01732 759725 magazine 31 will stop typing functions and start describing intent. Agents will build, test and validate systems on their own, communicating in their own compressed dialects. Humans will design outcomes, not loops, and the last person who can actually read the code will sound like a historian. Woodson Martin CEO, OutSystems Most AI agents will fail in production Demos of autonomous AI agents are spectacular. Unfortunately, these demos crumble when they meet an enterprise production environment, and this is likely to get worse soon. Why do they fail? Because real world environments involve constantly changing APIs, incomplete or messy data, AI not just to source tools, but to engineer outcomes. Instead of recommending SaaS vendors, CIOs will assemble multiple LLMs to build solutions to solve today’s problems while anticipating what’s next. The IT function will no longer be just about infrastructure; it will be about delivering corporate intelligence with AI-driven solutions and providing leverage across every critical business platform. AI will redefine the CIO as a business innovator, not just a technology operator. Tony Gentilcore Co-founder, Engineering at Glean Code is the new Latin: historic, but mostly obsolete Natural language has replaced syntax as the interface of creation. In the next five years, developers common pipelines, shared SLOs and unified accountability for the full lifecycle of AI-enabled services. To reflect this evolution in how modern software behaves, end-toend observability will be essential to making sure that AI operates reliably in production i.e. within the workflows, guardrails and delivery pipelines used for the rest of the cloud-native system. Mike Blandina Chief Information Officer, Snowflake AI will redefine the role of the CIO The role of the CIO will shift from IT leadership to ET (enterprise technology) leadership. Traditional metrics like ticket counts will still matter, but forward-looking CIOs will adopt a solution mindset. The modern CIO must leverage AI PREDICTIONS continued... Derek Thompson Senior Vice President & GM, EMEA at Workato The agentic era will give CIOs new strategic power If AI agents are set to completely transform efficiency within an enterprise, it follows that the CIO responsible for deploying and managing that network of agents will reach new levels of influence. In 2026, ambitious CIOs will see this new responsibility as an opportunity to build their profile across the organisation and with the board. However, to be successful, CIOs must be hands-on and build agents themselves, not just delegate to their team, so that they understand AI through practice not presentation. AI will be aligned with core KPIs The past year consisted of agentic experimentation and pilots. In 2026, for AI projects to stick they must be directly driven by core business KPIs. Too many agent stories are being developed on the periphery of the organisation and are not anchored to critical business operations; their impact will therefore be minimal. Agents should be designed to serve the enterprise’s most important KPIs as that is where value is concentrated. Middle managers are more at risk from AI than grads Contrary to headlines declaring that AI is rendering graduate roles obsolete, the younger generation is at an advantage. They question established processes, asking ‘why do you have to do it this way’. In 2026, middle managers, often set in their ways, need to be open to learning and evolve their thinking. This year will be about hiring people who genuinely understand how AI transforms work and life, for example those who use ChatGPT to plan a three-week vacation, rather than simply listing it as a technical skill on their CV. Agents risk re-creating tech siloes One of the biggest problems to emerge from the era of digital transformation was SaaS sprawl and an un-manageable, disconnected tech stack. If agents are developed in isolation by different teams, organisations risk repeating their past mistakes, now rebranded as agent sprawl. In 2026, there must be a focus on ensuring that AI is orchestrated across the enterprise, so all processes and data are connected towards shared goals. Any organisation that believes the more agents you have the better, will face a new variation of technical debt.
RkJQdWJsaXNoZXIy NDUxNDM=