding you need both.
Named for this: Semrush · Ahrefs · Yext
Results nobody can tie to revenue
Do AI visibility tools prove ROI? Not yet, in these interviews. Buyers can watch mentions and citations move, and some can count the visits and conversions their analytics credit to AI referrals, but those are different levels of evidence: a citation does not guarantee a visit, a referred conversion is not necessarily one that would not have happened anyway, and some AI-driven visits arrive without the tracking that would credit them, the smart-home company's doubt in the third story below. Buyers mostly put the gap on the category rather than on one vendor. The vitamin and supplement brand above asks of its Profound numbers “is it moving the needle?” A consumer-hardware brand says “there's no clear playbook that we can work on to ensure a concrete way of getting a lot more mentions.” A pet brand chose Novi for a proof of concept partly on case studies that tied visibility gains to revenue, and says “They have not given insight into how they are actually attributing that revenue”; whether it continues depends on what its own analytics show at the end. A global hotel company holding off on dedicated vendors puts the doubt plainly: “I don't know how credible they are yet,” and “I just don't know if the industry is set up to be really, really ready to go to help brands figure this out.” An unproven return is not the same as no return: AI referral traffic is still a small share for several of these buyers, which makes any effect hard to read. A biopharmaceutical buyer building its own attribution around Profound is on the lessons page, and how buyers budget against an unproven return is on the pricing page. Our recommendation, not a practice buyers describe: before a pilot starts, agree what evidence would justify continuing it. Track visibility on a fixed prompt set, AI referral visits and conversions in your own analytics, and the work your team did in response to the tool's findings, as separate indicators rather than one proven chain. Where you can, compare against a baseline or a staged rollout, and write down what stays unmeasured; a rise in citations or referred conversions does not on its own establish new revenue.
Named for this: Profound · Novi · Gumshoe · Scrunch
One further complaint sits outside the chart. Price is as common as any complaint above among Profound users, who describe it as the premium option — a global food company calls it “a relatively expensive solution,” and a luxury fashion brand weighing it names pricing as its main drawback — while others say it is worth paying for. How buyers fund the work, and where the cheaper routes go, is on what buyers actually pay for AI-visibility tools.
Which AI visibility complaints appear alongside actual switching decisions? The clearest replacement evidence here is companies moving AI-search reporting off an SEO suite onto a dedicated or broader platform. Some judged the suite's AI features behind; others were consolidating overlapping tools or had new requirements. A software company cancelled Semrush after moving its AI tracking to Profound; a consumer-goods company that finds Semrush's AI features less mature is running a request for quotation for a dedicated platform; a data company plans to retire Semrush once a content-and-visibility platform is in, because the two overlap; and a hotel company is consolidating its search reporting on Conductor partly for its AI reporting. Among the dedicated tools, buyers describe re-evaluating: a luxury fashion brand is weighing alternatives to Scrunch over coverage and limits, a women's health brand is reviewing Profound at renewal over its method, and a skincare brand expects to keep evaluating during its Scrunch trial. No completed move here was driven by a data or method complaint alone, and some pilots end without a purchase for reasons buyers do not give. This supports an association in these interviews, not a measured cause of churn; a complaint that accompanies a move is not necessarily why it happened. Many complaining users stay. For the SEO suites themselves, see Semrush alternatives and top complaints about SEO tools.
Want your own stack read the same way?
Get mine →The stories behind the complaints
The counter-current: some of the complaint belongs to the category
Why do buyers rate tools well while complaining about them? In these interviews, many separate what a vendor controls from what nobody in the category can yet do. That is what they report, not something we measured. The large retail group that distrusts the method calls Profound “definitely the best solution available at the moment for enterprise customers like us.” A furniture retailer that had heard its tool might lose sight of one AI engine adds “I don't think that's specific to Profound”; Profound's pricing page currently lists both Claude and Perplexity among the engines its enterprise plan tracks. The same tools draw strong praise: a home-goods retailer says of Profound “We're very satisfied with it. Maybe even a nine,” and a spirits company rates Bluefish a ten for “analysis, measurement, KPIs, and content creation in a flywheel.” The implication, ours rather than a buyer's: before concluding that a tool is the problem, separate what it measures badly from what no tool can yet measure, and judge vendors on the first.
