Cisco: AI-as-a-Service is the channel’s next major opportunity

Partners will increasingly manage AI infrastructure, consumption and economics rather than just deploying technology, says Cisco CTO

As enterprises begin moving AI from experimentation into production, Cisco believes the channel's biggest opportunity will be helping customers operate, optimise and govern AI environments rather than solely deploying the underlying technology.

Speaking to CRN UK, Cisco’s CTO UKI Rob Lay said the next generation of managed services could extend beyond infrastructure management to include AI operations, consumption monitoring and even helping customers manage token-based AI spending.

“I don’t see any reason why we wouldn’t see managed services partners looking at delivering a true AI-as-a-service type capability to our customers that includes all of that,” he said.

“I think we’ll have a whole chunk of partners who will deliver elements of that – maybe AI-managed infrastructure, for example.”

He also suggested entirely new specialist services could emerge around managing the commercial impact of AI.

“We might end up with partners who build your AI economics services to deliver that capability purely for our customers, and that’s part of the value that our partners add.”

Usage-based AI pricing reshapes the channel

Lay’s comments come as software vendors increasingly adopt usage-based pricing models for AI services, with customers paying according to token consumption rather than traditional user licences.

The shift has prompted growing debate across the channel about how partners should price, deliver and manage AI-enabled services.

Asked whether those pricing models fundamentally change the partner business model, Lay said they would require service providers to evolve their offerings rather than reinvent themselves.

“It’s certainly going to drive an adaptation in terms of how our partners deliver services to our customers,” he said.

He compared the transition with previous shifts towards automation and managed services, arguing that “AI is just the latest evolution of that.”

Cisco is also building capabilities that allow customers to understand the financial impact of AI consumption, creating opportunities for partners to build advisory and managed services around AI operations.

“We can give you as a customer a view on what your economics looks like – the economic impact of the token burn that you have within your AI estates.

“There are capabilities there that our partners can also wrap services around, wrap value around, to deliver to our customers,” he said.

While AI may be changing how services are delivered, Lay said human expertise would remain essential.

“We still need human in the loop. We’re not in a world where we are simply going to hand over everything to AI.”

Every AI project becomes a networking challenge

Unsurprisingly, Cisco also believes the infrastructure conversation around AI is often overlooked.

While many organisations remain focused on identifying compelling AI use cases, Lay argued that successful deployments inevitably become networking and security projects.

“Every single AI use case today will end up being a network problem,” he said.

Whether customers are building AI infrastructure in their own data centres or deploying AI-enabled applications across campuses and branch offices, he said every deployment ultimately depends on modern, AI-ready networks.

“None of these AI projects will truly get to fruition without a really solid AI-capable, AI-ready network.”

He said that creates another significant opportunity for partners, who increasingly need to bridge the gap between business transformation and infrastructure modernisation.

“We need our partners to be going into customers, going, ‘How do you need to transform your business? What’s it going to look like? Here’s how we build the use case for you. Here’s how we take you on that journey from use case... to scale production.’”

Despite the excitement surrounding AI, Lay said few organisations have yet reached that stage.

“We’ve got a small handful who have really, really well-defined use cases that deliver genuine business outcomes.

“The vast majority of our customers are still trying to figure out what the true use cases are,” he said.

Tackling pricing and channel supply chain pressures

Lay also addressed partners’ more immediate concerns of rising infrastructure costs, dynamic pricing and ongoing supply chain pressures. He acknowledged that AI hardware remains expensive, particularly around servers, networking and memory, but argued that customers are ultimately willing to invest where there is a clear business case.

“When we’ve got customers who have well-defined use cases, then they can build that return on investment model because they know the tangible benefit that they’re going to get back.”

He also accepted that pricing volatility remains a challenge for the channel.

“Dynamic pricing is an unfortunate reality of the world that we’re living in today,” he said.

To improve stability, Lay said Cisco has diversified its supply chain since the pandemic and has been redesigning hardware to reduce dependence on specific components. Earlier investments in developing its own silicon are also providing greater flexibility when responding to changing market conditions.

“We’re doing lots of different things to try and improve that engagement with our partners and try and give them that level of stability that they’re looking for from a pricing perspective,” he said.