SAP AI Report 2026: Enterprise AI Now Handles 30% of Tasks

What happened

SAP has published its Value of AI Report 2026, produced in partnership with Oxford Economics and drawing on a survey of 2,600 business leaders across 13 countries. The headline finding is that artificial intelligence now supports nearly 30% of all tasks in a typical organisation, up from 25% the previous year — a sign that enterprise AI has crossed from experimentation into operational deployment at meaningful scale.

The report also tracks shifting expectations around agentic AI, the category of AI systems capable of taking autonomous, multi-step actions. Return-on-investment expectations for agentic AI rose from 10% last year to 17% this year, yet the report's central argument is that most organisations believe they are still leaving substantial value on the table. According to SAP's chief AI strategy officer Sean Kask, the gap between current returns and potential is not primarily a technology problem — it is a strategy, data and governance problem. AI deployed without adequate contextual grounding in business processes, clean data and clear oversight structures tends to generate activity rather than outcomes, and can introduce new risks.

Why it matters

For customer experience practitioners and service designers, the SAP findings reframe a question that has dominated boardroom conversations: the bottleneck to AI value is no longer model capability but organisational readiness. This is a classic behavioural-economics insight applied at enterprise scale — the tool is available, but the surrounding system (incentives, data hygiene, governance norms) determines whether it produces genuine outcomes or merely the appearance of progress. Organisations that invest in AI access without investing equally in the contextual infrastructure risk a form of automation theatre: visible activity, invisible impact.

For CX leaders specifically, this matters because customer-facing AI — from intelligent service agents to personalisation engines — is only as good as the data and process architecture behind it. A poorly governed AI deployment in a contact centre or digital channel does not just underperform; it can actively erode trust by delivering inconsistent, contextually wrong or opaque interactions. The report's implicit warning is that the organisations pulling ahead are those treating AI governance as a CX discipline, not an IT compliance exercise.

By the numbers

• 2,600 business leaders surveyed across 13 countries for the SAP Value of AI Report 2026, conducted with Oxford Economics.
• 30% of tasks in the average organisation are now supported by AI, up from 25% the prior year.
• 17% ROI expectation for agentic AI in 2025, compared with 10% the year before — a 70% increase in anticipated returns in a single year.

The Renascence take

The ROI gap the SAP report identifies is real, but the framing of "strategy, data and governance" risks becoming its own form of corporate abstraction. The more precise diagnosis is a context deficit — and context is fundamentally a human and organisational problem, not a technical one.

Most organisations will read this report and commission a governance framework. The smarter move is to audit the customer journeys where AI is already deployed and ask a blunter question: does this system actually know enough about this customer, at this moment, to be helpful? The behavioural principle underneath is ecological rationality — intelligence is only as good as the environment it operates in. A customer-obsessed operator should insist that every AI touchpoint is evaluated not on task-completion rates, but on whether it left the customer feeling understood. That is the governance metric that actually matters.