SitecoreAI™ is the most capable DXP platform I have worked with in 20+ years of digital delivery. The personalization depth, the content intelligence, the agentic workflows are all genuinely exciting and the potential for organizations that have made the move is significant.
But potential and delivery are different things. And the gap between them, in my experience, is almost never the platform.
Over five workshops in Sydney in early March 2026, I worked with the global digital and marketing team at Colliers International to build the strategic and structural foundations for their migration from Sitecore XP to SitecoreAI™. Here is what that preparation looked like, and why it’s important to any organization preparing to make a similar move.
Start with the experience you want to deliver
Every migration begins with a platform decision. The smarter ones begin earlier than that.
Before Colliers made a single decision about UX, content structure, or migration sequencing, they asked a more fundamental question: what should this platform actually do? Not just at launch, but in 18 months, when the optimization is running and the AI features are doing their job.
A pre-workshop survey of approximately 30 Colliers digital and marketing leaders gave the program a clear mandate:
76%
named SEO and LLM visibility as a top 18-month priority
56%
named generating more qualified leads as top priority
64%
identified overall customer experience as their biggest friction point
#1
most critical factor in DX success: a clear and shared vision
None of those priorities are solved by a platform migration alone. They are solved by designing the right experience, structuring the right content, and connecting the right data, and then giving a capable platform the foundation it needs to perform.
By the end of the first session, Colliers had a shared North Star: a single, agreed statement of direction that gave every subsequent decision a common reference point.
We knew where we wanted to end up but we were not sure of the right path to get us there. Working through the strategy sessions brought the whole team across regions and functions together to align on a north star. That alignment is what makes everything that follows move faster and with less rework.
Define success before you build
SitecoreAI gives teams the tools to optimize, personalize, and measure with real precision. But those tools only deliver meaningful value when you have clearly defined the metrics you are trying to move. Not just traffic and engagement, but the commercial indicators that leadership and finance recognize: lead quality, conversion rate, organic visibility, content efficiency, and editor confidence.
The second session translated Colliers' experience strategy into measurable outcomes by audience and business objective. We agreed on three audience groups — customers, users, and machines — and ran an OKR exercise for each. We assigned objectives and key results that are measurable, meaningful, and reportable in language the C-suite recognizes immediately.
This work also surfaced which data connections needed to be in place at launch, not retrofitted afterwards, meaning the platform architecture could be designed to support measurement from day one.
Plan for AI and optimization from the start, not after launch
SitecoreAI's agentic capabilities are genuinely powerful, but they are not independent of the foundations you bring to them. Structure, taxonomy, and connected data are what turn platform capability into delivered value.
Agentic Studio can automate the assembly and personalization of page experiences at scale, connecting the right content to the right audience without manual intervention for every variant. The Sitecore.com team recently used this to run a targeted ABM campaign, assembling nearly 100 personalized page experiences automatically from structured content. This is only operationally viable when the content model is designed to support it.
The third session mapped which audiences and journeys warranted optimization investment, what data signals were needed, and how content should be structured for contextual delivery by agentic workflows. Those are not questions you can answer after the platform is live. They are decisions that shape what you build.
Map your data architecture and address AI governance upfront
Nervousness about AI, data privacy, and how structured content will be consumed by AI systems is entirely reasonable and it’s one of the most common concerns I hear from senior leaders preparing for a migration of this kind. The right response is to build the answers into the architecture from the start.
SitecoreAI's Model Context Protocol (MCP) exposes your structured content and entity relationships to AI systems in a controlled, deliberate way. You’re not leaving it to AI to guess what your content means or who it's for. You’re providing that context explicitly, within a governed taxonomy and schema framework that you control.
The fourth session mapped the current and future-state data architecture, identifying where data existed, where connections needed to be built, and how the MCP opportunity could be realized cleanly.
The nervousness about AI and content exposure is real in our organization, as it is everywhere. What changed for us was moving from a vague anxiety about it to a specific answer: here’s the framework, here’s what it governs, here’s what we control. That shift from uncertainty to structure is what the data architecture work actually delivered.
Structure your content before you migrate it
Migration risk is real, but it’s rarely where organizations expect it to be. The platform risk in a well-managed SitecoreAI migration is manageable. The content risk is where programs most often encounter the unexpected.
A migration that carries poorly structured, inconsistently tagged, metadata-incomplete content into a new platform delays the moment when AI, search, and personalization features can do their job. Migrating poorly structured content and fixing it on a live platform is far more expensive than fixing it upfront and moving it programmatically. Not to mention the embarrassment of needing to do rework after you’ve already done a rebuild!
Colliers committed to the content audit and taxonomy work before the build: establishing a global taxonomy, rationalizing content models, and cleaning metadata at scale. That taxonomy will underpin search, DAM, campaigns, analytics, and AI content generation. It’s infrastructure rather than housekeeping.
We could have pushed the taxonomy and schema work to phase two. A lot of programs do. What made us commit to doing it first was understanding that these weren't just migration tasks, they were the foundations that search, personalization, and AI all depend on. We'll see the benefit before we've migrated a single page.
What Colliers gets from doing this right
Collaborative intelligence at its best
Colliers are not a cautionary tale. They’re a model for how to approach a migration of this scale and complexity. But the more useful question is not what they did. It’s what the organizations that skip this work discover later.
The platform risk in a SitecoreAI migration is manageable. The content risk, the data risk, the taxonomy debt that arrives with you on day one, those are where programs stall, and where the distance between a platform that launches and a platform that delivers is quietly decided.
The window to do this work is before the build. Not at launch. Not in the first optimization sprint. Before.
Colliers invested in getting that right. The question worth asking is whether your program has.