2026: The year of AI execution
As I said in Davos earlier this year, I believe 2026 will be the year of AI execution in Europe. In 2024, we saw widespread experimentation. In 2025, this gave way to patchy implementation, with AI applied to discrete business functions and processes – our analysis last summer found 48% of large companies across the region scaling large-scale, transformative gen AI investments.
But in 2026, I expect a step change. As business processes are completely reinvented with agentic and physical AI, growth and efficiency gains are starting to crystalise. Already, our latest Pulse of Change findings reveal that 85% of c-suite executives across the region plan to boost investment in AI over the next 12 months (vs. 2025), with those expecting growth benefits outnumbering those focused on cost reductions by four to one.
Based on the work we’re doing with clients, we clearly see the focus shifting to execution. Many companies are prioritising strategic and large-scale transformational programs – typically involving cloud, security and data modernization, and often combined with operating model and talent transformation.
This shows that clients are moving beyond simply layering AI onto existing processes or settling for incremental gains, such as those delivered by a customer service chatbot. Rather, they are making bold, large-scale bets – reinventing their entire operating models from the ground up with agentic AI. For these companies, AI isn’t just a faster way to cut costs; it’s a way to free up investment capacity and drive growth.
A stronger, sovereign AI ecosystem
On the supply side, Europe’s rapidly maturing AI ecosystem is another source of optimism – especially as sovereignty needs rise. Take the increasingly dynamic startup scene in Europe. Data collected by CB Insights shows deal volume decreasing but deal size increasing. In 2023, total AI start-up funding of $9.6bn was spread across 1,358 deals. In 2024, this increased to $12.6bn across 1,557 deals. However, in 2025, while total funding jumped sharply to over $21.1bn, the total number of deals fell to 1,494.
The upshot is that average deal size in the period 2023-2025 has almost doubled to $14.1m (see chart below), placing Europe still some way behind the US ($51.6m), but ahead of Asia ($8.6m. These larger funding rounds reflect investors placing later-stage bets on companies with proven traction and focusing on a smaller number of firms expected to lead a competitive market.
Recommended by LinkedIn
Average AI funding deal size in Europe ($m)
These numbers do not include our recent acquisition of Faculty AI in the UK, but our investment is emblematic of where the market is going: companies are looking to reinvent core and critical business processes with safe and secure AI solutions that result in tangible outcomes. Integrating the Faculty team of more than 400 AI native professionals, including data scientists and AI engineers, with our existing Accenture teams will allow us to better help clients execute on their AI ambitions.
Humans in the lead
A final, critical point is that execution remains a human endeavour. European companies are increasing AI investment, but too often we see tech trumping talent. Small expert teams experiment with AI in isolation, or companies roll out generic training while most employees lack the right incentives and access to work with AI. Our data highlights the issue: just 14% of European employees say that their leadership has clearly communicated how AI will affect jobs and skills.
The result is a talent strategy execution gap: nearly half (46%) of European organizations plan AI upskilling, but only 6% are redesigning roles. Business leaders need to think about how to reinvent work and teaming and how they refresh skills, mindsets and culture. Additional Accenture research reveals that organizations which deliberately put people at the centre digital reinvention outperform their peers.
Technology alone doesn’t deliver transformation; People do. This is why it’s critical that organizations move from the passive ‘humans in the loop’ to the active ‘humans in the lead’. And it’s how Europe can capitalise on both encouraging demand- and supply-side momentum to make 2026 the year of AI execution.
The accountability shift i.e., from AI capability to AI value, is the right frame for 2026. One dimension worth naming: the gap between organizations that close this accountability loop and those that do not is increasingly an architectural gap, not just a talent or adoption gap. Most enterprise AI deployments today are agentic at the workflow layer but classical at the computational layer beneath it. For the vast majority of use cases, that is appropriate and sufficient. But for a specific class of problem - combinatorial optimization over densely correlated networks where classical methods degrade under precisely the conditions of highest business pressure - the architecture beneath the agent determines whether the system delivers value or just reports it. The organizations building quantum-ready agentic infrastructure now are not making a speculative bet. They are ensuring that their 2026 execution investment remains relevant in 2030, when the computational substrate beneath today's agent architectures begins to evolve. Execution in 2026 is right. The question is whether that execution is designing for the compute of today or the compute of the next decade. QUAOS - http://linkedin.com/pulse/quaos-introduction-tilak-mitra-1pjse
Too many still treat AI literacy like a training module. The teams pulling ahead are the ones who make it part of how they actually work day to day. Not just something they learned about in a session last quarter.
Mauro, the 46% upskilling vs. 6% redesigning roles number is the one that should worry every executive reading this. You can train people on new tools. But if the role hasn't changed, nothing else does either. The hard work in 2026 is deciding what humans should own when agents are running the operations. Most companies haven't seriously asked that question yet.
The talent gap observation is the most critical point here. Europe can deploy AI infrastructure at scale, but if employees don't know how to work alongside agents, the ROI collapses. The challenge isn't just upskilling on tools — it's redesigning the human role itself. What does a project manager do when agents handle task decomposition? What does a data analyst do when the LLM queries itself? Organizations that answer these questions concretely and rebuild their operating models around them will be the ones that turn 2026 investments into durable competitive advantages.