Topics & AI Prompt Estimates, a Report Builder & Agents & API

Features

• Topics and Estimated AI Prompts — On Pro and above, prompts are grouped into topics and can show about how many times a month people ask AI that question. With enough Google Analytics data, the estimate is tuned to your own site, and the tip says so when that changes the range. A prompt with too little data says "No estimate available", and a small range reads "Under 10".
• AI Visits per Prompt — With Google Analytics connected, the Prompts overview, Your prompts and each prompt's page show the visits AI engines sent to the pages cited for that prompt.
• Buying Stage — Each tracked prompt is tagged with the stage of the buying journey it belongs to, so you can see where AI mentions or cites you and where it doesn't.
• Report Builder — Build a PDF report step by step: pick the dates, sections and notes, check the data, and save its sections and settings as a team template. Agencies can build one report across every client. Reports have a new cover and an At a glance page, and show a dash instead of 0 where nothing was measured.
• Agents & API — On Pro and above, a new Settings page connects your own AI assistant to your data (Claude, ChatGPT, Cursor, VS Code or any app that supports MCP) and gives you API keys to read your data from your own code. Organization admins can switch it on or off and see recent calls from the last 30 days.
• Start Here — Improve opens with a Start here tab that tells you what to do first, in plain words, grouped into ranked page and article steps.
• Export Your Visibility Data — Download the Prompts overview as a CSV.
• One Google Analytics Sign-In for Every Workspace — Agencies sign in to Google Analytics once for the whole account, then pick a property for each workspace.
• Shared Prompt Pool for Enterprise — Enterprise accounts can share one prompt allowance across all their workspaces, with a Prompt pool page that shows what is used and a notice when it's full.

Improvements

• A Calmer App — The shell, Home, Prompts and every dashboard page have a cleaner look, and charts are easier to read, with daily bars, a score dial and labels at the end of each line.
• Improve Pages, Simplified — Pages opens with a simple summary and a map of your pages, and a page's fixes open in a side panel. To do shows one row per page for its open fixes, and page scores and speed match the Speed tab.
• Brand Setup and Site Health in Tabs — Brand setup is one page with tabs for Your brand, Competitors, Markets, Files and Apps. Site health has tabs for Overview, Can AI reach you and How AI describes you.
• More AI Engines Checked — Can AI reach you now also checks Meta AI, Alexa, DuckDuckGo and Mistral, with example fixes.
• Presence and Perception — Each prompt shows the two halves of its visibility score, and every score and rate uses the same name in the app, PDFs and alerts.
• Your Plan Follows You to Checkout — The plan you pick on the Pricing page carries through sign-up to checkout.
• Phones and Foldables — Pages fit better on foldable phones and small tablets.
• Smoother Updates — If a tab left open through an update can't load a page, it reloads into the new version on its own, once nothing you're typing would be lost.

Fixes

• One Brand, One Row — When AI writes a brand two ways, Share of Voice counts it as one brand, and AI engines no longer appear as brands in brand lists.
• Share of Voice Markets — Share of Voice says which markets the latest analysis measured, asks you to set a market before the first run, and All markets engine cards read every market together.
• Room for Longer Answers — Perplexity, Claude and Gemini have twice the room to finish long answers.
• Deleted Prompts Stay Deleted — A deleted prompt's answers leave live scores, rates, answer totals and search, and the prompt can't be run or brought back by mistake.
• Not Measured, Never 0 — Unscored prominence, tier and certainty show a dash instead of 0.0, and a page that fails to load says so instead of showing empty or partial data.
• Smaller Fixes — A prompt's Last run is dated by its newest answer, Journeys no longer fail on sites that are working, the visibility trend line fills the chart from day two, the portfolio PDF lists clients with no AI answers yet instead of ranking them, and Brand setup keeps a saved name when you come back to it.