Nazia Pillay
By 2026, artificial intelligence will no longer be judged by its promise, but by its impact. For much of the past decade, AI has lived in labs, pilots and PowerPoint decks. The next phase is different. AI is moving into the operational core of organisations, reshaping how decisions are made, work is executed and value is created.
AI is becoming the most significant technology shift enterprise leaders will face in this generation. Not because the algorithms are new, but because the operating model required to make AI work at scale is fundamentally different.
One of the clearest changes heading into 2026 is the move from AI that assists humans, to AI that acts on their behalf.
Early enterprise AI tools functioned as copilots: surfacing information, generating insights or suggesting next steps. Increasingly, organisations are deploying autonomous AI agents that recommend actions – and take them – executing multi-step business processes within defined roles and controls.
This transition matters because it forces leaders to confront new questions of trust, accountability and governance. Autonomous AI can deliver significant productivity gains, but only if organisations are prepared to define where machines can act independently, where human approval is required, and how exceptions are handled.
In practice, this means treating AI agents less like software features and more like a digital workforce: assigned roles, clear permissions, monitored performance and escalation paths when things go wrong. Without this discipline, autonomy becomes risk rather than advantage.
Intelligence must be built in, not bolted on
Another defining trend is the move toward AI-native systems. Many organisations still treat AI as an add-on: a layer of intelligence bolted onto processes designed decades ago. That approach is reaching its limits.
AI-native architecture embeds intelligence directly into core workflows, allowing systems to understand intent rather than simply execute transactions. Instead of navigating complex menus and dashboards, users express what they want to achieve, and systems orchestrate the necessary steps across functions.
For leadership teams, this is not a user-interface upgrade, but a shift in how work gets done. Ideally, decision-making accelerates, organisational friction reduces, and the boundary between analysis and execution begins to disappear.
However, this only works when underlying systems are clean, standardised and integrated. Which leads to a harder truth many organisations are discovering.
Data quality is the real AI constraint
The biggest barrier to AI success is not model sophistication, but data reality. AI systems amplify whatever foundations they are given. Clean, consistent data produces reliable outcomes, while fragmented, poorly governed data produces confident nonsense.
This is why data has become the strategic nervous system of the modern enterprise. AI depends on shared definitions of customers, products, suppliers and processes. It requires transactional integrity, accessible historical context and the ability to combine internal and external information in real time.
Organisations that have postponed data discipline are finding that AI exposes weaknesses instantly, often in ways that affect customers, regulators or financial performance. In the year ahead, leaders will increasingly be judged on whether they treated data as a strategic asset early enough, rather than as an IT hygiene issue.
Closely linked to data readiness is a simple but central principle: keeping core enterprise systems clean.
Years of excessive customisation have left many organisations with fragile ERP environments that are difficult to upgrade and harder to integrate with modern AI capabilities. The shift toward standardised cores with extensions built outside the core system creates an environment where innovation doesn’t break operations.
For boards and executive teams, this requires a mindset shift. Standardisation is not a loss of competitive differentiation, but the price of adaptability. The differentiation moves to how organisations use data, design experiences and make decisions, not how many lines of custom code they maintain.
Technology alone will not deliver results
Perhaps the most underestimated factor in AI success is change management, which consistently accounts for a larger share of AI outcomes than technology itself.
AI changes roles, not just tools. Finance teams move from processing transactions to managing exceptions. HR shifts from administrative workflows to skills intelligence. Operations leaders rely more on forecasts and simulations than static reports. These changes demand new skills, new incentives, and new ways of measuring performance.
This year, leaders must invest in adoption with the same commitment and focus as they invest in new capabilities. AI literacy should be a core leadership competency not just a specialist function.
As AI initiatives multiply, so does the risk of fragmentation. Different business units experimenting independently can create inconsistent standards, duplicated effort and unmanaged risk.
This is why many organisations are establishing AI centres of excellence that coordinate AI innovation. Effective governance frameworks address five questions: how AI systems are approved and retired, how decisions are logged and audited, how policies are enforced, where human oversight is required, and how performance is measured.
In 2026, AI governance will be viewed much like financial governance: a prerequisite for trust, not a brake on progress.
From pilots to production or paralysis
A final challenge looms large: scaling. Many organisations are stuck in what has become known as “pilot purgatory”, where successful experiments never reach enterprise impact.
The reasons are consistent: poor integration with core systems, unclear ownership, lack of user trust, weak data foundations and vague ROI metrics. Moving from pilot to production requires deliberate planning, phased rollout and visible executive sponsorship. Leaders who expect AI to scale itself will be disappointed, while those who design for scale from day one will pull ahead quickly.
As we accelerate into 2026, AI is an operational reality. The real strategic question for leaders is whether their organisations are structurally ready for AI, with clean systems, trusted data, skilled people and disciplined governance. With these foundations, AI becomes a durable source of advantage.
In a volatile global environment, leadership is increasingly defined by the ability to move forward without perfect certainty. Business AI, deployed responsibly and at scale, is becoming one of the most powerful tools leaders have to do precisely that.
Nazia Pillay, Managing Director at SAP. She writes in her personal capacity.
