
Artificial intelligence isn’t just changing the technologies Cisco partners deliver. It’s changing the kinds of businesses they can build.
That’s a central theme of this episode of In The Channel, recorded at Cisco Partner Summit 2026 in Vancouver, where Alex Pujols, Cisco’s vice president of sales solutions engineering, joins us to discuss the opportunities emerging as AI becomes increasingly embedded in enterprise infrastructure and operations.
Pujols identifies security as perhaps the most immediate opportunity. He discusses Cisco’s five-stage Vulnerability Operations Maturity Model, designed to help partners assess their capabilities and develop the increasingly sophisticated services customers require. He also offers his assessment of how quickly partners are progressing through those stages.
The conversation turns to infrastructure, including emerging edge AI use cases, before exploring how Cisco Cloud Control could simplify managed services delivery. Pujols confirms that Canada is among the next markets targeted for Cloud Control availability, although no specific date has been announced.
Another major theme is partner-developed intellectual property. Cisco’s AI Studio and Model Foundry aim to make it easier for partners to build AI agents and specialized solutions, potentially creating new opportunities for recurring revenue.
But if AI agents become easy to create, how do partners differentiate their offerings and protect their intellectual property? Pujols discusses the importance of customer trust, creativity and the commercial protections Cisco is developing for its marketplace.
Finally, Pujols reflects on how AI is transforming Cisco’s Partner Innovation Challenge, with AI-related submissions rising from fewer than 20 percent just a few years ago to more than 95 percent today.
It’s a look at how Cisco sees the channel evolving from selling technology to building solutions around it.
Podcast: Play in new window | Download
Subscribe: Apple Podcasts | Spotify | Amazon Music | Android | iHeartRadio | Youtube Music | RSS
Read Full Transcript
Robert Dutt: Hello, and welcome to In The Channel from ChannelBuzz.ca, bringing news and information to the Canadian IT channel for the last sixteen years. I’m Robert Dutt, editor of ChannelBuzz.ca and your host for the show.
We’re coming to you today from Cisco Partner Summit 2026 in Vancouver, where Cisco is laying out its vision for how partners can capitalize on the next wave of artificial intelligence.
And while there’s no shortage of technology announcements this week, one of the more interesting themes is how those technologies are changing the kinds of businesses Cisco partners can build.
We’re talking about partners moving into increasingly sophisticated security services, building and managing AI-powered infrastructure, and developing their own AI agents and applications.
Cisco is also making a significant push to lower the barriers to creating those solutions, potentially opening up opportunities that were previously limited to partners with substantial development resources.
But making technology easier doesn’t necessarily make it easier to build a differentiated business around it.
That’s one of the questions I put to Alex Pujols, Cisco’s senior vice president of sales solutions engineering in its global partner sales organization, when we sat down here Monday at Partner Summit.
We also discussed Cisco’s new Vulnerability Operations Maturity Model, which is helping partners assess their capabilities and develop more sophisticated security practices; the implications of Cisco Cloud Control for managed services providers; and the opportunities emerging around specialized AI models and edge computing.
And Alex shared some interesting observations about how quickly Cisco’s partner community is evolving, including a dramatic increase in the proportion of Partner Innovation Challenge submissions involving AI.
It’s a conversation about where Cisco sees the channel heading and what partners need to do to take advantage of the opportunities ahead.
Let’s get right into it, my chat with Alex Pujols.
Robert Dutt: Alex, thanks for taking the time. I appreciate it. I’m sure this is a busy week for you.
Alex Pujols: Happy to be here, of course.
Robert Dutt: You talked this morning about the three major shifts you’re seeing in the market: securing agentic AI, scaling infrastructure for AI, and bringing agentic AI into business operations.
From a partner’s perspective, which of those shifts do you see creating the most immediate opportunity today? Like, the right-now opportunity?
Alex Pujols: I think there’s an opportunity for all of them right now.
It’d be really hard to ignore the role that security is playing within AI, not even just for agentic operations, but just more broadly.
