2026 Is the Year We Stop Experimenting

Pascal, you’ve said that 2026 is the year asset management stops experimenting. Why now?
Because experimentation has run its course. We’ve tried enough pilots to understand what works and what doesn’t. The problem today isn’t a lack of ideas. It’s a lack of decisions.

Decisions about what, exactly?
‍About how work gets done. For the last years, AI was mostly added on top of existing processes. Helpful, yes. But it didn’t change the system.

Why doesn’t that approach scale?
Because automating fragments doesn’t scale. You take a broken workflow, add a smart assistant, and hope efficiency shows up in the P&L. Sometimes it does locally. Across the organisation, it doesn’t.

So what changes in 2026?
The unit of change. We stop talking about tools and start redesigning operating models. The real question becomes: how should this process exist if AI is native?

What does that mean in practice?
You decompose workflows, research, risk, distribution, compliance, operations and rebuild them intentionally. Some steps disappear. Others get automated. Some stay human, but with more leverage.

You’ve said the bottleneck isn’t AI capability anymore. What is it?
Coordination. You hear this very clearly from the large platforms that run asset management at scale. AI doesn’t create value through isolated pilots. It creates value once data, governance, and workflows across investment, risk, and operations are aligned.

And without that alignment?
AI remains assistive. Useful, but supervised. Organisations still rely heavily on dashboards, controls, and manual checks because trust in the underlying system isn’t there yet.

Is that why data has moved to the centre of the discussion?
Exactly. Data quality is the single biggest predictor of whether AI outputs are trusted or questioned at every step. Across industries, poor data quality is the main reason AI initiatives underperform or fail. It’s not the models, it’s the preparation.

Even in finance, where accuracy and auditability are critical?
Especially in finance. AI only creates value here if data governance is strong enough that outputs can be traced, explained, and defended. Without that, you don’t scale AI — you contain it.

Does the same “stop experimenting” logic apply to blockchain and tokenization?
Very much so. Tokenization is following a similar path to AI. For years it lived in proofs of concept and innovation labs. That phase is ending.

What changes now?
Tokenization is moving from experimentation to infrastructure. Once tokenized assets reach sufficient scale, they force standardisation — of data models, compliance logic, and ownership records. That has real operating impact.

So this isn’t really about blockchain technology?
‍No. It’s about operating leverage. Blockchain is the mechanism, not the objective. The real value comes from cleaner data, programmable rules, and tighter integration between product, distribution, and operations.

Stepping back, where is the industry heading?
Toward fewer experiments and more integration. Toward systems designed end to end, rather than stitched together. And toward AI embedded into how decisions are prepared, monitored, and reviewed — not bolted on as an extra layer.

Final question. What should CEOs focus on as they look into 2026?
They already know enough to act. Now it's time to commit to a coherent end-to-end design: Their operating model. Their data foundation. And the role AI is allowed to play.

In one sentence?
2026 is the year asset management stops experimenting and starts delivering.

‍