By Jakub Koba, CDO, Kogifi
On 11 November, at the Sitecore Symposium 2026 keynote in Orlando, Sitecore announced Marketing IQ: an agentic solution for running marketing at scale inside SitecoreAI. It is available to all SitecoreAI customers from the day of the announcement.
I was in the room. The moment that stayed with me came from a customer. IFS shared on stage how they cut a piece of marketing work from two weeks to two hours.
In Sitecore's announcement, Oliver Pilgerstorfer, Chief Marketing Officer at IFS, said: "It's one thing to understand how we show up in AI-based search results, but it drives little value unless you act upon it. That's the real value of MarketingIQ, which combines the insights of Scrunch with the publishing capabilities of Sitecore."
Here is what was announced, how Scrunch fits in, who decides what the agents are allowed to do, and what it means if your sites still run on Sitecore XP.
What Sitecore announced
Sitecore describes Marketing IQ (written MarketingIQ on its product page) as a SitecoreAI module that connects your brand knowledge with AI agent orchestration. The name first appeared in July, when Sitecore called it "the intelligence layer behind SitecoreAI", connecting content, brand context, customer data and performance. At Symposium it was the headline.
Three points from the keynote matter most:
- It delivers work on your guidance. You give it a goal, and agents prepare the brief and draft the content and experiences, using your approved messaging, product facts and existing assets.
- It proactively suggests what to do next. It picks up signals and proposes the next piece of work without waiting for someone to ask.
- It is designed to work in a closed loop. As Sitecore described it on stage, every delivered experience and its measured results become input for the next action and for the guidance the agents work from.
The commercial facts, from Sitecore's own FAQ:
- Marketing IQ comes with SitecoreAI CMS and SitecoreAI DAM. It is not sold on its own.
- Existing SitecoreAI customers get access. Your organisation sets up the connections and permissions first.
- An XM or XP licence on its own does not include it.
- You can work with it inside SitecoreAI or from an AI assistant you already use, connected through Model Context Protocol (MCP).
Scrunch finds the opportunity, Marketing IQ acts on it
Sitecore acquired Scrunch in June 2026. Scrunch tracks how brands appear in AI-generated answers from tools such as ChatGPT, Gemini and Perplexity: where you are cited, where you are missing and where competitors are winning. Its Agent Experience Platform (AXP) serves AI crawlers a clean, structured version of your pages while people see the normal site.
At the keynote the two products were shown as one flow. Scrunch identifies new marketing opportunities and feeds them into Marketing IQ. Marketing IQ can then suggest and build new content, media and components, and deliver campaigns on its own. Add the closed loop and you get a full cycle: opportunity, build, deliver, measure, and the results feed the next action.
Most AI features in a CMS so far help with a draft someone already decided to write. This cycle can start the work, build the parts and run the campaign.
Scrunch also remains a standalone product that works with any stack, so you can measure your AI search visibility before any platform decision. Since October it has been rolling out its recommendations inside the SitecoreAI CMS, where teams with a Scrunch subscription can apply them page by page or across a domain. Scrunch states that every change is staged for human approval before it is published.
What "suggests what to do next" means for a marketing team
In practice, the work starts somewhere new. Most marketing work today starts from a request: a campaign brief, a product launch, a page someone asked for. With Marketing IQ, part of the queue starts from the platform. It might flag that a competitor now owns the AI answer for one of your key product questions, link that to the page that should be winning it, and arrive with a brief and a draft update ready.
That changes the team's job in three ways.
Someone has to triage. A system that suggests work will suggest more than you can ship. Somebody needs to own the queue and decide what gets done this week, what waits and what gets rejected.
Your source material decides the quality. Agents work from your brand guidelines, approved messaging and product facts. If those are out of date, scattered across decks or missing, the drafts will be confidently wrong in your own voice. The groundwork is one current set of messaging, product facts and guidelines that someone owns.
Measurement has to be in place before the suggestions start. The loop learns from measured results, so track the outcomes you care about, such as leads, bookings or sign-ups. Page views alone will teach it very little.
Who decides what the agent may do
IFS is not the only example of this kind of speed-up. Sitecore reported on 5 November that CWS Workwear, after moving to SitecoreAI, reduced content launch times "from six weeks to just six hours". Results like these come from removing the waiting and hand-offs around a piece of work. The same speed applies to mistakes, so the approval design matters more than before.
Sitecore stresses control. According to its product pages, every draft is reviewed before anything is published, agent access, actions and approval points are set by your team, and there are audit trails. That is the right starting point. The hard decisions are still yours. Once an agent proposes the next step, someone has to decide what it's allowed to do. A loop that feeds itself needs clear limits on what it can change without a human.
Before switching it on, write down answers to four questions:
- Which actions are suggest-only, and which can go to staging without asking? A metadata fix on a blog post and a new claim on a pricing page carry different risk.
- Who approves what? Brand, legal and regional owners need to be named in the workflow in the tool, where the agent will actually route work.
- What is off limits? Regulated claims, prices, legal pages and anything in a market with its own compliance rules.
- How will you know what changed and why? Use the audit trail, and decide who reviews it and how often.
Mature Sitecore teams already run this kind of content governance. The difference is volume: when agents propose, build and deliver, the approval step becomes the bottleneck, so it needs to be designed for that.
If you are still on Sitecore XP
This is the practical question for many of the teams we support. Marketing IQ needs SitecoreAI CMS or DAM. An XP licence on its own does not include it.
Timing matters too. Sitecore XP and XM 10.0 and 10.1 leave Extended Support on 31 December 2026. After that, Sitecore ships no further security fixes for those versions. So many XP teams now have two decisions on the table at once: what to do about support, and whether they want the agentic features that only exist on SitecoreAI.
As we see it, you have three realistic paths:
- Stay on your current version for now and reduce the risk with compensating controls (monitoring, hardening, WAF rules). This buys time, and you get no Marketing IQ.
- Upgrade to XP 10.4 or 10.5 to get back into Mainstream support. You keep your current setup and still get no Marketing IQ.
- Move to SitecoreAI, site by site, so marketing keeps publishing during the move. This is the path that gives you Marketing IQ. Sitecore also announced expanded Sitecore Pathway support for migrations from XP this year.
Whichever path you choose, the groundwork for agents is the same and can start now on XP: clean up and consolidate your brand guidelines and product facts, fix the content structure and markup that AI crawlers struggle with, and agree on your approval rules. Teams that do this first will get useful output from Marketing IQ much sooner after they move.
What to do this quarter
- Check which Sitecore version and licence each of your sites runs, and its support end date.
- Measure how your brand appears in AI answers for your top ten buyer questions.
- Write down the four governance answers above.
- Decide between staying, upgrading and moving to SitecoreAI, with cost and risk next to each.
If you want help with step four, our Sitecore Support Readiness Check compares the three paths for your instance, with cost, risk and a dated plan, using read-only access. It includes what you would need before AI agents work in your CMS.
Related reading