We also talked about this whole post-Mythos era that we’re in, and defending at machine speed, at machine scale.
And so I think, 100 percent, the agentic opportunity. If I’m a partner right now, and I’m looking at these shifts, and I have to invest in a single area, I would certainly look there because the demand that we’re seeing from customers is overwhelming.
The activity from a compliance and legislative perspective means there’s probably going to be even more opportunity around this moving forward.
And then, selfishly, I feel that we have an amazing portfolio to address the opportunity. And so I certainly would say the security side of it.
Robert Dutt: I have to imagine that’s accelerating, too. AI has been the priority for a while now, but over the last six months or so, drip, drip, drip — Mythos, [unclear], OpenAI — all of that is just adding to the awareness, the appetite, all of it.
Alex Pujols: A hundred percent.
And it’s interesting. It’s to the point where family members who know nothing about technology are asking me, “Is it over? Is the end near?”
And so, in between therapy, I also just do a little bit of education and kind of help them understand where things are.
But the point is, it’s mainstream. It’s mainstream.
And so Mythos did that. The news reports of AI autonomously attacking other AI and stealing secrets have people spooked.
But while I think, in my opinion, a lot of that is inflated, what’s not inflated is the opportunity more broadly around security.
And so I think it certainly is real, and it’s going to continue to grow.
Robert Dutt: You talked a bit today about the Vulnerability Operations Maturity Model, helping partners understand where they are today and what capabilities they need to develop.
Can you give me that kind of elevator pitch on that? Just the real quick walkthrough of it: how it works, and what it tells a partner about how and where they should be investing right now?
Alex Pujols: Yeah. So how the maturity model came about was initially a little bit before Mythos, but certainly as the news dropped, we really accelerated the conversations we were having with partners.
And in my conversations, I was just noticing that everyone was on a different journey, or they were at different stages of that journey.
And so a lot of partners, we’d say, “Hey, how can we help?”
They’d say, “Get out of the way. We’re rocking and rolling. We’re cooking. We’re good.”
Then we had a lot of partners that said, “I have no earthly idea how to have a conversation with my customer about this. I have no idea what capabilities I need to build out. I’m not sure how to build them out.”
And so, “I’m struggling with just understanding the gravity of the situation and how I should respond.”
As we saw so many different levels, we said, “Look, let’s put this into a model, and as part of that, we can help partners get from one stage to the next.”
And so the model has five stages.
You could say Level 1 is probably partners with the least amount of sophistication around their practice.
And then Level 5, which is our most advanced partners, have full-fledged, what we’re calling Vulnerability Operations practices.
And so they can manage customer vulnerabilities really at machine speed, instantly.
They have super-advanced capabilities, like being able to go out and proactively scan customer networks, almost leverage AI to create automated patching on the fly, even when the vendor hasn’t supplied patching on their end.
So, super-advanced capabilities. But ultimately, that’s not needed for every customer.
And so the idea isn’t that every partner has to develop that capability. They need to find out where on that maturity model their customers need them to be, and then build to that.
So if your customers need you to be at a three, and you’re at one, then you can move to Stage 2 and Stage 3.
From a vendor perspective, what we want to do is equip them with the skills and the capabilities and the tools, and ultimately the guidance, so that they understand clearly how to get from one level to the next and build out the practices their customers need.
Robert Dutt: Especially for those partners who are finding they’re at the Level 1, perhaps, and their customers are saying, “Maybe two or three would be more representative of what we need, what we want from a partner.”
Are there any common threads or common first steps that you find work best for helping them move up that?
Alex Pujols: Yeah, I think there’s a couple of things.
Number one, of course, even though we’re talking about this more in the context of agentic AI and post-Mythos and things like that, what is just the fundamental understanding of AI at the organization?
Generally, for our Level 1 partners, AI means that a few people have ChatGPT subscriptions or something.
