Guided Setup, Prompt Pages & GA4 Outcomes

Features

• Guided Setup and Your First AEO Brief — New workspaces open a full-screen setup in five steps: Website, Your business, What to measure, Baseline and First finding. It suggests questions buyers ask AI about what you offer, with monthly AI searches where we have an estimate, ticks five, and tracks only the ones you keep. Nothing runs until you click Run my baseline; the first ChatGPT and Gemini answers then arrive within minutes, with progress shown for each engine, and setup ends on your first brief: one finding, the answer behind it and one next step. Your first Share of Voice analysis runs by itself a day later.
• Prompts and Sources — Five sidebar items become two. Prompts has three tabs: Overview, Your prompts and Find new. Sources has two: Cited by AI and Found in research. The Prompts Overview adds a grid of where each prompt wins by engine, with Searches/mo and AI names instead columns, and each site in Sources opens a panel listing the prompts it shows up for. Old links redirect.
• A Page for Every Prompt — Each tracked prompt now has its own page with its visibility score, how often you are mentioned and cited, the brands AI names instead and its monthly searches. Tabs cover Answers, By engine, Sources, How AI searches, Fixes and Perception, and Run now, Edit and Delete sit in the header. Each rate says how many answers it rests on, such as "6 of the last 13 answers".
• Google Analytics 4 Outcomes — With Google Analytics 4 connected, GA4 Traffic has a new Outcomes tab. Pick 7, 28 or 90 days, or a single channel, to see sessions, engaged sessions, key events and revenue against the period before, with a day-by-day chart, your landing pages, channels, devices, countries and the key events set up in your property. Up to 90 days of history fill in over the first nightly syncs. A day Google hasn't reported shows as a gap rather than a zero, revenue Google withholds reads "Hidden by Google", and connecting no longer shows Google's unverified-app warning.
• Questions Your Market Asks AI — Prompts › Find new suggests questions people ask AI about what you offer, grouped by intent, with monthly AI searches where we have an estimate. Track adds a question; Swap pauses a prompt that names your brand and adds the question in its place, so your count of active prompts stays the same. When nothing in the question list relates to your prompts, it shows questions asked by people looking for what you offer instead of an empty page.
• Opportunities: Real Priorities, One Card per Idea — Priority now reflects how often people ask AI about the topic, how ready your page is and whether you are mentioned but not cited, so not every idea reads High, and confidence rises with the number of engines that raised an idea. The same idea from several engines is now one card whose Model column shows each engine.
• Your Latest Finding on Home — Home shows your latest finding as one line under the Visibility trend. Open it for the quoted answer, how many answers it rests on and the next step, or go to the full brief. It updates as new answers arrive.

Improvements

• See What's Running and When — Home, Opportunities, Share of Voice and GA4 Traffic show when their data last updated and when the next run, scan or sync is due, in your own time zone. Empty pages say why they are empty, such as no scan yet or a Share of Voice run in progress. When an engine's answers are running late, a notice at the top says so and confirms that runs finish on their own and nothing is lost.
• AI Search Volume for Every Prompt — Every tracked prompt now gets its monthly AI search estimate in the background, so the number is ready on each prompt's page and in the Prompts Overview without opening Your prompts first.
• More Page Fixes — If you haven't added pages from your own site in Improve, up to 10 of its crawled pages are tracked for you after a crawl, so Opportunities can suggest fixes to them. Pages on www. and on subdomains of your domain count as yours, and prompts match their page despite punctuation, plurals and your brand's own words, so more ideas arrive as a fix to the right page.
• Run Now Runs Once — Run now runs a prompt once even when it is clicked in two tabs, on two devices or during setup, and a prompt that is already running is no longer reported as failed. When we're busy, the request waits and retries for up to two minutes instead of failing.
• ChatGPT and Gemini as People See Them — Run now and Share of Voice now take ChatGPT and Gemini answers from their apps, as your scheduled checks already do, and fall back to the engine's API only when an app answer doesn't come back in time. Since October 10, DeepSeek, Qwen, Kimi and GLM look things up through a new web search service, so their sources may shift.
• Buttons That Say Why — Workspace viewers, and anyone in the agency portfolio view, now see write buttons disabled with the reason instead of getting an error after clicking, across Data Hub, Share of Voice, Your prompts, prompt import, Opportunities and Generated FAQs. At your plan's client limit, Add client says so instead of failing.
• Share of Voice Every Week — Share of Voice refreshes itself once a week on every plan, and Run Analysis still runs it any time within your plan's monthly analyses. A new workspace's Share of Voice page says when its first analysis will run.
• How We Measure — Our measurement guide is now one short How we measure page in the docs: the engines on each plan, how often each checks, what each number means and why numbers move. It replaces the in-app Methodology page, and old methodology links lead there. Model updates and your own chart annotations still show on your trends.
• A Sharper Dashboard — Every dashboard page has a cleaner finish. Tables are ruled with your own row highlighted, every chart has the same readout, ranked lists replace pie charts, and engine logos appear in their own colors, with separate logos for Google AI Overviews and AI Mode. Home, Prompts and each prompt page now fit a phone.

Fixes

• Deleted Client Workspaces Stop — Deleting a client workspace now stops its scheduled checks, crawls, journeys, Share of Voice runs and scheduled reports right away, including work already under way. Restoring the client picks up where it left off.
• Prompt Rates Follow the Time Range — Since October 5, a prompt's 7, 30 and 90-day views could show the same mention, citation and source rates. Each view now covers its own range, and the change from the previous period shows again.
• Indexed Pages Stay in Their Workspace — With several client workspaces, a page indexed in one could overwrite another workspace's site-audit result for a page at the same address. Indexed pages now stay in their own workspace, and audit page lists and counts include audited pages only.
• Agency Checkout Confirms the Right Plan — After you buy the Agency plan, checkout opens the organization you paid for and waits for its Agency plan instead of confirming a personal plan.
• Pages That Load, Labels That Fit — Found in research no longer fails to load when one cited link is malformed, and Journeys findings filter by journey again. The report builder offers Model Comparison only when there are answers to show, and Documents shows file sizes and flags files that couldn't be read. Prompt cards no longer read "Unnamed prompt", the prompt page names your brand rather than the client workspace, and the Models tables show ChatGPT and Gemini as one row each.
• Omni Agent — New Chat stays new after a reload instead of reopening the previous conversation, and charts in the agent's answers no longer lose a line, bar or slice.