D365CC – Wave 1 (2026), incoming waffle…

I usually hold off on the release wave write-ups for a while. If you’ve followed any of the previous ones, you’ll know the pattern: features get announced, GAs get pushed, previews linger longer than anyone expected, and by the time something lands in your tenant, it’s been through three rounds of changes. That said, Wave 1 2026 has a few things worth flagging early, particularly if you’re mid-deployment or planning something for later in the year. So here we go…

Also, as always, take the timelines with a healthy dose of ‘mayyyyybe’. “GA” in Microsoft-land and “GA” in the real world are two very different things. In my experience, try, try and keep trying until it breaks, and only then wait 6 months before making it prod!

Ben’s post thought of the day… (subject to change, like these updates 😆)

The agentic AI conversation is getting louder, and I’m having it with clients a lot at the moment. There’s a real split between those who get excited by the idea of AI handling interactions end-to-end and those (usually the ones who’ve been around a bit longer) who immediately ask “but what happens when it goes wrong?” Both are valid, and honestly, both need to be in the room. The tech is moving fast, faster than most businesses can adapt their processes to keep up.

Anyway… onto the stuff and waffle….

The Four…

Four areas are now the focus; they’re going to be the buzzwords, but they do sound pretty fun to get our hands on, and there’s no need to worry too much about change. It’s all about adding value, keyword!

• Agentic automation across self-service and assisted service, higher containment is the goal.
• Omnichannel intelligence for voice and messaging, with a focus on emerging channels.
• Supervisor experience – reimagined (fancy word), with richer real-time AI-driven insights.
• Extensibility – this one’s been around for a while; they keep using it.

Even I cringe at some of these terms, but they are accurate.

Copilot, COP Studio and Skynet (AI) – still coming in first…

No surprises. Copilot continues to be front and centre. For agents, it’s about speed: finding the right resource faster, drafting contextual replies in email and chat, and generating case and conversation summaries with one click. The aim is to cut handling time without cutting quality, which is the right drive, provided the underlying knowledge base is good (please start to build this early). That’s the bit that often gets glossed over in the demos. Garbage in, garbage out – something we should all be looking at!

The shift I’m watching is the move from Copilot-as-assistant to Copilot-as-agent, AI that doesn’t just help the human but actively manages the interaction on its own. That’s a meaningful change in how contact centres will operate, and it’s the kind of change that needs proper management, not just a Teams announcement and a hope for the best.

Standout: Enhanced RT Translation – nice to see this getting focus, opening up more agents without the language barriers!

 

IVR Control – It’s getting bigger

Conversational IVRs powered by Copilot Studio continue to offer more to experiment with. The pitch is containment, with more customers resolving their issues without speaking to a human agent (personally, I don’t think it’s fully there yet), which obviously reduces cost-per-contact and (in theory) improves CSAT. However, trying to adopt CSAT is very different in practice. Most customers now don’t want to give feedback, so using the framework to do this will open up a new form of data gathering.

In practice, this only works if the bot is well designed and the intents are properly mapped. I’ve seen plenty of poorly built bots that do the opposite, frustrate customers to the point that they’re hammering zero to get out. The capability is there; the discipline to build it sometimes isn’t.

Standout: Only the one here, but having more Custom Voice models is only a good thing, BUT I’m still a fan of recorded voice, albeit harder to manage it does bring that personal edge!

 

Omnichannel – Deeper, not wider! 🤨

The omnichannel layer already covers voice, live chat, SMS, Teams, and social. Wave 1 isn’t about adding more channels so much as deepening the intelligence across the ones already there, better conversation insights and greater consistency in the agent desktop experience, that sort of thing.

For those of us who’ve sat through the pain of trying to achieve a consistent agent experience across five different channels, this is genuinely welcome. I like the way I can reuse data and connections for this, piping multi channels into one agent can actually work (voice and text aside).

Standout: Consult, Transfer Securely – BIG one for me, we’re still handing off customers to payment services and there is some really good ones out there still, huge market, this gives us that scope for lockdown!

Standout: Data Masking – Another one that needed attention, control over mistakenly said information and agent controlled choice is huge!

 

Supervisors – We are getting more, but slowly.

I’m pleased to see this. The supervisor experience has historically been something of an afterthought compared with the agent-facing stuff (and that’s genuine feedback from my clients), and it’s usually the first thing a customer mentions in a review session.

Real-time, AI-driven insights, sentiment monitoring, staffing guidance, and the ability to intervene quickly, are things supervisors have been asking for since the dawn, and they’re in there, access and use are another story.

The extensibility piece is also handy here: custom metrics, custom entity reporting, bespoke dashboards. Useful for clients with non-standard workflows (which, in my experience, is most of them).

Although not cemented in the release notes there are musings of this becoming better!

 

Unified Routing – Still doing it’s thing

The routing engine continues to evolve, AI-based assignment across chat, digital messaging, and voice, with features like percent-based routing, overflow management, and preferred agent routing helping to optimise without manual queue babysitting.

Unified Routing has really stepped up over the past year or so and is genuinely solid now (well, they built on top of it!). It’s worth keeping an eye on how the agentic layer interacts with it as that matures. I really like the Intent and Work Classification logic on this. Don’t overcomplicate, but it’s pretty cool to mess with.

With the data inside the exposed APIs it’s nice to be able to pipe this into other systems, I still maintain dashboards with an external service is a great tool for customisation (I’m a huge fan of Grafana) so this is an nice addition.

Workforce Engagement Management (WEM)

WEM covers the full picture: scheduling and forecasting, performance management, quality monitoring, agent engagement, and training. Wave 1 continues to build this out. It’s one of those areas that often gets left to the end of a deployment conversation and then suddenly becomes urgent.

If you’re planning a D365CC rollout, then think about how it can help. When it first came in, it was a bit of a mess, but it’s now coming to be a bit more integrated with the data.

Wrapup!

As always, the full release plan is on Microsoft Learn and is worth a read if you want the granular detail: Dynamics 365 Contact Center 2026 Release Wave 1

Features roll out between April and September this year – maybe…. some are auto-enabled, others require action. Remember what I said at the top, have control!

Who knows, hopefully these will start to make an impact, but you need time to see if they are going to take hold well!

That was some writing power, I know it’s a lot but hopefully worth some value!

Ben