And so it’s kind of tapping into, where are you? What are the specific skills and capabilities that you need to build? And then also, how do you apply them?
Because a lot of partners maybe work with a specific subset of customers, or customers focused on a particular vertical. And so there are different flavours of what those customers need.
Can you provide those base-level capabilities to those customers in that particular area? And if you can, then you can kind of graduate to the next steps.
We actually have a really cool tool that a partner can leverage and go in and basically self-assess to find out where they are.
So it’s like 20 questions. It sort of tabulates and gives them initial guidance.
Then they can go and leverage a lot of the documentation and collateral that we’ve put out, and they can follow some concrete steps, almost like a journey.
“These are the things that I need to do, or I need to develop, to build out a capability so that I’m Level 3 or Level 4.”
And then the big one is — and I heard this from a lot of partners that I would probably say are on the two or the one maturity level — “What does a five even look like?”
What does somebody like that look like? The entire business, the practice, the operation.
And so we provide guidance across that entire continuum. And then we also have things like workshops where we can engage with partners to help them more.
Ultimately, again, get to wherever their customers need them to be, not necessarily where any particular vendor wants them to be.
Robert Dutt: I know it’s obviously early in the game, but how are partners grading out on that one-to-five kind of scale? Is it about where you’d expect?
Alex Pujols: It’s moving quickly.
I would say, generally, from a Vulnerability Operations perspective, that type of practice was almost relegated to some of the largest partners in the world.
And so, yeah, there were a few fives out there, but there were very few.
Generally, if I averaged it out, the partner community was probably more at a two.
And I can say, just probably within the last half-year, I think that’s a strong three at this point, because the demand is so intense.
For example, one of the characteristics of a Level 4, Level 5 partner is they have to have a managed capability, because you have to be able to protect in real time.
And so there are a lot of partners that didn’t have that, that are saying, “In this new world, I actually have to become a managed services partner. There’s no other way for me to do that.”
And so, yeah, I would say initially two, definitely a strong three.
I think the average would probably settle at around a four, with just some very special partners that are servicing clients that have multinational, complex, regulatory-adjacent types of environments where they need to be a five.
Robert Dutt: Yeah. I wanted to ask how broadly you saw five being, and that sounds like it makes sense that the five, that very tippy top, remains the cream of the crop who’ve really invested heavily in there.
Alex Pujols: Absolutely.
I mean, you even take it a step further for some of the fives, and they’ll fully outsource a lot of that work.
So it’s interesting. There are so many transitions over the last couple of years.
It makes me think of COVID as well. After COVID, there was this strong move of, “Okay, uncle, I can’t do this anymore. Somebody step in and manage my infrastructure.”
And that created a ton of opportunity for our partners.
Post-Mythos is almost doing the same thing, because it’s so complex for a customer to try to stand this up on their own that you almost have to work with the partner community. You have to leverage a strong partner to deliver that.
And so there are some customers, just like in post-COVID, that are saying, “You know what? I’m just going to go back to building cars. Real-time automated patching with this level of intelligence is just not my thing.”
And that’s okay. Again, it presents a huge opportunity for our partners.
Robert Dutt: You talked about AI infrastructure becoming more diverse, going from the gigawatt-scale data centre deployments to smaller AI pods to inference at the edge.
How does that change the opportunity for partners, especially those who traditionally build and manage enterprise infrastructure?
Alex Pujols: I think it creates another pathway of opportunity for them.
There certainly is enough business where partners can sustain themselves on the traditional data centre buildouts, but there’s a rich set of opportunities on the horizon.
A lot of them have already dropped around edge inference use cases.
It makes me think of a partner I was talking to that was working with their client, and they were trying to figure out how they could use AI at the edge to reduce traffic fatalities at railroad intersections.
Their client was in railroads.
And one of the issues is that some of these fatalities, they believe, could be prevented if they could algorithmically find out what areas are potentially most prone, and whether they should develop more safeguards, things like that.
