Author transcript
Rob Wilkinson: Hello and welcome. I'm Rob Wilkinson at CX Today, and today we're taking a closer look at the contact center that learns with your business. If you're challenged with fragmented customer data, disconnected service operations, or the familiar problem of AI looking impressive in a demo but being a lot harder to make useful in production, then stay with us. Today I'm joined by Darya Mazandarany, Corporate VP, Dynamics 365 Contact Center at Microsoft. Darya is well placed to talk about where contact center transformation is heading right now – not just in terms of channels and automation, but in terms of how service organizations can use governed data, AI and business context to improve how the work actually gets done. Welcome, Darya. Thanks for joining me.
Darya Mazandarany: Thanks for having me, Rob. It's a pleasure to be here.
Rob Wilkinson: Let's set a little context for the audience. Most contact center transformations are framed around channels consolidated or costs reduced. Microsoft's story right now starts somewhere completely different: a single data layer that lets a business learn from every interaction. Where did this ambition come from? What made a unified data layer feel like the right starting point right now, and why wasn't it the contact center platform itself?
Darya Mazandarany: At Microsoft, we felt that focusing on the customer experience and customer outcomes was the way to delight everyone and win our share of the market. So we looked at what our advantage would be. It's really the access to business collaboration data – Teams messages, meetings, documents on SharePoint, emails that go through Exchange and Outlook – and combining that with the business systems of record to create a data layer that can fuel the context and IQ for all of these workloads. That allowed us to create scenarios backed by AI, both human-to-AI and human-to-human with AI assistance, that move the business outcomes for a contact center in the right direction across the whole customer journey.
Rob Wilkinson: Is that a big ask, though? Trusting a single governed data layer – customer profiles, case histories, business processes – is a big thing to ask of an enterprise, especially when those systems have been kept deliberately separate for years, often carrying a lot of fragmentation and accumulated technical debt. It's a real pain for people to move away from. So how is Microsoft making that a credible ask, and what does it take for an organization to get to the point where it can achieve that shared context?
Darya Mazandarany: Great question. The first thing to note is that we understand the lack of homogeneous setups in these contact centers. They're very fragmented. They typically have multiple contact center systems, multiple CRMs, multiple systems of record. So we started with a complete but composable story, so we can meet every customer where they are in their journey and derive a lot of value.
One of the advantages we have is that most organizations already use Microsoft for their collaboration. The back office already lives in Teams, SharePoint and Exchange. Think about conversations like, "Should we give Contoso the extra 20% discount on this deal?" That doesn't happen in the CRM or in the contact center. It typically happens in meetings, emails, documents and Teams chats. We have access to that information as a starting point, which is a huge advantage for customers already using the Microsoft ecosystem.
We also construct our own context graph within the contact center system itself. The platform creates that data layer through every interaction and combines it with the rest of the Microsoft IQ ecosystem. That's how we build the system as we go and make every interaction better than the last one.
Rob Wilkinson: It's exciting. We're seeing lots of industry players move in this space with the advent of AI – lots of new offerings, lots of competition, CRM companies becoming different businesses. For years, the industry was framed as a CRM versus contact center battle, with customers having to pick a side. Now Microsoft seems to be arguing that's the wrong way round. Why do you think customers are now treating this as a platform decision rather than an application decision, and what's driving that?
Darya Mazandarany: There are a couple of factors at play. Customers are talking to us about a platform decision. It's not a CRM or CCaaS decision anymore. They're looking at a much broader problem and how to solve it. Typically, the conversation shows up as, "I'm looking for an AI platform," or "I'm looking for a customer experience platform." Many customers are experimenting with the frontier labs and with other competitors.
It comes back to our advantage: context fuels the AI. Our ability to assemble that context across heterogeneous systems, and to use back-office data to power front-office scenarios, is a huge advantage. That's why we've reframed it as a customer experience platform, with the data platform as the key decision point. And Microsoft has a history in compliance, security and governance. We've invested heavily in IP and technology to ensure your data is secure and that you're safe from a regulatory perspective, and that all feeds into the decision-making.
Rob Wilkinson: There's also a big appeal in not having clunky bolt-ons on top of bolt-ons. There's been a rush on those over the past 12 months, with new tools deployed and sold very quickly everywhere. So I don't think asking people to go down this route is a bad idea – I think it will be welcomed. It's probably still a scary step, though, unless you're going with a company like Microsoft that has that legacy as a safe pair of hands.
Let's say this vision fully lands. A lot of what defines today's contact center architecture might look outdated. Do you think that's the case? And which part of Microsoft's approach will prove most durable – the piece still standing years from now, regardless of how the technology around it changes?
