Serpstat AI Playbook: LLM Brand Monitor, AI Overview Tracking and What They Cost

Three AI products, three separate bills

Serpstat sells AI in three places, and each one is billed differently. LLM Brand Monitor tracks how chatbots answer questions about your brand and has its own Free and Pro plans. AI Overview tracking lives inside Rank Tracker and spends the credits of your normal SEO plan. The MCP server lets AI assistants query Serpstat's data and spends API credits, which start on Team. In practice that covers two jobs: measuring how AI systems show your brand, and using AI to do SEO work faster. Keep the two apart. They are priced apart.

The verdict is simple. The AI Overview tools inside the SEO plans are the best part of the offer: cheap, inside the plan, open through the API and well documented in the changelog ^[1]. LLM Brand Monitor is a capable, separate product with a small free tier and an introductory Pro price. Some of its marketing numbers disagree with each other, so test before you trust them. A team that wants all three pays for Team and for LLM Brand Monitor Pro, which comes to $168 a month at monthly rates by our arithmetic. Start with the free checks.

What LLM Brand Monitor actually measures

LLM Brand Monitor tracks how AI assistants describe your brand when users ask questions. It shows which models mention you, how often, with what sentiment and how competitors compare. It runs on serpstat.com and on its own site, llmbrandmonitor.com, which says the product is built by Serpstat. Same company, separate checkout. Each scan runs your prompts against the models you chose and records whether, where and how your brand appears, with recommendations attached to the results. The help article lists five metrics ^[2]:

• Mention rate: the share of monitored queries in which a model names your brand. Track 20 prompts, get named in 6 answers on one model, and your mention rate there is 30%.
• Share of voice: your mentions as a proportion of all brand mentions in the query set. If three rivals share 40 mentions and you have 10, your share is 20%, by our arithmetic.
• Sentiment: each mention is marked positive, neutral or negative, and a July 2026 update added a filter to isolate each group.
• Competitor positioning: up to five competitors per project, with the models that favor them, how often and with what framing.
• Mention trends: how your rate moves across scans. A sudden drop may signal a model update or new content shaping the model.

Reading sentiment and trends without fooling yourself

Sentiment is where teams misread the dashboard. The help article warns that "Neutral framing in a comparison context can still put you at a disadvantage" ^[2]. Neutral is not safe. A negative mention may be a fair limitation, and a positive one may come with a wrong price, so read the answer before you react and fix facts first. Trends need patience as well. Monthly snapshots tell you little, while weekly scans held for two or three months start to show a pattern you can act on.

140 models, or 100?

The May 28, 2026 help article says the tool checks 140+ models, ChatGPT, Claude, Gemini and Perplexity among them, and the pricing pages agree. The LLM Brand Monitor page on serpstat.com is less consistent: one block says 100+ AI models are monitored, while its own FAQ says 140+. We take 140 as current. The count matters less than which models your buyers use, and you choose those per project. Every extra model multiplies checks, so start with the four the help article names and add a fifth only if your audience uses it heavily.

Setting up a project that tells you something

You create a project, enter your brand and up to five competitors, add prompts or generate them, choose models and set a scan schedule. Serpstat promises five minutes to the first actionable insight. Plan an hour instead. The Generate button suggests prompts from your brand and category, but generic prompts return generic answers. Track the questions buyers ask before they buy: comparisons, best-of lists, price questions and problem questions, phrased the way your sales team hears them. Serpstat's Perplexity suggestions show real AI search phrasing. Avoid prompts that name your brand. They measure recall, not recommendation.

The August 2026 housekeeping features

Three August 2026 updates make larger projects manageable. Brand name variations let you add several spellings, and all of them count as one brand during tracking; add abbreviations and common typos, or a model that writes your name differently counts as a miss. The tag filter now scopes every metric across the report, so one project can hold prompts for several product lines, which matters on a Free plan limited to two projects. And since August 16 you can paste two columns from Google Sheets, prompt and tag, with no CSV formatting. Keep the prompt list in a sheet. It doubles as team documentation.

