Industry voice: Start 2026 the right way: the playbook for agentic success (and why most enterprises will miss it)
Agentic AI is no longer a “future bet”.
The challenge is that agentic AI rewards the deliberate, not the loud.
In our 2025 global research, 95% of major enterprises told us they intend to become theirown AI and data platforms— “like Amazon”—in less than 750 working days. That’s a dramatic recognition that the agentic future will be built fast, and built in-house. But the same data shows a widening gap between intent and outcomes:
- 23% of CEOs believed in 2025 they were already operating as their own agentic AI and data platforms
- Yet only 13% of the total universe are doing it well — achieving up to 5x the ROI of peers still stuck in “intent only”
That 13% represents more than 16,600 enterprises that have moved beyond pilots and prototypes into durable performance. The list spans banking and insurance to telecoms, healthcare, energy and manufacturing—organizations such as Danske, BNP Paribas, Aviva India, Banco Santander, Commerzbank, Deutsche Telekom, Eni, JPMorgan Chase, MetLife, Nippon Life, Oslo University Hospital, Tesla, Wells Fargo, and many more.
If you compete with them, assume they’ve already discovered how to be agentic at scale. If you work with them, ask what they did first. The most important lesson is this: agentic success is not one big play. It is a sequence, built on foundations, with realistic expectations.
The three decisions that separate winners from the rest
Across hundreds of variables we analyzed, the enterprises thriving with agentic AI in 2025 consistently made three simple decisions. Make them early in 2026, and you dramatically increase your odds of joining the 13%.
1) Treat AI and data sovereignty as mission-critical—because it is
The biggest differentiator is foundational: AI and data sovereignty.
Top performers built their agentic strategies around the ability to access trusted data anywhere, anytime, in any way, while remaining secure and compliant.
Crucially, most did not force everything into a single cloud. In our findings, sovereignty was achieved through a mix of architectures:
- Hybrid (cloud + on-prem): 44%
- Multi-cloud: 31%
- On-prem only: 20%
- Fewer than 10% relied on a single cloud data strategy
Why does this matter? Because agentic systems don’t just query databases. They reason across processes, tools, policies and context. That requires consistent governance, access controls, lineage, auditability, and performance—without compromising regulatory obligations.
Sovereignty over AI and data wasn’t just a technical preference. It was the only defining variable (out of 500 analysed) that statistically correlated with greater optimism about an organization’s future. In plain terms: the companies that got sovereignty right felt more confident, moved faster, and scaled agentic outcomes more reliably.
2) Set realistic ROI expectations—then win through compounding, not heroics
It’s tempting to justify agentic investment by demanding massive ROI from a single use case. The data suggests that approach is a trap.
The most successful 13% typically target 10–25% ROI per agentic workload, then expand horizontally across the organisation—compounding value through reuse of the same sovereign AI and data foundation.
By contrast, around 50% of enterprises pursuing agentic AI aspire to 50%+ ROI across multiple economic metrics but end up delivering less than 5% because they haven’t organised for systemic success. They over-index on the “one killer app” and under-invest in the repeatable platform, operating model, and sequencing required to scale.
Michael Gale, Wall Street Journal best-selling author on digital transformation and CMO at EDB, puts it:
“This sense of value in the aggregation and not just in specific agentic wins means the 13% have become agentic learning machines. They build scale, performance, differentiation and, maybe most importantly, a confidence that gives them the agentic flywheel for years to come.”
3) Sequence outcomes to build an agentic flywheel—fast
Even with the right platform and expectations, success still depends on what you do first.
Among thousands of possible paths, one sequence of AI use cases dominates the playbook of the 13%:
- Start with HR
- Move to IT operations and DevOps automation
- Expand into content and marketing automation
- Optimize supply chains
- Apply agentic AI to legal (document analysis, contract management)
- Extend into finance and expense automation
- Add digital simulation
- Improve workplace collaboration and productivity enablement
This sequence works because it builds confidence and capability across a wide range of roles without betting the company on a single transformation moment. Each step strengthens the next, creating an agentic flywheel that accelerates adoption, governance maturity, and ROI.
Do agentic “the right way” in 2026
Agentic AI will define winners and losers faster than many leaders expect. The enterprises already scaling it aren’t simply “more advanced”. They are more deliberate:
- Sovereign foundation first
- Realistic ROI expectations
- A proven sequence that compounds success
If 2025 was the year of intent, 2026 will be the year of execution.
Build your sovereign agentic and data platform on proven, enterprise-grade foundations, including Postgres, which already underpins a meaningful share of successful sovereign AI and data platform deployments.
Because in the agentic era, speed matters. Sustainable speed only comes from getting the fundamentals right.
This article is sponsored by EnterpriseDB.