What's New in Microsoft Dynamics 365 Customer Insights — 2026 Release Wave 1
Jack Amin
Digital Marketing & AI Specialist

Quick Answer
Microsoft's 2026 Release Wave 1 for Customer Insights (rolling out April–September 2026) is genuinely useful for working marketers, not just buzzword-filled. On the Journeys side, the headline additions are Copilot-powered conversational SMS, message expiration dates, automatic content refresh on queued emails, and journey actions that can create CRM records (lead-to-sales handoff, finally done properly). On the Data side, you get native Microsoft Fabric OneLake as a source, smarter audience targeting based on past campaign engagement, and agent grounding on unified customer profiles for Copilot Studio agents. Most of this is incremental polish of real pain points — the conversational SMS and record-creation features are the two that change workflows. If you're running CI-J today, your May and July upgrade windows are the ones to plan around.
Why care about another Microsoft release wave?
Because unlike most Dynamics release waves, this one fixes things marketers have been working around for a year, rather than bolting on more AI that nobody asked for. Customer Insights – Journeys is being framed by Microsoft around end-to-end, agentic customer engagement across sales, marketing, and service — the marketing is predictable — but the actual features shipping are practical.
If you're using CI-J day-to-day (I use it at Nexacu for upsell/cross-sell journeys to training graduates), this wave is worth paying attention to because a few of these changes directly reduce how much manual scaffolding you've been doing around the platform's limitations.
Here's the honest walk-through.
What's changing in Customer Insights – Journeys?
Microsoft's official release plan covers features planned to ship between April 2026 and September 2026, split across three themes: Agents and Copilot, Elevate customer experiences, and Turbocharge your pipeline.
Translated into plain English, there are five features worth your attention.
1. Copilot-powered conversational text messages — GA April 2026
The headline feature, and the one that actually changes what you can do. Until now, SMS in CI-J has been broadcast-style: you send, the recipient maybe clicks a link, the conversation ends. This feature lets Copilot handle back-and-forth SMS conversations inside a journey — acknowledging replies, routing based on intent, and keeping the thread alive without a human picking it up.
Why it matters: SMS reply rates are 4–6x email open rates for most Australian SMBs, but the operational cost of actually replying has kept most of us using SMS only for one-way reminders. If Copilot can handle the basic "yes / no / reschedule / more info" branches without a human, SMS becomes genuinely viable as a conversion channel, not just a notification channel.
What to watch out for: Conversational AI on SMS has a narrow band between "helpful" and "uncanny." Expect to spend real time tuning the Copilot prompts and fallback paths before you let this loose on real customers. Do a soft pilot with your own staff first.
2. Set message expirations to keep communication relevant — GA May 2026
Public preview since March, general availability May 2026. A small but overdue feature: you can now set an expiration date on a queued email or SMS, so that if the message hasn't sent by then (because a journey is taking longer than expected, or the recipient only just qualified), it gets dropped instead of delivered with stale content.
Why it matters: Anyone who's run a seasonal campaign in CI-J knows the pain of someone entering a journey in late December and getting a "Christmas sale" email in February. This fixes that. It's a quality-of-life feature, not a transformation, but you'll use it constantly.
3. Automatically update emails with the latest content — Preview April 2026, GA July 2026
Right now, once you build an email in CI-J and queue it in a journey, the content is effectively frozen. If you update the underlying template — fix a typo, swap a CTA, update pricing — the in-flight journey keeps sending the old version. This feature lets the platform re-sync emails to pull the latest template content before sending.
Why it matters: Big deal for long-running journeys (onboarding flows, multi-month nurture sequences). It means you can maintain one canonical template instead of duplicating and versioning. Watch this one closely — the GA version will be the one to trust, previews of features this plumbing-heavy often have edge cases.
4. Transform customer journeys into action with record creation — Preview April 2026
This is the one that quietly reshapes how CI-J integrates with Sales. Until now, getting a journey to create a lead, contact, or custom CRM record required a Power Automate detour or custom plugin. This feature lets journey actions create records natively — so when a contact hits a qualifying behaviour, you can just… create the lead. From inside the journey. No workaround.
Why it matters: If you've ever built the "marketing qualified lead handoff" flow in CI-J, you know it's been the single biggest source of fragility in most deployments. This is Microsoft finally closing the gap. Worth planning your handoff workflows around this when it hits GA.
5. Boost customer confidence with branded content links — Preview Sep 2026, GA Dec 2026
Further out — public preview in September, GA not until December — but flagging because it matters for email deliverability and click-through rates. Instead of your tracked links resolving through a generic Microsoft tracking domain, you'll be able to use a branded subdomain (e.g. links.yourbusiness.com.au). This is table stakes in other marketing platforms and it's good to see it coming to CI-J.
Why it matters: Click-through rates measurably drop when links look like spammy tracking URLs. Branded links typically lift CTRs 5–15% and reduce the chance your emails get flagged by filters. If you're in a regulated industry (finance, legal, health) where deliverability really matters, this one's worth the wait.