Darya Mazandarany: I think we can take a couple of things for granted. People have been interacting over the base channels for decades. Telephony has been around a long time, and I don't see it going anywhere. In fact, I see a bit of a reversal. People were gravitating towards digital channels because the capabilities were superior and it was easier for customers – they didn't have to wait on hold. Now, with voice AI capabilities dramatically better, I don't have to wait on hold, and I can do it hands-free while I'm in the car. There are a lot of advantages to that. So the channels themselves, while new ones will come, will be quite durable.
But the data layer, the context and the intelligence on top of it are the key. Nirvana for every business is to truly understand the customer in their journey. What are they trying to do in the moment? What have they struggled with? What have they done to fulfil that intent? And how can I use that in the moment they're interacting with me to delight them?
If I go on a website, start searching for a product, then stop because I'm having an issue and end up coming through a support channel, then as well as resolving that quickly, having context about what I was just doing – and using that to improve my NPS or CSAT, whatever I'm measuring – is a huge advantage. Context and intelligence like that will be durable, because they let us push the business outcomes of the contact center and the wider business in a way that's very difficult without this technology.
Rob Wilkinson: That's so true. That's the holy grail: being able to measure every touchpoint of the customer's journey, act on it, and take it into account when you're interacting with a customer. It's what companies have tried to do with voice of the customer and the other programs we've had in the past. Now everyone understands we're very close to that holy grail. There are opportunities to capture every element of every interaction – both sides of the conversation, the agent's (whether human or AI) and the customer's – put it all in one place where you can interrogate it, and find out what's really going on. That's exciting.
On the flip side, no layer of any technology stack stays stable forever. Which part of today's contact center architecture – or Microsoft's own stack, because I don't think you get away with this either – has the shortest shelf life, and why?
Darya Mazandarany: I would say it's the models themselves. We've already seen proof that models, while incredibly powerful and amazing, have commoditized. Look at the pace at which new functionality is released, the competition between the frontier labs, and open-source and open-weight players showing incredible results without trillions of dollars of investment to produce frontier models. What's durable is the harness the models run in, and the ability to give them the right tools, intelligence and context. Models are going to come and go, but what's fuelling them is the key.
Rob Wilkinson: So true. I've never seen a pace like the one we're moving at today. It creates a lot of opportunities, but also a lot of potential risks, so I'm glad you've boxed some of that off today.
For someone considering this – perhaps looking to get started with a voice AI agent – much of this vision depends on a mature, unified data foundation. Migration can sound like a heavy lift for an organization that isn't there yet. What's the lowest-friction way to get started today, and what should they expect in terms of initial changes?
Darya Mazandarany: Like you said, our real-time voice agent is the place to start. Voice is one of the most challenging channels to operate in, and for most companies it's the predominant channel of engagement. So we started there as a way to get your journey started. It doesn't require you to lock into our contact center as a service platform or our CRM. You can attach it to your existing CCaaS or on-prem system, with whatever CRM you're running, and take the AI for a spin. See how we leverage a variety of model families, along with the context and intelligence I was talking about, to delight your customers. It's on the order of weeks to months to show business outcomes. It gets you hands-on with what I've been talking about, and then you can decide how far you want to go on that journey, and how quickly.
Rob Wilkinson: That's a very straightforward way to get started. Who would have thought, with all this clever technology, it would be that simple? Thank you, Darya. I appreciate the clarity, particularly on the difference between adding AI features to a contact center and building a system that learns from operations over time. That stands out to me.
What comes through is that this new wave of contact center modernization is no longer just a question of replacing one platform with another. It's bigger than that. It's whether organizations can create that trusted layer of context you talk about, because that's now the foundation. Data always has been, but even more so now. AI agents and business systems then work from the same understanding, and teams looking at AI in customer service can find the business value – not just generating insights, but making sure those insights change how things happen. Otherwise, it's just a nice new toy. Driving real change is the key thing here.
Darya Mazandarany: Spot on. Business outcomes, right?
Rob Wilkinson: Yes. And that's not always resolution, although some organizations focus only on that. It's important to get those outcomes. Before we wrap up, Darya, for anyone watching who wants to learn more about Microsoft Dynamics 365 Contact Center and how this approach works in practice, what would you suggest they do?
Darya Mazandarany: Go to the Dynamics 365 Contact Center landing page, start one of the trials, and get playing with the voice agent if you want. Then contact us or one of our trusted partners to get you going. It's an exciting journey. The outcomes are real, and the technology is in a place it's never been before. I couldn't be more excited to be in this space.
Rob Wilkinson: Fantastic. You can discover more articles, resources and videos like this one at cxtoday.com. That's all we have time for. Thanks again to Darya at Microsoft for joining us. I've been Rob Wilkinson – thanks very much for watching.