And so they ended up taking some edge compute and basically having it in proximity to the railway, where a lot of the intelligence and the analysis was actually happening at the edge.
Because those satellite links for the railroad are very expensive. They don’t transmit a lot of data, so you have to push the intelligence to the edge.
And they were really nice use cases where they were actually able to reduce traffic fatalities at intersections by leveraging that.
And that’s, in essence, an edge AI inference use case.
But that’s not something that you get in the standard book.
Some creativity is required to go out there and uncover those opportunities, but they’re there.
The technology is ready, the use cases are ready, and it’s really about making sure that the partner’s ready.
Robert Dutt: How much does that, especially as things become more personalized, bespoke at the edge, as you’re describing, change who the partner involved is?
I’m just curious how broad an opportunity this becomes beyond the big integrators who maybe have traditionally had that play.
Do you see it becoming a broader channel play?
Alex Pujols: I think so.
And I think there’s an opportunity for some new partner types: developers, for example, advisor-type partners that may not make their business on selling a particular piece of infrastructure or hardware, but are more involved in the broader consulting around how to address these unique use cases.
I know one of the things that we get asked quite a bit when we work with partners is specifically around solution development.
And when we’re developing something, we’re co-developing with them, instead of just focusing on a traditional use case, like just sitting down and doing a design-thinking session for a week.
Let’s just sit down, put all the technology aside. Let’s just sit down and have a conversation around what it is you do, what we do, what the opportunities are, and what we can build together.
Not what can we sell together, but what can we physically build?
And I think there are a lot of partners that are starting to have those conversations with customers.
Like, “Hey, Mr. and Mrs. Customer, what do your accounts need? What do they need from you? What could we build together?”
And it sort of democratizes technology more broadly.
It makes it more about the group that can build, that can brainstorm, be creative, and build more effectively, as opposed to the partner that can just do a traditional resale-type motion.
Robert Dutt: One thing that caught my attention when you were talking about Cloud Control and the idea of partners using agentic operations to manage increasingly complex customer environments.
You talked about availability in the U.S. and then into other regions.
Where does Canada fall in that rollout? Are we in lockstep with the U.S., or is that somewhere else?
Alex Pujols: So Canada is in a soon-to-follow next phase.
The current GA is for the U.S., minus federal accounts, but that’s going to scale out significantly.
I don’t necessarily have a date I could give you, but certainly Canada, Latin America, rounding out North America, is next on the list.
Then EMEA, then APJC.
Robert Dutt: Okay. And when partners do get access to these capabilities, how does it change their managed services business?
Is it mostly about being more efficient, or does it enable services that they couldn’t do before?
Alex Pujols: Oh, I think if I was a managed service provider, I would be in love with this thing.
One of the number one pieces of feedback we get from MSPs, especially when they’re incorporating Cisco’s products more broadly, is that whether it’s APIs or however they interface with our technology, it may be different product to product, product family to product family.
That creates complexity. It creates challenges for them to build out their services.
One of the great things about Cloud Control is it provides that single front end so that those MSPs can basically tap into Cisco’s entire portfolio from a common front end.
And so it makes it so much easier.
It’s not even just the beautiful interface that we show on stage. It’s also APIs.
And so, if you have a customer that’s embracing Cloud Control more broadly, as an MSP, I can leverage the APIs that I’m given by Cisco to tap into that customer data to build out more effective services for those same clients.
So when done right, it’s going to dramatically reduce the friction and the time to value for partners that are trying to build and develop capabilities for customers.
Robert Dutt: All right. You talked about partners being able to build agents in AI Studio in seconds without programming expertise, kind of vibe-coding agents, basically.
And then potentially commercializing them, bringing them to the marketplace, which is obviously something that we’ve talked about in the channel for a long time: the ability to do repeatable solutions, and the power in that, and productizing, for want of a better word, solutions.
Powerful capability, but I guess the question it brings to mind for me is, if I can build that agent in seconds by describing the business process I want to solve, presumably my competitors can, too.