Scans, transcripts and exports

Pro includes weekly automated scans. The standalone site queues them and runs every prompt against every selected model in one pass, and you can also trigger a scan on demand. Each run spends checks. Chat Mode shows every answer as a conversation, filterable by tag, model or sentiment, and since June 2026 a full dialogue exports to PDF, Markdown or DOCX with its date, model and project. Read the transcripts. Scores hide the sentence that hurts you. Results also export to PDF or CSV in one click.

What the checks really cost

The unit to watch is the check. One check is one query run against your selected model set, and the help article gives the example plainly: "Running 5 queries across 4 models uses 20 checks" ^[2]. Checks therefore scale with prompts times models times scans. Ten prompts across five models scanned weekly use about 217 checks a month, by our arithmetic. Count before you choose models.

The Free plan includes 10 checks a month and two projects, with email support ^[3]. Ten checks is two prompts across five models, once a month. Enough for a first look. Not for tracking. Pro is listed at $39 a month as a limited-time offer, with 200 checks, unlimited projects, automated scans and Chat Mode. Serpstat's own page shows the list price as $99. Budget for $99, and enjoy $39 while it lasts.

Top-ups, and two different minimums

Credits top up any plan, Free or Pro, and they do not expire. The pricing page on llmbrandmonitor.com lists a minimum top-up of $10 at 25 cents per credit. The help article and the product page instead price the smallest top-up at 100 credits for $25. Both give $0.25 a credit and differ only on the minimum, and we treat the pricing page as current. Top-ups suit a one-off audit better than a standing monthly plan.

A worked budget

Take a brand tracking 10 prompts across 5 models every week. That needs about 217 checks a month. Pro covers 200, and the other 17 cost about $4.25 in credits, so the monthly bill lands near $43.25 at the offer price or $103.25 at the $99 list price, by our arithmetic. Trim one model and it fits Pro. Each added model costs about 43 more checks a month at that prompt count, which is why the model list deserves more thought than the prompt list.

AI Overview tracking inside the SEO plans

Since September 25, 2025, Rank Tracker can track your domain in Google's AI Overview on desktop and mobile, with reports in Project Overview, Position History and Pages ^[1]. The Rank Tracker page adds that you can compare organic and AI Overview rankings and follow competitors' AI Overview positions. The cost is double a normal check: each one consumes 2 credits per keyword and region, because it needs two Google requests. On Individual's 10,000 monthly checks, AI Overview tracking alone covers about 166 keywords a day, by our arithmetic. Choose the keywords where AI Overviews appear. The rest waste credits.

Brand Opportunities and the API

The AI Overview Brand Opportunities report, launched December 11, 2025, sorts keywords into three groups: mention with link, link only, and mention without a link. It started in the US and Ukraine and was temporarily available to all users, but the pricing page now lists it among Team features ^[4]. The API followed the same path. It added AI Overview filters to domain keywords in August 2025, AI Overview fields in domain history in March 2026, and two methods, getAIOSummary and getAIOBrandOpportunities, in April and May 2026. Agencies can pull that data into their own dashboards on Team.

The rest of Serpstat's AI, feature by feature

Beyond the two monitoring tools, AI shows up across Serpstat in smaller ways. Since July 23, 2026, Market Research shows AI Traffic, AI Score, AI Keywords and AI Citations next to SEO metrics, with a toggle for SEO-only, AI-only or the default combined view, which is useful for sizing who wins AI answers in a niche. Since August 2025, traffic estimates use machine learning, which Serpstat says lands closer to real clicks from Google Search Console. It changes every traffic number you see, so compare it with your own analytics.

Keyword Research gained semantic AI Search in June 2025, which finds keywords close in meaning even when phrased differently, for all users. In February 2025 Serpstat added more than 16 million Perplexity suggestions to its US database. The AI content tools write articles in 26 languages, meta tags, FAQs and paraphrases on Team and above, but they do not use your Rank Tracker or LLM data. The AI content detector estimates how much of a text AI wrote. Treat its score as a hint.