What's changing in Customer Insights – Data?
Smaller but strategically important wave for CI-D. Three features worth knowing about.
1. Microsoft Fabric OneLake as a data source — Preview May 2026, GA July 2026
If your organisation has invested in Microsoft Fabric (as many larger Australian businesses are doing via their Microsoft EA), this is significant. You'll be able to use OneLake as a native data source for CI-D, which means you can stop building and maintaining the middle-mile pipelines that move data from your lakehouse into Customer Insights. It's also a strong signal about where Microsoft is heading — Fabric as the unifying data layer under everything.
Practical take: If you're on Fabric, this simplifies your architecture and lowers the cost of keeping CI-D data fresh. If you're not on Fabric, nothing changes for you this wave.
2. Target audiences using signals from past campaigns — Preview April 2026, GA July 2026
CI-D will start using engagement signals (opens, clicks, conversions) from earlier campaigns as inputs when you're building new segments. So instead of "everyone who bought in the last 90 days," you can build "everyone who bought in the last 90 days AND engaged with our last two email campaigns but not the most recent one" — without manually stitching that together.
Practical take: For mid-sized marketing teams running 10+ campaigns a month, this closes a real analytics loop. For smaller teams running one or two campaigns a month, the effect is modest — you already know who's engaged.
3. Grounding AI agents in unified customer profiles — Preview June 2026
This is the strategic play. Microsoft wants the AI agents you build in Copilot Studio to reason over the unified customer profile in CI-D, not just whatever table you point them at. In practice this means an agent (for example, an AI SDR built in Copilot Studio) can pull the complete known history of a contact when deciding how to respond, rather than working off fragments. ServerSys notes that Customer Insights is getting a new connector that lets AI agents in Copilot Studio tap directly into unified customer data, which is the practical manifestation of this.
Practical take: Very powerful if you're building agents. Mostly irrelevant if you're not (yet).
What does all this mean if you're using CI-J today?
A few genuinely practical takeaways:
| If you're... | ...this wave changes |
|---|---|
| Running nurture journeys longer than a few weeks | Content refresh on queued emails (July GA) removes a real maintenance pain |
| Doing lead-to-sales handoff | Native record creation (in preview now) lets you retire Power Automate workarounds |
| Using SMS as a channel at all | Conversational SMS (April GA) fundamentally changes what SMS can do in a journey |
| Running seasonal or time-sensitive offers | Message expiration (May GA) stops the "I got a Christmas email in February" problem |
| On Fabric | OneLake data source (July GA) lets you retire bespoke pipelines into CI-D |
| Building Copilot agents | Unified profile grounding (June preview) makes agent responses measurably better |
| Worried about email deliverability | Branded links (December GA) is the one to plan for Q1 next financial year |
What's notably not in this wave?
Being honest, because most vendor release announcements never tell you what's missing:
- No major WhatsApp or RCS improvements. If you've been waiting for richer messaging channels, this wave doesn't deliver.
- No obvious improvements to the journey designer UX. The builder still has the same friction points it did a year ago.
- A/B testing capabilities remain thin compared to dedicated ESPs. If serious email optimisation is your world, you're still better off with a specialist tool for that workload.
- Reporting and analytics are again incremental. If you've been frustrated that CI-J's native reporting doesn't match what you can do in a proper BI tool, this wave doesn't change that. Most mature teams still pipe CI-J data into Power BI for real analysis.
- No material changes to pricing or licensing model. Which is either good or bad news depending on your renewal.
Should you enable the public preview features?
Depends on what it is. My rough framework:
- Preview features that affect core delivery (content refresh, branded links, OneLake sources): Wait for GA. A broken send in production is expensive and these features touch the send pipeline.
- Preview features that are additive (record creation in journeys, message expiration): Worth enabling in a sandbox immediately, and in a limited production journey once you've tested the shape of it.
- Anything Copilot-powered (conversational SMS, agent grounding): Enable in preview because the behaviour is genuinely new and you need hands-on reps to understand where the edges are. But don't let it drive customer-facing interactions until you've put it through its paces.
The general rule with Microsoft releases: the documented feature almost always works. The edge cases don't always work. Pilot narrow, roll out broad.
The bottom line
This isn't a transformative wave for Customer Insights, and that's actually the good news. After several releases that piled on AI features, Wave 1 of 2026 is mostly Microsoft cleaning up known rough edges — message expiration, content refresh, native record creation, branded links. These are the "finally" features. The Copilot additions (conversational SMS, agent grounding) are the ones worth experimenting with, but the bulk of the value in this wave comes from the boring-sounding fixes.
If you're running a Dynamics 365 Customer Insights deployment and want a second pair of eyes on how to roll these features into your existing journeys — or you're weighing up CI-J vs a specialist ESP for your next financial year — get in touch. It's the kind of practitioner call we have with Australian businesses on the Microsoft stack every week.
Frequently Asked Questions
Let's discuss your project
Need help upgrading to CI-J?