How do I maximize the differentiation and commercial value from those, especially if I’m looking to bring them to the marketplace and go beyond the specific-to-one-customer kind of opportunity?
Alex Pujols: Well, there is an element in there of he who is most creative will win.
And so the creativity is still going to be required.
But there’s an aspect of this where, if I’m a partner and I’m building this out for my client, my client, in essence, is authorizing me to view their information in Cloud Control.
So number one, for a partner to be able to build that out for a customer, they have to be trusted by the customer.
And so it isn’t nearly as simple as any partner can log in and basically create anything that they would need for any client.
There’s still some trust required on the front end, coupled with the creativity and the innovation.
Then I do think that they can create some unique services, ultimately, that they could offer to their clients.
Where it gets interesting is when a partner wants to post that to the marketplace.
We’re currently working through this. It’s sort of in beta right now.
But we’re working through what type of protections we can offer that partner so that they can obviously benefit from potentially monetizing that capability without opening it up to their competitors that may otherwise try to go in there and replicate it.
We make it really simple, but there are still some elements that are required, and I think that’s the value the partners are going to focus on.
Robert Dutt: All right. You also talked about the shift towards smaller, specialized language models, particularly for customers concerned about cost, or where the application demands or works best with a small language model.
You guys are introducing Model Foundry to help partners work with those models.
What does that enable partners to do, specifically, that they couldn’t easily do before?
Alex Pujols: Yeah. So there’s an inherent complexity with providing security and governance, and in essence a harness, around some of these localized language models, whether they’re getting them from a marketplace like Hugging Face, for example, or whether they’re developing them internally.
And so that’s something the frontier providers have had to deal with privately on their side.
It’s not something our customers, the average enterprise, have had to do with small language models.
And so what this is going to do is allow our customers, or our partners on behalf of our customers, to drive a level of governance around those local language models that otherwise, really, right now, are deployed without safeguards for the most part.
And so there’s Model Foundry, but there are also a couple of other pieces that plug in that sort of act as a bit of a security wrapper around these SLMs.
Because the last thing you’d want to do is include that capability in your enterprise to reduce cost, only to create a cost elsewhere by opening up more risk for the business.
And so the idea is that you should be able to take advantage of that while ensuring that all of the actions with that SLM are secure.
And so that’s one of the things that isn’t at scale. This is not an issue with the enterprise, but it’s certainly on the horizon as more enterprises look to negate some of that cost.
And also, by the way, take advantage of more intelligence out of those systems, because a lot of those small language models are calibrated for specific domains, whether it’s medicine or finance or whatever it might be.
So you can get the best of all those worlds, but you have to do it in a secure fashion, and that’s where that helps out.
Robert Dutt: As we get into that side of things, does that bring AI development more within the reach of the traditional Cisco infrastructure partner, as opposed to having previously a pretty specialist practice required for it?
Alex Pujols: I think so. I think so.
Time will tell how quickly those small language models can be tuned to a place where they’re effective and they’re outperforming some of the AI frontier model peers for specific tasks.
But I do 100 percent think that it is going to be easier.
We’re going to see much more of that.
And frankly, you’re probably going to see more from Cisco in terms of things that we’re releasing as open source to help support that transition.
Robert Dutt: A little bit of a tangent, but you’ve had the Partner Innovation Challenge for a number of years under your belt.
I’m curious how the current milieu, the current transformations going on across business and the channel everywhere around AI, are changing what you’re expecting and what you’re seeing in the context of those awards.
Alex Pujols: Oh, goodness.
Was it two years ago? Trevor, was it two or three years ago?
The percentage of submissions that included anything AI-related was less than 20 percent.
And I think we’re over 95 percent now.
Robert Dutt: There are five percent who are holding out?
Alex Pujols: Yeah, well, very narrow use cases on the submissions.
But yeah, no, it’s here to stay. It’s permeated through everything.
Even our Partnering for Purpose awards, which are the solutions that have the best humanitarian impact, basically, even those are AI-related or AI-driven.
