‘You can only move at the pace of the trust you can build’: Brendan Mooney on government AI push

Kainos CEO reflects on the agentic AI demand, hyperscaler partnerships and building a UK technology ecosystem

Kainos CEO Brendan Mooney says public sector customers are taking a measured approach to agentic AI, with the IT services group finding that government organisations are more interested in tangible improvements to existing services than adopting the technology for its own sake.

The Belfast-based organisation has been working with AI and its predecessors for close to a decade, building on its wider position across UK government and healthcare.

Mooney tells CRN that developing those capabilities has been a gradual process rather than a sudden response to the generative AI boom.

“We were doing our research into AI almost ten years ago,” he says.

“Back then the term was much more around machine learning. We were looking at how do we understand this technology, where it’s going and how can we give advice to clients.”

One of Kainos’ early projects involved using machine learning for fraud detection around vehicle inspections, allowing potential cases to be identified for further investigation.

Mooney notes that work with organisations including the Driver and Vehicle Standards Agency (DVSA) and National Crime Agency subsequently helped Kainos establish a track record it could take into conversations elsewhere in government.

“It’s not an overnight success,” he says.

“It’s hard to fast track that. The reality is that you can only move at the pace of the trust you can build with clients.”

Agentic AI yet to take off in government

Despite the explosion of interest around AI agents, Mooney says Kainos is seeing considerably less demand for agentic technology among its government and healthcare customers.

The question for government, according to Mooney, is what additional value would be created by redeveloping an already effective digital service around agentic AI.

“There’s a sensible assessment to be made about the value of building something agentic today, versus simply doing something really well from a digital perspective.”

That does not mean government customers are uninterested in AI.

Mooney says much of the demand Kainos is seeing is instead focused on suppliers using the technology to increase productivity and deliver greater value.

Customers may also choose to embed AI when building new services but are less inclined to retrofit it into existing systems without a clear benefit.

AI productivity boost and limits

The company has spent around four years increasing its own internal use of AI.

Mooney describes the objective as making employees “AI augmented or AI enabled”, with the biggest impact so far coming within software engineering rather than consulting.

The company is seeing productivity improvements of between 22 and 25 per cent, according to the CEO.

But for Kainos, those gains do not automatically translate into higher margins.

Much of its government and healthcare work operates on a time-and-materials basis, meaning increased productivity can instead allow customers to receive more work for the same expenditure.

He adds that the benefit can still be significant for customers undertaking digital transformation programmes stretching across seven or ten years if AI allows that work to be completed faster.

There are also areas where Kainos cannot deploy the AI tools it would otherwise use.

Mooney says some customers explicitly prohibit their use because of the sensitivity of the data or applications involved.

Kainos operates across areas including policing, defence and immigration, leaving the decision over which tools can be used ultimately with the customer.

“We can give advice and guidance, but if they’re, you know, perhaps, we do work in police, we do work in defence, we do work in immigration, if those customers prefer not to be using tools, then that is obviously their decision to be made.”

The restrictions sit alongside a broader conversation around sovereign AI, with Mooney saying concerns around security, safety, data location and information potentially being used within AI models are increasingly being raised by customers.

Hyperscalers remain central as AI partnerships shift

Changing AI requirements are also having an impact on Kainos’ vendor ecosystem.

The company has longstanding partnerships with AWS and Microsoft, alongside a relationship with Google around Google Cloud Platform (GCP).

Mooney says those hyperscaler platforms remain highly capable of supporting customers’ AI requirements, but Kainos is also developing relationships elsewhere in the market.

“We have a growing relationship with Anthropic,” he says.

“There are changing or shifting partnerships, but we still would see ourselves as a very adept partner with the likes of the hyperscalers.”

Its broader partner model is similarly driven by filling specific skills gaps rather than distribution.

Mooney says between 98 and 99 per cent of Kainos’ business is conducted directly with end customers.

But as projects increase in scale and complexity, the provider works with specialist organisations where it cannot reasonably maintain every required capability internally.

Government and healthcare customers are also more and more identifying specialist suppliers they already trust and asking Kainos to incorporate them into larger projects.

“The partnership model for us is not about distribution,” Mooney tells CRN.

“It will be about how do we bring together for the client the very best set of skills, the best set of organisations that can really help deliver against that business requirement.”

Building a UK tech ecosystem

That approach is also feeding into Kainos’ efforts to develop a broader network of UK-based technology companies.

Mooney argues that the way government builds its technology stack should leave room both for global hyperscalers and alternative tech stacks where greater sovereignty is required.

The CEO describes Kainos as the largest UK-owned and operated provider of digital automation services to government and says the number of major domestically owned suppliers to the public sector has fallen substantially over the past two decades.

For Kainos, the response is not to abandon international technology providers, but to ensure there is also an ecosystem of UK-based companies capable of responding where government demands a more sovereign approach.

“For us, it’s about how do you build a partnership, a network of UK-centric organisations,” Mooney says.

“If sovereign becomes a really important thing for government, then there is a really strong set of organisations that can respond to demands placed on us.”