Assistants, Skills and agents

Serpstat offers two custom GPTs in ChatGPT, which need a paid ChatGPT plan; one of them, the SEO Copywriting Brief Generator, needs no Serpstat account. The MCP server connects Claude, ChatGPT and Gemini to Serpstat's data, with all 65 tools across 9 categories in the May 2026 help article, and our Serpstat MCP guide covers setup. On top of it sit Serpstat Skills for Claude, released March 19, 2026, which produce five complete SEO reports from one prompt and import from GitHub in one click. They need MCP, and MCP needs Team. Product pages also advertise stus.ai, with 10 AI agents called by commands such as @contentgap. We did not test it and found no pricing. Treat it as a preview.

Serpstat next to Semrush, Ahrefs and SE Ranking

On monthly billing, Team at $129 plus LLM Brand Monitor Pro at $39 comes to $168 a month. With Team on annual billing it is $139 a month, or $1,668 a year, and at the $99 list price the year costs $2,388, by our arithmetic. The offer matters. Semrush builds prompt tracking into its plans: 50 prompts a day on Starter, 100 on Pro+ and 200 on Advanced, across Google Search, ChatGPT, Perplexity and Gemini, with Starter at $199 a month. So Serpstat's pair costs less at the offer price, while Semrush tracks daily.

Ahrefs sells Brand Radar AI to research brands across more than 476 million organic prompts and lists its starting price at $199 a month, and even Lite includes 5 tracked AI prompts updated daily. That is a research database, while Serpstat monitors your own prompts. SE Ranking sells an AI Search add-on with 200 prompts at $71.20 a month on annual billing or $89 monthly, covering AI Overviews, AI Mode, Perplexity and ChatGPT, and its Core plan already tracks 100 prompts daily. On included prompts, SE Ranking is more generous. Our Serpstat vs Semrush comparison covers the wider picture.

A 30-day plan for AI visibility

The plan below uses Free for discovery, then Pro for four weekly scans, with AI Overview tracking on Team. It assumes one brand, five competitors and ten core prompts. Adjust the counts to your budget, but keep the order, because each week builds on the baseline from the week before. By the end you have a baseline, a list of competitor gaps and a first report.

1. Week one, baseline. Run the Free plan's ten checks on your two most important buyer questions across five models, and note which models name you and which name rivals. Then add your top 50 commercial keywords to Rank Tracker with AI Overview tracking on. You now have two baselines.
2. Week two, competitors. Upgrade to Pro, build ten prompts tagged by product line, add five competitors and run the first automated scan. Sort models by the gap between your mention rate and the leading competitor's; the help article flags gaps of 20 points or more.
3. Week three, content gaps. Open the Brand Opportunities report and list keywords where you get a link but no mention, or a mention but no link. Read the transcripts for the weakest model, fix the pages those answers draw on, then rescan.
4. Week four, report. Export the share-of-voice chart and the changed transcripts, put AI Overview visibility next to them, and report three numbers: mention rate, share of voice and AI Overview presence. Keep the same prompts next month, or the trend breaks.

What AI monitoring cannot tell you

LLM Brand Monitor shows what models said to your prompts on the day of the scan. It does not show what millions of users asked, or how answers change per user. The product page claims that 73% of users now prefer AI over search engines, with no source given, and its testimonials name people but not companies we could check. We could not confirm the current model count beyond 140-plus, the end date of the $39 offer, or stus.ai pricing. The standalone site calls itself privacy-first; we did not review its policy and terms. Prompts are stored with your project. Read those documents first.

Who should use it, and who should skip it

Brands and agencies that already see buyers arrive from ChatGPT or Perplexity are the natural users, along with SEO teams that need to report AI Overview presence next to rankings. Serpstat Team users get the AI Overview half at no extra cost. Skip LLM Brand Monitor if your buyers do not use chatbots to research purchases, or if you want daily prompt tracking inside one SEO plan, which Semrush and SE Ranking offer, or if $99 a month is out of reach.

The bottom line

Use the Free plan to see how AI describes you today. If those answers matter to your sales, take Pro and track ten prompts weekly, then turn on AI Overview tracking for the keywords that actually show one. At the offer price that pairing costs less than Semrush's entry plan, with less daily tracking. Report the same three numbers every month and keep the prompts fixed. For the plan-by-plan math, read our Serpstat pricing guide.