On site at ServiceNow Knowledge: ServiceNow puts its AI leader in charge of the channel and the partner model is changing to match

Robert Dutt: Hello and welcome to In The Channel from ChannelBuzz.ca, bringing news and information to the Canadian IT channel community for the last sixteen years. I’m Robert Dutt, editor of ChannelBuzz.ca and your host for the show.

I’m recording this from Las Vegas, where I’ve spent the last few days at ServiceNow’s Knowledge 2026 conference. The big theme of the week has been what ServiceNow is calling the Agentic Business – the idea that we’ve moved past the AI pilot era and into a world where autonomous AI agents are doing real work, governed by a platform layer that sits above everything else running in your enterprise.

That governance layer – ServiceNow’s AI Control Tower – is central to how the company is repositioning itself. And it raises a real question for the channel. If ServiceNow is now the arbiter of how AI gets governed and orchestrated across the enterprise, what does that mean for partners? Who do they want to work with? How does the delivery model change? And how do you build a services business in a world where the traditional implementation and configuration work is being compressed by AI itself?

My guest today is the person who has to answer those questions for a living. Michael Park is the senior vice president of global partnerships and channels at ServiceNow. What makes his appointment interesting is that he didn’t come up through channel sales or alliances. He most recently led AI go-to-market strategy for ServiceNow before being tapped to run the partner organization. It’s a deliberate choice, and it says something about how the company is thinking about what the channel needs to be right now.

Let’s get right into it – my chat with Michael Park.

Michael, thanks for taking the time. I appreciate it.

Michael Park: Great to be here.

Robert Dutt: You have an interesting background coming into leading the partner organization, coming from the product strategy side – and particularly given where we sit, with the topic du jour being AI product strategy. Was that an intentional path on ServiceNow’s part? And if it was, what does it say about how the organization at large is thinking about where the partner organization needs to be right now?

Michael Park: Yeah, I think it was intentional by Bill and the management team. The AI stuff is relatively new. And having the opportunity to lead AI go-to-market from the beginning, and helping the organization as a whole go from zero to where we are now with it – when you’re first in, you learn a lot. The good and the bad. And I think that first year in the job, I did over 500 actual customer meetings, just explaining what AI is, how to get to value with it, what the platform is.

And so I think when we started looking at where the market’s going, we realized that ServiceNow is on a wonderful trajectory for where we’ve come over the last twenty years. But for the next generation, what we need to do is take everything we know and figure out how we build for scale to enable the partners to execute with even greater scale. And so a big part of that was taking what I know, moving it over to the channel, and then driving the channel.

And so as you heard yesterday from the mantra of our whole operating strategy – we must be first customer obsessed, and we do it in a way that we’re AI led. But then we deliver it in a way that we’re partner empowered. And so it’s a very simple mantra that keeps the GPC, the Global Partner and Channel organization, grounded in what we’re here to do.

Robert Dutt: A lot of channel chiefs tend to come up through sales, maybe through alliances, occasionally through marketing. When you’re coming in from the lens that you have – having led AI product – what do you see that you think a traditional channel chief might not? Where are the blind spots, since that was one of the words of the day on the main stage earlier, in how the industry has been thinking about how partner programs and partner value are structured?

Michael Park: Well, if you go back all the way through my resume, I’ve been in and out of product, marketing, and sales jobs. And so my special skill is that I’ve been able to sit at the intersection of customer demand, the business model, and technology innovation – and I can translate between those three.

And so really the only difference with AI right now is the pattern for thirty years has really been the same as new technologies come in. It’s just happening a lot faster. The floor of adoption has become easier because of the way you can administer an LLM, which has raised the ceiling on what the art of the possible is. And so I think what I would say is it’s not so much about the skill set of a channel leader or a product leader. It’s the ability to think in an agile way, to really free your mind from a locked-in pattern – to understand the need, the technology, and the business model, and say, hey, is there a better way to create a new outcome here? And that’s where I’m seeing the real great leaders emerge.

I also think the other thing about AI is that the change management required to administer value creation at this kind of speed requires real conviction and the ability to really understand change management. Because the tech is the tech – but changing the behavior, the mindset of somebody, while there’s also the looming threat of job insecurity, everything else going on in this world, things changing so fast for a lot of people that aren’t used to changing fast – it’s as much about the leadership mindset of change management as it is the adoption of the new technology.

Robert Dutt: With AI Control Tower, it seems like there’s a real effort to position ServiceNow as the governance layer – even more so than before – sitting above everything else running in an enterprise. Does that change what you need from partners, who you partner with, or how you deepen relationships as you build out the orchestration-plus-governance story?

