Signals, powered by Scrunch, tracks how AI platforms such as ChatGPT, Gemini and Perplexity mention your brand, and links each significant change to the pages that contributed most to it. Monitoring, a recommended response and the page edit now sit in one platform.
Your team still has to judge each signal: whether the change matters, what caused it and how much to do about it. Name the person who makes that call before you switch Signals on, and give them the authority to decline a recommendation.
What Signals, powered by Scrunch, does
Sitecore acquired Scrunch on 3 June 2026, describing it as a platform that shows where a brand is “appearing, missing, or misrepresented” in AI answers. Three months later the first integrated capability arrived in SitecoreAI.
According to the changelog entry of 1 September 2026 and the Signals documentation:
- It monitors AI platforms. Scrunch tracks mention rate, citation rate and competitive presence across platforms including ChatGPT, Gemini and Perplexity.
- It raises significant changes as signals. Each signal explains what changed, its impact, and “a recommended direction for action”. Signals appear on the SitecoreAI Strategy page.
- It links signals to pages. Only signals with associated pages reach SitecoreAI. Each carries page-level opportunities: AI-generated recommendations for the pages that contributed most to the change, reviewed and applied in Page builder.
Two related features followed on 16 September 2026, both in early access at first. AEO opportunities lists a site’s top opportunities in Channels, with the affected pages and “projected lift, where available”. Its per-page recommendations appear in Page builder, where an editor applies or declines each one. Sitecore’s content optimisation feature, Optimize content with AI, scores content for “citeability and retrievability” in Page builder. Accepting a recommendation creates a new version of the content item, which then moves through the existing workflow and approval process before publishing.
Two terms need pinning down. A mention is the brand being named in an AI answer. A citation is a page being referenced as a source for that answer. They can move independently, and they matter differently depending on whether you need awareness or visits.
Where Signals fits in the loop from evidence to action
Sitecore describes a continuous cycle of monitoring and improvement. Laid against the chain we use to move from evidence to execution, the platform covers the stages at either end well. The highlighted stages are where the feature produces evidence and where people have to decide what the evidence means.
Context
What the organisation needs from AI discovery: which topics, audiences and products matter commercially.
Held by the team. The platform cannot tell which topics are worth defending.
Signal
A significant change in how AI platforms mention or cite the brand, or in competitive presence.
Signals, powered by Scrunch, on the Strategy page.
Interpretation
Whether the change matters, and what may explain it.
Each signal suggests an impact. People test it against competitor activity, campaigns and platform changes.
Decision
A proportionate response, which can be to watch and wait.
The recommended direction is one input. Priority against other work is a human call.
Orchestration
Who does the work, in what order, under which governance.
AEO opportunities in Channels group affected pages. Ownership and sequencing stay with the team.
Execution
The content change itself.
Recommendations applied in Page builder, through existing workflow and approval.
Measurement
Whether the change had the intended effect over a sensible period.
Scrunch keeps monitoring. Attributing a movement to one edit stays uncertain.
Measurement becomes the next signal, so the quality of the loop depends on the interpretation and decision stages between each turn.
Why AI discovery changes the shape of content operations
In a typical enterprise content operation, search visibility analysis, content planning, CMS editing and performance reporting sit with different people, in different tools, on different cadences. A report identifies a problem, a plan gets written, and weeks later somebody edits a page.
Signals compresses that sequence. The observation, a suggested reason and a recommended edit arrive in the platform where the edit is made. When we wrote about the launch of SitecoreAI in November 2025, Sitecore’s framing was agentic orchestration across the content lifecycle. This is one of the first places that framing reaches an everyday editorial task.
Plan for the operational change. When analysis and editing sat with different people, the handover gave someone time to question a recommendation before it reached a page. In one platform an editor can accept it within minutes, so build that review in as a deliberate step.
A signal reports a change, and strategy decides what it means
Sitecore’s own documentation is candid here. It notes that a change in AI visibility can also come from competitor activity, campaigns, or changes in how AI platforms work. Those causes sit outside your own pages, and a page edit may do little about any of them.
