Throughout 2025, enterprise software providers have launched and spoken of agentic AI, claiming it will revolutionise the operating model of organisations. With agents available in key enterprise software-as-a-service (SaaS) applications, organisations require visibility, and an ability to plan for, protect and manage agents effectively.
For enterprise architects, this poses a major opportunity and a challenge for their role. Enterprise architecture is adapting to map the new software landscape, such as the interdependencies between systems and agents.
Nick Tzitzon, vice-chairman of ServiceNow, described enterprise architecture as a blank screen in his recent ServiceNow World Forum keynote in London, giving a clear sales pitch to organisations that platform providers want to be the foundations for the new architecture, saying: “The more flexibility and choice we can give you will be how you see us as valuable.”
Femi Bamisaiye, COO and technology leader for pensions provider Nest, says: “The role of the enterprise architect is even more important now. That person will look at agentic AI in partnership with the CIO or CDO, and be the person that can articulate and translate its business impact and opportunity.”
He adds that the articulation skills are of particular importance: “Much of my career has been about simplifying technology, cutting tech debt and reducing complexity, but now with agentic AI, all the complexity is coming back, so I need somebody that can look at the enterprise holistically and then make it hang together and connected up.”
Because of its association with architecture in the built environment, it is often assumed that enterprise architecture is built in stone. Those blueprints and complex architectural drawings cannot be changed, or to redraw them is too difficult and likely expensive. As a role, some in enterprise architecture have fostered this belief, but modern enterprise technology and the nature of modern business require a far more flexible approach to enterprise architecture.
“In 2022, we realised the need to transform, as customers were getting more demanding,” said Mark Pearson, chief architect for defence manufacturers BAE, at the 2025 Gartner Symposium. He explained how the core business of BAE is changing, as it increasingly develops unmanned aircraft, which are heavily reliant on technology, and is currently developing the new Tempest military aircraft.
With major conflicts raging in Europe and Africa, and a demand for NATO nations to increase their military funding, the suppliers must deliver ever more modern military might at a rapid pace. “We need to be agile and to change the operating model to support that.”
We need to be agile and to change the operating model to support that
Mark Pearson, BAE
Just like a pure digital business, Pearson and his colleagues in enterprise architecture have a constant need to deliver business outcomes faster and cheaper. When customers need to change direction, the enterprise architecture of BAE has to be able to switch course just as quickly. Before Pearson began re-architecting, this was not the case at the UK headquartered business. Pearson adds that not only must modern architecture be flexible, “we need to do things fast and first time right, so they are correct by design”.
This approach reflects an ongoing trend in enterprise architecture called the Lego bricks methodology. Since the arrival of service-oriented architecture (SOA), then APIs and platforms, organisations have embraced an ability to connect in services and technologies as and when needed. Enterprise architects have, largely, embraced these developments. Agentic AI differs in that its pace and ease of use is disruptive and – in the worst cases – a risk to the organisation.
Kit Cox, CTO and founder of Enate, a supplier of process orchestration software for the services sector, says organisations and enterprise architects will need to balance delivering on the needs of the customer with savings and the business benefits of standardisation. He believes enterprise architects will need to learn from US coffee chain Starbucks, as it has elements of standardisation, but also distinctions per geographic market it operates in: “It is the consistency that is delivered.”
Interestingly, in the second half of 2025, the major suppliers took a much more careful and pragmatic approach to agentic AI. Early in the year, the hype suggested that agentic AI was going to rapidly transform the enterprise, but by the autumn suppliers were focusing on their agent management technologies, saying they wanted to empower business and technology leaders.
Clear vision
With agentic AI potentially making it easier to create business processes to alleviate workloads from staff, the enterprise architect has to provide clarity for business, its people and its suppliers. At BAE, Pearson defined the role of enterprise architecture’s as digitising the enterprise through greater visibility of value chains and capabilities across the business.
It is clear that as the enterprise has abilities and skills that are both human and agentic code, there is a need for a person or department to map, monitor and utilise these. Without a role that categorises and understands all the capabilities of the organisation, duplication of effort will run rife. CEOs are hoping that agentic AI will increase productivity and eventually the bottom line, therefore duplication is the enemy of productivity.
Pearson also sees enterprise architecture as central to fostering collaboration and cross-functional working in the organisation. Again, if an organisation has a strong culture of cross-functional working, AI agents can be deployed safely, with the right questions asked of its role by a (hopefully) diverse range of staff.
Pearson says the much-desired demands to be data-driven and agile as an organisation stem from the abilities and outputs of the enterprise architect. This is delivered, he says, though some of the traditions of the role – in his own organisation, it is the enterprise architects who have a holistic view of the technology stack that includes SAP, ServiceNow, JIRA, Apptio and others.
However, previous complaints that enterprise architects are too rigid and use doctrines such as The Open Group Architecture Framework as blueprints that cannot be challenged still exist.
“All too often, we find ourselves on a call with a large number of people that call themselves enterprise architects and they don’t add any value,” Cox says. In his experience, the enterprise architect is now responsible for managing legacy technology estate and making that “AI friendly”.
At BAE, Pearson invested in upskilling the enterprise architects to ensure this was not the problem. As new and more dynamic technologies and methodologies were introduced, the investment in skills development paid off with high levels of adoption.
Cox adds that enterprise architects looking to be at the forefront of agentic AI adoption also need to invest time and energy in getting close to the business operations: “It is really easy to get focused on processes, but people get too hypothetical – when you talk to people about the work they are doing, you get a much crisper conversation. It is a frustration that enterprise architects are part of the problem, not the solution, and that is a brush that IT gets painted with.”
Just before the global financial crisis in 2007-2009, IT departments were accused of not being aligned to the business outcomes and for being the department that said no. Agentic AI offers organisations a great deal of potential, but there are risks, especially in information management. Cyber attacks, such as the one on Tata-owned car manufacturer Jaguar Land Rover resulting in a loss of £485m in the quarter ending 30 September 2025, showcase the risk. The need to comply with laws such as GDPR mean the adoption of agentic AI has to be well-planned and executed.
Safer with architects
Notwithstanding some need for change, the role of enterprise architect remains vital if organisations are to navigate the adoption of agentic AI.
Bamisaiye says that an enterprise architect is “really important for running end-to-end technology”. At his organisation, a third of employees work in technology, data and operations, indicating the importance of tech to the operations of the business and services to customers. Alongside this, there is a significant outsourced partnership.
“That means I need someone that looks at the entire landscape, outsourced or not, and at our IT and operations, so they understand all of the business functions and business processes,” he says, adding that the enterprise architect is not just a planner, but is responsible for making sure that the technology solutions deliver on the expectations of the business lines.
Agentic AI is not the only force reshaping business processes. Oganisations are operating in a period of dramatic market changes as tariffs and geopolitics change supply chains on a weekly basis. Agentic AI will therefore enable organisations to operate processes for both long and short periods of time. Staff will be freed to use their experience in helping the business navigate this changing landscape.
The enterprise architect in an agentic AI era could, in fact, become the enterprise gardener, selecting agentic AI plants that suit the weather of that year – for example, a cold year shaped by tariffs or a warm year benefiting from the warm winds of growth. A garden changes according to the seasons or from shade from the environment just beyond the hedgerow. Crafting the garden beds, crops and displays to deliver something for everyone that uses the garden is a form of architecting for change.
