Frequently asked questions

AI Peekaboo tracks whether, where, and how your brand gets mentioned when people ask AI assistants questions, across ChatGPT, Google Gemini, Perplexity, Google AI Overview, and Google AI Mode. Instead of tracking where your website ranks on Google, it measures whether the AI itself recommends you, how you're positioned against competitors, and which sources the AI cites while answering. All five models are included on every plan, including Starter, the difference between plans is how many brands and prompts you can track and how often they run, not which models are checked. See What is AI Peekaboo? for the full picture.

Every tracked prompt is asked live against each AI model with real-time web search turned on, the same way a real person would ask it. Google AI Overview answers are captured at the city and geo level for accuracy, and if one provider has a temporary issue, the system automatically falls back to an alternate method rather than failing the check. Full detail: How we get the data.

It depends on your plan: Starter checks every 2 days, Peek and Grow check daily, and Enterprise is custom. Refreshes run on a rolling schedule from your last completed run rather than a fixed clock time, so a run can start up to about an hour after it's technically due, and a brand with a lot of prompts can take a few minutes to roughly an hour to fully complete once it starts.

Yes. The free report and the paid dashboard run on the same underlying engine, but the free report checks only 2 models with a lighter extraction pass and skips sentiment entirely, so treat any paid analysis as the more accurate, current number. See Free report vs. dashboard.

In short, it's the percentage of tracked AI answers where your brand appears, weighted by how prominently you're positioned, averaged across every model and every run in your selected time window, including the runs where you weren't mentioned at all. The full formula and a worked example live at The AI visibility score, explained.

Visibility measures whether your brand's name was mentioned in the AI's actual answer text. Citations measure whether your website was used as a source the AI drew from, which is a separate thing that shows up under the Sources tab, not in your top-line visibility percentage. You can be cited without being mentioned by name, and mentioned without your site being cited, so check both. Full explanation: Visibility vs. citations.

No, it's two different metrics working as designed. A 0% visibility score means your brand name wasn't mentioned in that answer, but your site can still show up as a cited source on the same prompt. Check Dashboard → Sources, or open the prompt's detail view, to see citations that don't show up in the visibility number. This is the single most common point of confusion in the product, so it's worth bookmarking Score is zero but you were cited.

AI models aren't deterministic, the same prompt asked twice can return a different answer, a different brand order, or different citations. Your score is an average across every run in the window, which smooths this out over time, but a small day-to-day wobble with no real-world change behind it is expected model variance, not a bug. More detail: Why scores fluctuate.

Average position is your average rank (shown as something like #1.8) across only the runs where you were actually mentioned. Unlike the visibility score, misses aren't counted here, since "how good is my rank when I show up" and "how often do I show up at all" are deliberately reported as two separate numbers. See Average position and market share for how market share is calculated alongside it.

You choose prompts per brand directly, up to 500, as part of your configuration. Tracking frequency is a separate lever, not tied to prompt count: your price is based on the total configured checks a month (brands times prompts times models times cadence), so a lighter cadence at the same prompt count costs less, down to the $29/month minimum. See Plans and pricing for the full formula and the cadence table.

If your chosen brands, models, and cadence would price below that $29 minimum, your prompts per brand are raised automatically to the smallest number that fills the minimum bundle, so you are never billed the floor price for a near-empty configuration. See Adding prompts and Plan limits.

The best method is pulling real queries from Google Search Console and rephrasing them as natural questions a customer would actually ask an AI assistant, favoring commercial, transactional, and investigational intent over purely informational questions. Purely informational prompts ("what is X") tend to waste tracking slots on questions that don't reflect buying intent. Full method: Choosing what to track.

Yes. Prompts can be edited, paused, archived, or deleted. Pausing stops it from being analyzed and it no longer counts toward your score, but keeps its history; archiving is restorable later. If you try to add more prompts than your plan allows, you'll get a clear message and the add action is blocked rather than silently failing, upgrade your plan or delete existing prompts to free up room. Details: Managing prompts and Plan limits.

You can add a competitor from real brands AI already mentioned in your answers, from AI-generated suggestions, or manually by name and domain. Competitors are scored with the exact same formula as your own brand, so the comparison is fair. Full detail: Adding competitors.

There's no fixed list of tiers. You build a configuration from brands, prompts per brand, which AI models to track, and how often to run, and the price is $16.67 per 1,000 configured checks a month with a $29/month minimum. Full breakdown: Plans and pricing.

A configuration priced under $100/month gets a 14-day free trial that requires a card at checkout; at $100/month or more, billing starts immediately at checkout. You get one trial per workspace ever, and your trial runs at the exact brand count and configuration you checked out with, not a scaled-down preview. See Free trial.