And so it’s everywhere. It’s here to stay.
It’s fascinating. It’s interesting.
What it’s also done is — this is probably because of Codex and other capabilities — it’s even further democratized who can build.
Because there was a moment in time where building was really relegated to our largest partners.
You couldn’t stomach the cost or hire the expertise to do that if you were a smaller partner.
That’s not the case anymore.
Literally, you have a great idea, you have a couple of friends, you have a weekend, you build it.
But then you spend weeks — and what our partners are doing is, they’re doing that fast, but they’re still spending a ton of time refining it, pressure-testing with their clients, rebuilding it, enhancing it, and then eventually submitting it to us as part of the competition.
So the quality of what we’re seeing, it was always high, but it’s even higher now.
And what partners are doing with AI is just ridiculous. It’s awesome.
Robert Dutt: All right. Last one for me.
You mentioned that partners are getting involved earlier in developing many of the capabilities, many of the products, many of the technologies you’re bringing to market.
Can you give me an example of something that changed along the roadmap as a result of a partner’s feedback that this is what they want or need?
Alex Pujols: Well, it might be a little secret sauce, but I’ll touch on the high level.
Cloud Control is a perfect example.
So we were engaged, we were involved with the business unit on this product for a long time.
And even at the earliest phases, when there were many Cisco employees who weren’t familiar with it, we had already brought in GSIs and managed service providers, basically to walk through the concepts and for them to share with us a day in the life of being a managed service provider.
And ultimately, what they would need from the product to realize the goals that we wanted.
And so they were very vocal, very supportive.
A lot of those partners became our alpha partners, our beta partners, and EFT partners, and kind of helped us early on.
So definitely, the partner community had a strong opinion and definitely shaped where we ended up with Cloud Control, and frankly, where we’re still going.
Robert Dutt: All right. Look forward to seeing where it goes next.
And once again, I appreciate your taking the time.
Alex Pujols: Awesome. Thanks. Appreciate it, Rob.
Robert Dutt: There you have it, Alex Pujols from Cisco.
Thanks again to Alex for joining us here at Partner Summit in Vancouver, and thanks for listening today.
A few things that stuck out for me from that conversation.
First, the pace at which AI is changing the security business.
Alex described Cisco’s five-stage Vulnerability Operations Maturity Model and suggested that the average maturity of partners has moved from Level 2 to a strong Level 3 in just the past six months.
That’s a pretty significant change, particularly when you consider that some of the most advanced capabilities require partners to develop managed services practices.
Second, there’s a growing opportunity for partners to build their own intellectual property.
Cisco is making it easier to create AI agents, customize specialized language models, and potentially bring those solutions to a much broader market.
But I thought Alex’s response to my question about differentiation was really interesting.
If creating an AI agent becomes something almost anyone can do, the value isn’t necessarily in creating the agent itself.
It’s in understanding the customer’s business, having access to the right information, and developing something that genuinely solves a problem.
Cisco’s still working through some of the commercial protections that will help partners monetize those solutions through its marketplace.
Finally, I was struck by how much broader Cisco’s definition of innovation is becoming.
Alex told me that just two or three years ago, fewer than 20 percent of submissions to Cisco’s Partner Innovation Challenge involved AI, and today that figure is above 95 percent.
And the partners building those solutions aren’t necessarily the largest organizations with the biggest development teams.
That’s an interesting change in what’s possible for the channel.
Thanks for listening.
If you enjoyed the conversation, please subscribe or follow the podcast wherever you get your podcasts — Apple Podcasts, Spotify, YouTube, the usual suspects — and consider leaving us a rating or review.
It helps other members of the channel community find what we’re doing here.
We’ll have more conversations from Cisco Partner Summit, along with our regular coverage of the people, technology, and business opportunities shaping the Canadian channel, later this week.
Until next time, I’m Robert Dutt for ChannelBuzz.ca, and I’ll see you in the channel.
