Changelog #0008

Seeing a competitor recommended for a question your business can answer is useful. Deciding what to do about it, getting the work published, and checking whether anything changed takes much more effort.

Since our last change most of our work has gone into joining those steps up. Here's what's new.

A brand Overview that follows the whole journey

Overview is now the first page for every brand. It follows your brand from its website to the visits AI sends you, in six stages of the AEO pipeline:

• Readable: Your site readiness score.

• Read: Visits from AI crawlers.

• Cited: How often answers cite your own site.

• Named: How often answers mention your brand.

• Recommended: How often answers recommend your brand.

• Visited: Referral visits from AI engines.

The pipeline narrows where you lose the most between one stage and the next, and marks the biggest leak. Select a stage to see its breakdown, or follow the link to the page behind it. Stages you haven't set up yet, such as AI traffic before your site is connected, link to the step that switches them on.

AI answers vary from run to run, so every rate on Overview now shows the range it's likely to sit in, such as Likely 28–41%. Arrows only appear when a change is bigger than that variation could explain. Otherwise you'll see No real change. Brands within normal variation of each other share a place, such as Level with Acme for #1, rather than being ranked on noise.

That means fewer false alarms, and fewer wins claimed on a lucky week. Your weekly overview email follows the same rule.

Visibility is now simply your mention rate: the share of captured answers that name your brand. The old blend of mentions and position has gone, so your visibility figure may look different from before. Average position is still shown separately.

Press ⌘K/CTRL-K to jump to any brand or page, or paste a website to add a new brand. All brands shows each of your brands as its own pipeline in one list.

See how each engine really answers

Every engine now searches the web using its own app's search: OpenAI's for ChatGPT, Anthropic's for Claude, Google Search for Gemini, Meta's for Muse, and Perplexity's built-in search. If an engine's own search isn't available, the run fails rather than quietly using a different search engine. Google AI Overviews and Google AI Mode come straight from Google's results page, and you can track them separately from Gemini.

Each answer now shows more of how it was put together:

• Searches: What the engine actually searched for while writing the answer. Engines often rephrase your question, so this shows the wording your pages need to match.

• Citations: Only the pages the engine reported as sources. Where an engine ties a source to a sentence, a small pill with the site's icon appears after that sentence.

• Links in the answer: Links written into the text, such as Google linking a company name to its website, kept separate from citations.

• Read but not cited: For Claude, pages it found in its search results but chose not to use. If one of yours appears here, this indicates it was found but didn't answer the question directly enough.

Because citations now count only reported sources, some of your citation totals may fall. The new figures are a more accurate record of what the engines say they relied on.

We've updated the engines we track. Cleotic now monitors ChatGPT, Claude, Gemini, Muse, Perplexity, Google AI Overviews, and Google AI Mode, using current versions of each. Grok, Qwen, and Seed are no longer monitored.

Scheduled collection runs on Mondays, Wednesdays, and Fridays on every plan, with web search on every run. Charts plot points on collection days and show when the next collection is due. Prompts and prompt suggestions also show estimated monthly AI searches where data is available, to help you choose which questions to track.

Give each opportunity an owner and a next step

Opportunities is a shared action inbox for the work your evidence suggests. It covers:

• Content gaps: Questions where competitors appear and you don't.

• Readiness findings: Problems from your latest site scan.

• Source gaps: Third-party pages that AI answers cite, which list competitors but not you.

• Traffic gaps: Pages AI bots crawl often but rarely send visitors to.

• Coverage gaps: High-demand questions none of your prompts asks yet.

Each item shows its evidence, a score, and one suggested next step. Assign it to a colleague, comment on it, snooze it, or work through the list with keyboard shortcuts. Work moves through Now, In flight, Watching, and Won.

Once you publish a page, ship a fix, or pitch a source, the item moves to Watching, and Cleotic compares the evidence before and after. When the gap closes, it suggests calling it won. Only a person can make that call. If the gap returns, the item goes back to Now.

Turn website findings into fixes you can check

Site readiness now scans your site every week and can email you the results. History explains why a score moved, separating changes on your site from changes to the checks themselves.

The action plan lists affected pages, likely effort, and who should own each fix. Send findings, or individual pages, to Opportunities and assign them. Once you've deployed a fix, Verify fix re-checks those pages straight away, without using your checks allowance. If the issue is still there, the item goes back to Now.

With AI traffic tracking connected, By demand puts pages with the most AI crawls, referrals, citations, and searches at the top. Business and Custom plans can send branded readiness reports to clients. Custom plans can scan up to three competitor sites and compare every check side by side.

Connect your website once

All website connections now live on one Site page. For WordPress, one approval through the Cleotic plugin connects both AI traffic tracking and draft delivery. If someone else manages your website, send them a connection link and let them finish that part.

Smaller changes

• AI citations: Now has separate Overview, Domains, and Pages views. Every cited page has a signal score, and Source gaps shows pages that list competitors but leave you out.

• AI traffic: Now separates crawler visits from referrals and lets you open any page's recent activity.

• Brands and studies: Projects are now called brands, with studies kept for research. Adding a brand starts from its website.

• Connected Agents (MCP): We've rebuilt our MCP server so your AI assistant can do more of what you can do in the app: triage opportunities, write Content Studio pages, fix and verify readiness issues, add brands, and share reports. It now connects through browser sign-in only and follows your role and brand access. If you connected with an MCP API key, reconnect using https://mcp.cleotic.ai.

• CLI: cleotic auth login now signs you in through your browser. CI and scripts still use a CLI API key.

• Docs. The documentation is reorganised to more closely align with the app structure, with new guides on how answers are collected, MCP tools, and the CLI.

• Mobile navigation uses Find and menu buttons instead of swipes.