Michael Park: I think the channel is always evolving to meet the needs of the market. And what is happening now is you’ve seen this enormous surge of token consumption happen in the last two years. Now, whether you’re getting the return on investment for all those tokens you’re paying for is another question altogether. But the other constant we see in the enterprise is that there’s not one platform. There are multiple platforms people are using, multiple elements, multiple agents being built. And one of the inconsistencies there is – do you have a consistent way of setting thresholds on token consumption across platforms in a unified way? Do you have a way to administer compliance and security or risk management protocols over MCP servers that different groups are building on different platforms? Can you administer a regulatory or company-specific compliance policy across multiple agents on multiple platforms? And today’s answer is no, no, no, no.

So we see this incredible opportunity – because, Rob, in many ways we’re already in the game. We have this product called IRM, integrated risk management, where we do cross-company compliance and risk management and security modeling for many industries – banking, healthcare, manufacturing, etc. And so this is just a natural-order expansion of where we’re coming from in the context of integrated risk management, and even IT services management, because at the end of the day that’s a catalog of IT assets being managed, with life cycles being managed in relation to employees.

So AI Control Tower is just the next evolution, saying: regardless of what the asset is in the AI world, we will tag it, we will track it, and we will administer policy, compliance, performance, and risk management over it – the same way we’ve done for IT, the same way we do for integrated risk.

So those partners that have been with us on that journey, it’s a natural progression for them. But then with the acquisition of Veza and Armis, it takes us into an even deeper realm of security. With Veza, with identity – so that as every agent stands up and more non-human identities are set up, we have secure identity management over every one of those assets, easily administered as part of AI Control Tower. And then with Armis, the OT element – really being able to tag beyond just IT, any asset in the enterprise, and administer the same process with AI Control Tower – makes it very, very powerful in what we can do in a multi-vendor way. Like our game, as you heard from Bill today, is we’re friendly to all, because we can administer a common policy over anybody that wants to play with it. And that’s a slightly different approach right now.

Robert Dutt: Security and governance – such a big area, such a hot spot in the channel, and one that a really broad variety of partners play in. Do you see an opportunity to reach out to partners who maybe historically haven’t been in the ServiceNow sphere, as a result of going deeper into the security space?

Michael Park: Lots and lots of opportunity. If you look at the history of ServiceNow up until recently, most of our channel was activated as a mechanism to implement software – ServiceNow software that ServiceNow direct sales sold. But now that we’re surpassing the fifteen billion dollar mark and continuing to grow at twenty percent, the opportunity for us to scale has to be more leveraged through different partner ecosystems, many of which we’re not even tapped into yet.

So there’s going to be growth for the existing partners who continue to grow with us. And then as we get into new buying centers, there are going to be lots of partners already existing in the security domain who will be able to use Veza and Armis and the new AI Control Tower as a way to extend their security practice and build new practices on ServiceNow that they haven’t had before. And what I like about that is they have domain expertise in security that we don’t – but we have a platform and technology that they don’t. And the two make well together, just as much as one of our more traditional partners who really understands ServiceNow but is entering a new domain.

Robert Dutt: One of the big topics here at the event writ large has been getting past proof of concept – past “the board’s excited about AI, so we have to do AI” – to AI that actually proves its value in business outcomes. Can you tell me a little bit about what you’re doing in terms of enablement to help partners realize that opportunity and have the skill sets and tools they need?

Michael Park: Yeah, this goes to the operating strategy of what we call AI led. And for us, AI led starts with my own organization, GPC. When I took the job a year ago, one of the operating strategies I laid down is: if we cannot ourselves be AI fluent in the way we operate with our partners, we cannot expect our partners to be AI fluent. So we’ve been using the ServiceNow technology, we’ve been using Claude, we’ve been using Copilot, we’ve been using a couple of other vendors to basically operate the ServiceNow Global Partners and Channels group. The content we create, the policies we administer, the training we do – it’s all been agentified now.

In the last year, we’ve been able to identify about thirty-four percent of the redundant work that doesn’t add a lot of value. We’ve administered it away – either automated it or built AI agents to do that work. And we’re reorienting our people toward the more value-added work that is literally facing the partner, to help them drive business.

And if you go out and talk to partners, they will tell you – yeah, we’re kind of seeing that from ServiceNow. We expect that over time our partners will also pick that up.

So if you think about the opportunity for partners – all that we’ve done is also shaping into the enablement we’re building. For example, we have a hundred billion workflows today that are already operating, most of which are not yet agentified. So we don’t have to go build from scratch. We have to go agentify the workflows that are already running. It’s a huge opportunity for partners that we cannot possibly administer directly on our own without them.

One of the unique things about ServiceNow is that ninety percent of the deployments we do are actually done through partners. Only ten percent is direct. So the partner already plays an important role. But we want them to go beyond that – because in this new world of AI, as we talked about in the keynote, the installation, implementation, configuration, and upgrade work will get agentified in the next two years. So the services they’ve been driving for twenty or thirty years are going to get compressed into a smaller order effect.