What this means — whether you run one or you're buying one
Six problems recur in these interviews: figures whose method buyers cannot see, coverage limited to the prompts and engines a tool runs, reporting without direction, products still being built, SEO-suite AI modules that trail the specialists, and results nobody can yet tie to revenue. The first two concern the tool's data and method, the next three the product, and the last a category too young to have settled how its effect is measured. None of this is evidence that one vendor is worse than another, that any tool shows a complete picture of how AI engines describe a brand, or that any one complaint caused a move. If you run or are evaluating an AI-visibility tool, our recommendations — not practices buyers describe — sit under each complaint above, and come down to three. Compare platforms like for like: your own prompt set, on the same engines, for the same weeks. Before trusting a tool's AI-referral count, check how it identifies an AI referral against how your analytics does. And set the pilot's test before it starts, along the lines under complaint 06, without reading a rise in visibility or referred conversions as proof of new revenue. For the ratings and lessons, see what AI-visibility buyers wish they'd known; for the market, the scramble to be the AI's answer and what is answer engine optimization?; for a specialist against a suite, Profound vs. Semrush; for the SEO tools around it, why companies switch SEO tools.
Common questions
What are the most common complaints about AI visibility tools?
In these interviews six recur: visibility figures whose method buyers cannot see, coverage limited to the prompts and engines a tool runs, reporting that stops short of telling the team what to do, products still being built, SEO-suite AI modules that trail the dedicated tools, and results nobody can yet tie to revenue. The first two are about the tool's data and method, the next three about the product, and the last about a young category. These are patterns in the interviews, not failure rates. Price is as common as any of them among Profound users.
What are the biggest complaints about Profound?
Price, and trust in how its figures are made: users question how it finds answers and projects prompt volumes, and several call it expensive. After those come a learning curve, agents that one user says are not yet working as well as hoped, and API access that sits with its enterprise plan. Other users praise its interface, its reporting depth and its engine coverage. Profound rates around 8/10.
What are the biggest complaints about Scrunch?
One user reads its data as directional and another finds its reporting stops short of telling the team what to do. Others describe a product still maturing: one trial found it buggy and its onboarding confusing, and a long-term user wants broader engine coverage, more flexibility in adding prompts and a larger page allowance for its optimization product. Other users praise how easy it is to configure and use. Scrunch is rated by too few buyers to publish a figure.
Are SEO tools' AI visibility features as good as dedicated AI visibility tools?
Some buyers who compared them say the modules in their plans did not meet their needs. They describe the AI modules in Semrush and other SEO suites as less mature, slower to add AI features or missing from the plan they run, while rating the suites well for SEO. Several of the replacement decisions in these interviews move AI-visibility work off an SEO suite, though some of those are about consolidating tools, and none shows that a complaint caused the move. These are judgments about particular plans and dates, and the suites have since added AI-visibility features, such as Semrush's AI Visibility Toolkit and Ahrefs' Brand Radar.
Can you trust the data from AI visibility tools?
Buyers who use them mostly read the figures as directional. AI answers vary by prompt, engine and moment, and no tool observes every answer every engine gives every user. A visibility score rests on the prompts and engines a tool runs; an estimated prompt volume rests on evidence of what people ask, projected to a wider population. Buyers question both, and mostly how little of either method they can see. The implication, ours rather than a buyer's, is to ask each vendor for the two methods separately and to check its referral figures against your own analytics.
Your stack, read by the buyers who've run it
Peer-powered: built from hundreds of verified buyer interviews, and fitted to the tools you already run. You get a shortlist for your stack and your next steps.
Get my stack readFree · no password · your answers are anonymized before they inform research like this.