Here are a few predictions for 2026 from:
- Ben Sekhon – Regional Vice President of UK and Ireland for OutSystems and has a unique perspective to offer on how he sees the agentic and low-code industries evolving locally.
- Woodson Martin – OutSystems’ CEO. Previously serving 18 years in global leadership positions at Salesforce, Woodson now leads low-code pioneer OutSystems in its vision to transform enterprise software delivery through AI.
Ben Sekhon, RVP UKI at OutSystems, commented:
1. UK’s Hidden Edge: How Legacy Knowledge is key for 2026 – The UK has a lot of legacy enterprises who have been, historically, slow to adopt new technologies. However, we’ve seen them eagerly dive head-first into AI alongside the young startups. Fearmongering headlines may point to an AI bubble, but our on-the-ground feedback tells us the UK is going all in with AI mandates for 2026.
Old companies often mean old infrastructure; the number one priority for UK companies adopting AI in 2026 is app modernisation. Many existing systems are several decades old, patched together with countless upgrades. We no longer know how they operate and must therefore rely on AI to understand the system, rebuilding it to fit the modern architecture.
However, legacy systems are also a largely untapped goldmine of institutional knowledge. They contain decades of built-in business rules and process logic; knowledge that no documentation library can fully capture. AI-powered systems and agents, deployed on the right kind of platform, can turn that dormant knowledge into a strategic asset.
2. AI Phase one: Building a new-world architecture that’s agentic ready – Within the tech development bubble, we are innovating for the economy ten, even twenty, years in the future. However, what legacy enterprises like banks, insurance companies and telecoms need AI automation for immediately is for app modernisation, often at cost neutral – not abstract futuristic ideas that never leave the ground.
AI tools can help autonomously update outdated systems without out-sourcing or hiring a new team, as agents evolve automation as we know it. We’ve seen automation move from scripts to RPA to intelligent automation; unlike these previous iterations, agentic automation blends LLM reasoning, orchestration, APIs and human oversight to handle judgment based tasks.
That will be phase one: using AI to quickly and cheaply lift and shift old applications into a new-world architecture that can tackle specific, high-volume business pain points and, most importantly, be agentic ready. It is a solid foundation, equipped for today’s fast-moving economy, that can then be enhanced upon with later releases, agents, and AI features.
3. UK Tech Talent: Essential, Strategic, and Balanced – Developers are essential workers in the AI era. Fears of AI-driven mass lay-offs in 2026 can be abated as a human-in-the-loop is essential for governance, security and quality control. This oversight is crucial for ethical use and preventing AI errors in critical systems – vital assurance needed, for example, across the highly regulated UK financial and public sectors. Particularly with processes like vibe-coding, software developers are the best positioned for AI oversight as their way of thinking is the most aligned to understanding machine systems and programming. Their business knowledge ensures AI outputs meet all regulatory requirements.
That said, the job description for a software developer in 2026 will undoubtedly change. As humans and AI collaborate more, developers will have time to be creative – a skill not previously associated with the role – and evolve from code-writer to AI strategist. This shift allows developers to become internal consultants focused on high-value business problems, rather than spending 80% of their time on maintenance and debugging. The developer can drive innovation for the enterprise in the way they provide oversight to the AI, refine output and define its parameters.
Woodson Martin, CEO at OutSystems, commented:
1. AI will increase complexity before it reduces it – The potential for AI to accelerate and scale software development is huge. We are already seeing how vibe coding is turning what used to take weeks into minutes and even seconds. What most people aren’t seeing today is the potential for AI to help with the harder stages of the enterprise software development lifecycle. They are overindexing on the build phase, but creating bottlenecks downstream in quality control, security, maintenance and updates. 2026 will be the year IT teams turn their focus to containing and auditing ungoverned AI-generated apps and agents. Those who use AI to systematically govern their full portfolio will be the first to realize the true potential of AI-driven development.
2. Code becomes cheap → Architecture becomes expensive – For decades, the strategic challenge was writing the code to realize a given architecture. Now that AI can generate functional code, its strategic value is lost. Architecture, integration, data modeling, and lifecycle governance become the new strategic moat. The stack collapses upward. Value and expense will concentrate in these layers:
– System architecture
– Data modeling
– Integration strategy
– Lifecycle governance
As AI commoditizes the lower levels of the stack, solutions and talent that design, integrate, and govern these complex, AI-driven systems become the most valuable resources.
3. 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
– Conflicting business rules
– Complex identity and permissioning models
– Non-deterministic behavior leading to unpredictable outcomes
Most autonomous agents will need tight orchestration layers and human-in-the-loop controls. In other words, they’ll need new platforms. Autonomy only works in fantasy. It’s orchestration that wins in reality.
Be the first to comment