But the new services we need require the skill sets we were just talking about – knowing how to use AI Control Tower, administering data graph connections, learning how to use multiple models of inference to plug in and call the ServiceNow workflow. These are all new value-added services that will help re-engineer a company. And it was also why I was reinforcing the value calculator assessment tool – you can’t just sell AI. You have to be able to articulate what the ROI is, what the benefit is, quantified from the telemetry of what you’re getting out of the platform.

We’ve delivered that baseline telemetry and asked partners to take it, learn it, and make it theirs. It’s not a completed product – we expect them to put their special sauce on it and then bring it to the customer. And then on top of that, we announced outcome-driven services, which is using that analytics baseline to drive into defining the business outcome and quantifying it – rather than just billing time and materials the old school way.

Robert Dutt: As that shift happens – as workflows become more autonomous, more agentic, and per-seat starts to feel like a legacy metric – how does that change how you think about partner incentives and compensation? Does the model shift toward outcome-based, consumption-based, something else? And how do you make sure partners can actually build a business around that transition?

Michael Park: The way we measure partner value contribution today is what we call sourced. The partner is sourcing value to ServiceNow – they’re bringing us a customer. Whether that’s licenses or consumption of AI, it’s still sourced. So the metrics we hold partners accountable to – sourced and adoption – don’t change. What changes is the speed.

What we expect now is that partners won’t source something that takes three years to deliver. We expect partners to source in a hundred days, deliver the first point of value in a hundred days, then do the next one in a hundred days, and then the next one and the next one. So in five hundred days you’ve had five points of sourced value creation and adoption – versus the old way, where you do one source point and three years later you come out with some kind of value. And so that I think is the new model.

And this is where ServiceNow’s platform is uniquely suited. If you go learn the ServiceNow platform, you can start in IT, move to HR, move to CRM, move to procurement. But the tools to build the agents, the tools to do the data connectors, the querying and reporting, the declarative modeling, the tools to call MCP – it’s the exact same across every different buying center. So the business model leverage for the partner is: once they’re trained up on the ServiceNow platform, they can administer and monetize any kind of workflow above it using the same skill set. The economics will be a higher operating cash flow margin per dollar of resource invested in the ServiceNow practice versus any other technology out there.

Robert Dutt: You opened the door with the words “hundred days” there. One of the big things you teed up on Monday, and then Bill talked about this morning, was the Go Live AI guarantee – the idea of a hundred days to real, active, measurable ROI on AI, as a guarantee. How does that work with partners? What’s the mechanic? Is there a co-delivery model, a financial backstop? And what does a partner need to do to carry that guarantee to a customer?

Michael Park: We had a number of partners step into the offer. We shared it with our most accredited partners first and asked them if they’d be interested. And we had quite a few step in.

We will prime. So the first way out, ServiceNow will prime that particular service. It’s a paid service the customer pays for. Upon delivery of the hundred days of service is when the last leg of that service bill comes. The way the model will work with a partner is the partner will be sub-primed into the prime model for some period of time, until we can get the operating model strong. Then as we do that – as we get the tooling and the procedures right – we’ll syndicate that out to the partner network so they can do it on their own and actually prime on their own.

But the first way, we need to prime just to get it right. The key is making sure we’re taking that learning and thinking about scaling and syndicating the model so that it’s not taking services from the partner – it’s actually replacing the proof of concept. Because typically a proof of concept takes about a hundred days. But if the platform’s already in, on ServiceNow, and the workflow’s already running, and all we have to do is activate the AI agent – which as you saw on stage is already built – then the risk is just turning it on out of the box. So the precondition of this service is that you have to deploy the out-of-the-box stuff we built, because we’re confident and we’ve seen enough that we know we can get the partner and the customer to value under a hundred days.

And then the beauty is: once you’ve got that point of value in the hundred days and you’ve proven it, the next projects come online very quickly. What we believe will happen is the hundred days leads to an expansion of book of business for the partners, because once that’s in, the customer will want to start more projects.

Robert Dutt: By nature, you’re saying that over time this is going to trickle down to situations where partners are prime. I’m assuming this is an ongoing thing – more program than promo, it sounds like.

Michael Park: Yeah – it’s a program, it’s not a promo. The promo was the Control Tower offer. That was the promo, direct to customer. But the prime offer is going to be there for a while. The last meeting I was just in, a partner said, “we’re all in because we’re already doing it – we’re going to run this in parallel to you guys. But what we like about it is if you’re going to run a program around it and create that brand and your sellers are going to activate it, we will come in behind you.” So I think the smarter partners are already on board starting to do it that way. Because the key is we can easily get this into a customer who is already up and running with ServiceNow on the core workflows – all we’re doing is activating the agentic workflow on top of it.