Take a common case. A signal reports that citation rate for a product category has fallen, and a competitor’s presence on the same topic has risen. That pattern is consistent with several explanations: the competitor published stronger material, an AI platform changed how it selects sources, a campaign of yours ended, or your own pages drifted out of date. Each explanation points to a different response, and only one of them is a page edit.
| Question | What Signals and AEO opportunities supply | What the team supplies |
|---|---|---|
| Does this change matter? | The size of the movement and the affected pages | Whether the topic carries commercial weight this quarter |
| Why may it have happened? | An explanation of impact and the pages that contributed most | Knowledge of campaigns, competitor moves and platform changes the monitoring cannot see |
| Which response is proportionate? | A recommended direction and page-level edits, with projected lift where available | The choice between editing, restructuring, waiting, or escalating to a wider content decision |
| How does it fit wider priorities? | A ranked list of opportunities for one site | The trade-off against roadmap work, brand consistency and other channels |
| Did it work? | Continued monitoring of mention rate, citation rate and competitive presence | A measurement window agreed in advance, and caution about attributing a movement to one change |
Everything in the team’s column is context the platform does not hold. Treat each recommendation as evidence to weigh, and make one person accountable for acting on it.
Trade-offs to weigh before adopting the loop
Page-level edits and structural causes
Opportunities are expressed per page. Some causes of weak AI visibility are structural: information that exists only in PDFs or behind forms, thin coverage of a topic across the site, or an information architecture that buries answers. A stream of page-level fixes may improve individual pages while leaving the pattern intact. When several signals point to the same area, look at the content model or the information architecture before editing more pages.
Consistency across reused content
In SitecoreAI, one content item can appear on many pages. Once a recommendation is accepted, content optimisation shows which other pages reuse the same content item before the change moves on to approval, which is a useful safeguard. Teams still need a rule for who approves a change that affects pages beyond the one being optimised, and how recommended wording is checked against brand voice.
Optimising for what is measured
Monitoring covers named AI platforms, and each one selects sources in its own way. Google’s guide to optimising for generative AI features, published in May 2026, describes optimising for generative AI search as “still SEO”. It lists “rewriting content just for AI systems” among the things site owners do not need to do. Keep recommendations anchored to what makes content clearer for people, and question any edit whose only purpose is a better score.
An inside view of a competitive question
Signals measures your brand’s presence and competitive presence on tracked platforms. It is well suited to telling you that a competitor is gaining ground. Why they are gaining ground sits in their content, their site structure and their experience, which is a different kind of evidence. That outside-in comparison is what our AI discovery assessment looks at: what makes information understandable, retrievable and citable, compared across a primary site and its competitors.
Access and cost
All three capabilities need a Scrunch subscription on top of SitecoreAI. On 2 October 2026 Sitecore made Signals and AEO opportunities generally available, together with Knowledge Studio, which prepares approved brand documents as grounding for AI-generated content. Content optimisation arrived in early access. “Projected lift” is a vendor estimate and belongs in a prioritisation discussion as one input among several.
What to establish before you switch the loop on
For a digital leader deciding whether to adopt an AI-visibility optimisation loop, in SitecoreAI or elsewhere, settle these first.
- The topics that matter. A short, agreed list of the subjects, products and questions where AI visibility has commercial value. Without it, every signal looks equally urgent.
- An owner for interpretation. A named person or role who reviews signals on a regular cadence and decides whether each one warrants action. Editors act on recommendations once that review has happened.
- A threshold and a decision record. A rule for what size of change triggers a response, and a simple log of decisions, including decisions to wait. The log is what lets you learn which responses worked.
- Governance for AI-generated edits. Confirm that recommendations go through your workflow and approval, decide who approves changes to reused content items, and check recommended wording against brand guidance.
- A measurement window. Agree in advance how long you will wait before judging an edit, and accept that competitor activity and platform changes will blur attribution.
- An escalation route. A way to recognise when repeated signals point beyond individual pages, to content strategy, information architecture or platform capability, and a forum where that decision gets made.
None of these needs the feature to be switched on, and most are worth doing whichever monitoring tool you use. With them in place, Signals, powered by Scrunch, gives your team a fast route from evidence to a considered edit, and a named person decides what each signal means.