Yes. The one-time free report needs no signup or card and takes about 2 minutes. A free signed-in workspace (no subscription) lets you keep 1 brand with up to 25 prompts and 3 tracked competitors, with most paid-only areas shown as a preview with an upgrade prompt. See Free account.

Upgrades and new subscriptions run through in-app checkout. Downgrades, cancellations, payment method changes, and invoices are all handled through the Stripe-hosted Billing Portal, reached from Settings → Billing, which stays reachable even if your subscription has lapsed. See Managing your subscription.

Yes, custom volume plans are available, contact us to talk through brand count, prompt count, and tracking frequency for your specific volume. On custom plans, tracking frequency and prompt count trade off against each other at the same price point, more frequent tracking costs more, less frequent tracking with the same prompt count costs less.

Yes, the REST API v1 gives full read access to your brands, scores, competitors, sources, and prompts, plus write access to create and manage them if you generate a read+write key. It requires an active paid subscription on any tier. See REST API.

Yes, an MCP server lets you connect Claude Desktop, Claude Code, Cursor, or any MCP-compatible AI client directly to your own AI Peekaboo data using your API key, so you can ask it questions in natural language instead of digging through the dashboard. See Connect Claude via MCP.

Yes, a Looker Studio connector pulls your visibility score, citations, and prompt-level data straight into your own Looker Studio dashboards, included on every current plan. One common setup mistake: sentiment needs to be added as a dimension (positive, neutral, negative), not a numeric measure, or it will display as a single confusing number. See Looker Studio.

The most common approach is pulling data through the REST API or MCP and assembling it into your own branded report. See Custom reports for the pattern.

There's no direct integration with the Gemini app specifically, CSV export and import is the workaround if Gemini is your AI tool of choice. Agency Analytics isn't natively connected either, the interim workaround is pulling your data via the API into a Google Sheet, then feeding that sheet into Agency Analytics. Octoboard has been requested but isn't built.

Yes, every current plan (Starter, Peek, Grow) has an Agency pricing variant that trades fewer prompts per brand for the ability to track 2 to 5 client brands under one subscription, so you don't need a separate product per client.

Generate a public share link for their brand from Settings → Share Links. It's a revocable, read-only link with no login or invite email required, giving the client visibility into Dashboard, Prompts, Competitors, and Sources for that one brand only. You can rotate or revoke it anytime, and set an optional expiry. See Client share links.

Yes, where enabled, white-label replaces the AI Peekaboo logo and applies your own brand colors across the app and exports. It's not included on any self-serve plan, it's bundled with Enterprise or scoped as a separate agency engagement. There is currently no custom domain option, that's on the roadmap but not shipped, so a white-labeled report still lives on an AI Peekaboo URL today. See White-label.

Teammates can be added with an Admin or Member role, and there's no limit on the number of seats or extra cost for adding them, since billing is per workspace, not per user or per seat, everyone shares one subscription, one bill, and the same limits. Only Admins can start or change billing. See Team members.

No. AI Peekaboo tracks ChatGPT, Gemini, Perplexity, Google AI Overview, and Google AI Mode, but does not track or monitor Claude's own answers to prompts, that's outside what's supported today. See What AI Peekaboo cannot do for the full list of boundaries.

No, tracking starts from when you add a brand and prompts, there's no backfill of historical AI answers from before that point. If you need a "before" baseline for a new client, use current mentions and cited sources as your starting evidence instead of a trend line you don't have yet, see Reporting before you have a baseline.

Not today. There's no proactive email alerting for score changes or competitor movement, only two transactional emails exist right now: your report-ready email and a plan welcome email. Check your dashboard directly, or ask about it in Ask AI, to catch changes.

No, it measures visibility, ranking, sentiment, and citations, it does not attempt to prove a causal link to revenue or sales. Use it alongside your own conversion and traffic data if you need to build that case; see Measurement and decisions for how to think about tying the two together.

No. Content Briefs and the To-do/Action Plan tool tell you exactly what to write and why, grounded in real sources AI already cites for your category, but they deliberately stop short of generating finished articles for you. The product's stance is that AI-written filler content damages credibility, so briefs are handed to a human writer, not auto-published.

Email team@aipeekaboo.com with your workspace and brand name and what you're seeing, that's the fastest way to reach a real person. See Contact support.

Ask AI is an in-dashboard chat that can explain your metrics, tell you why a score moved, evaluate your prompt quality, and even build a quick custom chart for you, using only your real tracked data, it can't invent numbers. Free accounts get 3 total messages before an upgrade prompt; paid plans get unlimited messages.

There's no self-serve full account deletion button today. Email team@aipeekaboo.com to request full account or personal data deletion, and it's typically handled within 30 days. Deleting an individual brand or competitor inside the app is instant and self-serve. See Data privacy and deletion.