Robert Dutt: Action Fabric was another one of the big announcements – opening the platform to external AI models via MCP. It’s interesting – you’ve got that opening motion going one way with Action Fabric, and then the opening going the other way with AI Control Tower keeping an eye on things from above.

Michael Park: Correct.

Robert Dutt: Specifically around the MCP side – are you seeing partners start to build proprietary agent skills or vertical IP on top of ServiceNow that they’re bringing to market as their own? And is that something you’re actively encouraging as a route to market?

Michael Park: It’s actually called Action Fabric – I think Bill described it as Agent Fabric, but regardless of what you call it, it is basically calling the full power of the automation runtime of ServiceNow.

What I think will happen is lots of different kinds of partners will be able to use their domain expertise in a particular industry, geography, or segment, identify a problem to be solved, use an LLM to gain inference off the data sets that feed it, and then very easily call the ServiceNow runtime to deliver the workflow. And hopefully AI Control Tower will sit on top of that and administer all the other AI components that might be feeding other workflows around it.

The key is to set up for flexibility in different patterns. Some people will come in through an MCP server because they’re building something on the outside. Others will choose to use ServiceNow’s build-agent capability and build inside the ServiceNow platform. Others will basically say, I’m already building on five other agent platforms and I’m just going to put AI Control Tower on top to administer common control. The design point is to create flexibility for the customer – to give them the options they need without slowing them down or forcing them into a particular angle.

The flip side is: if we just say “here’s ServiceNow, build whatever MCP server you want” without Action Fabric, that’s going to create all kinds of problems. Everyone will create MCP servers against ServiceNow that aren’t properly administered, there’ll be performance issues, and then the question becomes, “that ServiceNow stuff’s not working” – and we’ll be saying, yeah, because you built an MCP server the wrong way. So part of this is about setting patterns that can be replicated with high security, high scale, and high repeatability, by either the customer or the partner, in safe, secure, high-performance ways.

Robert Dutt: Last one for me – a partner comes to you and says, I want to be one of your top AI delivery partners in the next couple of years. What do you need them to build, to be, to do, to have, that maybe they don’t today?

Michael Park: We have a certification path to all of the different kinds of skills a partner may choose to be in. And we just introduced our new SKU structure for ServiceNow products. In the past, you actually had to buy the highest tier SKU to get AI. What changed just a few weeks ago is that even in the base tier product, AI is naturally embedded. So AI – and certification – becomes a consequence of every single product we sell now.

What we’ve done in getting partners ready for that is: the AI certification used to stand separately. You had to go get it. Now it’s natively built into every product they’re getting activated on. And the beauty is that our products are all built from the same platform – so once you learn the AI capability natively in IT, it’s the same capability in HR, in procurement, in ERP. That makes getting the partner ecosystem up to speed technically much, much easier.

Robert Dutt: Michael, I’m sure it’s a busy week and a half for you, but I appreciate your taking the time.

Michael Park: Happy to. Thanks for the time.

Robert Dutt: There you have it – Michael Park, senior vice president of global partnerships and channels from ServiceNow, live at Knowledge 2026 in Las Vegas.

I’d like to thank Michael for his time, especially in the middle of what is clearly a very full week for the ServiceNow team.

And thank you for listening. A few things I’d pull out of this conversation as worth sitting with.

First, on the Go Live AI guarantee – Michael was pretty explicit that this is a program, not a promotion. The current model has ServiceNow as prime, with partners sub-primed into the delivery. But the stated intent is to syndicate that model out so partners can eventually carry it themselves. If you’re a ServiceNow partner and you’re not already thinking about how your practice gets certified to prime a hundred-day engagement, that’s a conversation worth starting now rather than later.

Second, the services compression point is real and worth taking seriously. Michael said it plainly – implementation, configuration, and upgrade work is going to get compressed in the next two years. The partners who come out ahead are going to be the ones who’ve already built the new competencies: AI Control Tower administration, Action Fabric and MCP integration, outcome-driven services built on telemetry. Those are the new billable skills.

And third, I found the economics argument compelling. The platform leverage point – that the same skill set applies across IT, HR, CRM, procurement, and every other buying center on the ServiceNow platform – is a real differentiator for partners who go deep on ServiceNow versus spreading across multiple vendors. Five sourced value points in five hundred days versus one in three years is a different kind of business.

More from Knowledge 2026 coming later this week, including the Canadian and GSI perspective on what all these announcements actually mean for the local market back home.

If you’re finding In The Channel useful, we’d love for you to follow or subscribe wherever you’re listening. We’re on Apple Podcasts, Spotify, YouTube, and most major directories. Ratings and reviews are always appreciated and always read.

Until next time, I’m Robert Dutt for ChannelBuzz.ca, and I’ll see you in the channel.