Community Summit NA 2026 brought the Dynamics 365 community to Music City for a week of sessions, demos, and hallway convos (to no one’s surprise, most of them centering around AI). The Coffee + Dunn team spent the week in sessions (including two of our own), talking with the marketers, admins, and leaders who run the platforms or are evaluating whether to buy it.
TL;DR: AI agents like the Journey Creation Agent, which Dynamics 365 VP Mike Morton highlighted during the CE keynote are promising, but we’re still judging them on how they perform in client environments, not on stage. The way marketers interact with Customer Insights is starting to change in a good way, but clean, unified customer data is still the foundation for all of it.
1. Customer Insights sessions were few, but the audiences were engaged
There were a whopping eight sessions on the agenda that mentioned Customer Insights outright—an obviously small slice of a very large event. But the people in those rooms? These were folks running live journeys, working through implementation problems, or in some cases, evaluating the platform. They came with detailed questions (exactly what we love to see)—a sign that the CI community, while not always the most visible, is no doubt an active one.
2. Shiny objects are nice, but fundamentals still get the biggest reactions
I chatted with another speaker during a Women in Dynamics lunch who delivered a session that was so packed, she actually had to tack on a second session on the fly. The primary topic? Power Automate. (You thought it was agents, didn’t you?)
Some of the strongest reactions in our deliverability session came from what can feel like fairly foundational guidance. Even experienced users found capabilities and best practices that already existed in their environment.
Candidly, this is a lesson we keep relearning. It’s no secret that Customer Insights isn’t always first in line for the slickest features, but getting more out of CI often has nothing to do with the newest feature (or even the product directly). It comes from understanding deliverability, consent, journey design, data, and governance well enough to use what’s already at your fingertips.
3. Customer Insights wasn’t the loudest product, but the investment continues
Mike Morton led a general session on the future of customer engagement at Microsoft. Did it focus on Sales? You bet it did: automatic data capture, the Sales Development Agent, and Account 360, etc. Marketing got less stage time, but I’d be lying if I said wasn’t heartened to hear him call Customer Insights one of the quieter forces in the Dynamics lineup—and more importantly, an area Microsoft is continuing to invest in.
4. The Journey Creation Agent has real potential. Test it in your own environment
In Microsoft’s demo, you describe a campaign in plain language, like a birthday offer. The agent asks clarifying questions, finds the right segment and content, and builds a journey in minutes.
The truth is that this agent is still a little…half baked. No demo is going to show you the timeouts, etc, we’re still seeing. That said, we’re optimistic about where this goes. For now, though, I don’t think this changes our recommendation to clients: pilot it on a low-risk journey you know and compare the output to what your team would build before anything goes live.
5. The bigger shift is how you work with Customer Insights
The most recent development that stuck with me most wasn’t another in-app feature. It was CI capabilities showing up via plug-in in Copilot Cowork, a callout that got surprisingly little attention at the event.
The idea that marketers don’t always need to open the Customer Insights app to use its capabilities and could work through a conversation instead feels increasingly more aligned with the marketer’s modern experience outside the Microsoft ecosystem, and something the keynote focused on as it related to Sales; why we didn’t give the marketing plug-in the same love feels a little like a missed opportunity.
6. Match the type of AI to the problem
One of the most useful AI sessions skipped the hype and offered a decision framework, built on Microsoft’s progression from Search à Assist à Recommend à Automate à Orchestrate. Each kind of tool fits a different job:
-
- Copilot (based on prompts) for judgment-heavy, variable, or one-off work
- Rule-based automation (e.g., Power Automate) for structured, repeatable processes
- Agents for multi-step, cross-system workflows with clear permissions and approvals
Before adopting any tool (double when it comes to AI) we ask: Does it solve a meaningful problem? Is it a fringe case, or will it work for the org? What happens if the AI tool gets it wrong? Just like automation, AI won’t fix a process no one’s bothered to define.
7. Data quality is still the foundation, and AI is starting to help
A session on customer data strategy focused on an all-too-familiar problem: the same customer showing up many times across many data sources, each with its own variation on name, etc. The session presenters showed an agent-based approach to matching and deduplication. As far as we could tell, it was a partner-built solution, not an out-of-the-box Customer Insights feature.
The idea still caught our attention. Much of the AI conversation centers on content and campaign creation, but some of the most valuable uses may be the less glamorous ones: matching records, catching duplicates, and keeping profiles accurate. For Customer Insights – Data users, unification and match rules need regular attention, not a one-time setup.
Looking beyond Summit
AI grabbed the headlines in Nashville, but practical value won the convos that mattered. We didn’t leave thinking AI makes Customer Insights fundamentals obsolete. We left feeling validated that our philosophy when it comes to the importance of strategy, clean data, and thoughtful implementation matter more than ever.
Want a second set of eyes on your Customer Insights setup, or weighing whether it’s the right fit? Let’s chat.
