Innovation Governance in Late 2026: The Supervisory Reset the Innovation Office Did Not See Coming

In November 2025, the ECB did something unusual with its supervisory priorities. It cut the list from three to two for the first time in five years, and quietly moved banks' digital and AI-related strategies, governance and risk management into the medium-term supervisory focus. The Single Supervisory Mechanism will now run horizontal workshops with selected banks on generative AI applications, and it will read the innovation agenda as a governance question, not a technology one.

For most European innovation offices, that is a category shift. Innovation used to be observed by the CFO, tolerated by risk, and negotiated with tech. It is now, formally, a supervisory concern. The mandate, the KPIs, the reporting lines and the stakeholder map that most innovation offices carry into 2026 were designed for a different regime.

The problem the headlines miss

The public narrative is that banks have 'matured' their innovation function. Pilots down, deployments up, ROI discipline in. That story is broadly true, but it misses the harder part. The pressure is now bidirectional. The board wants fewer, larger, deliverable bets. The supervisor wants documented governance around every AI and digital adoption above a low materiality threshold. And the delivery organisation wants innovation to stop generating work that lands in change portfolios already at 110 per cent capacity.

Inside the innovation office, three things happen at once. The horizon-scanning remit gets crowded out by adoption support for last year's decisions. The governance layer expands to cover DORA third-party controls, EU AI Act annex III systems, PSD3/PSR readiness and the FIDA implementation debate. And the peer benchmarking question, always fragile, becomes essential, because supervisors are effectively asking innovation leads to justify their choices against what comparable institutions do.

None of this is a productivity problem. It is a governance problem dressed as a workload problem. And it is why the innovation office restructures that have already begun across European banks look, from a distance, like they contradict one another. Some banks are folding innovation into transformation. Others are separating it more sharply. A third group is quietly building a small policy and governance function inside innovation itself.

The three common responses, and what each one costs

The first response is consolidation. Fold innovation into transformation, group strategy or the CTO office. It works for delivery discipline. It works less well for horizon scanning. Twelve months in, most of these teams have lost the muscle to say what they are watching outside the current portfolio. Peer benchmarking becomes something the strategy team commissions from a consultancy once a year, and the innovation function ends up defending its remaining budget against every other line under the same umbrella.

The second response is defensive separation. Keep innovation formally independent, but wrap it in a governance layer that requires legal, risk, compliance and procurement sign-off on anything that moves. It is the safest structure on paper. It is also the slowest. It creates a queue between 'interesting' and 'adoptable' that most fintech partners will not wait through, and it leaves innovation leads spending their week on internal negotiation rather than on external signal.

The third response is the split model. A 'core' innovation team focused on adoption support for in-flight programmes, and a 'forward' team scanning three years out. It looks elegant. It duplicates roles, splits the peer network, and creates a status hierarchy that quietly demotes the horizon-scanning work. Within eighteen months, the forward team is either absorbed or dissolved.

Each of these responses is a rational answer to a real pressure. None of them fully solves the underlying problem, which is that the innovation office has to hold three things at once: horizon signal, adoption governance and peer intelligence. And most banks have built structures optimised for one of the three at the cost of the other two.

A more workable route

The banks that appear to be navigating this best are treating innovation as a governed portfolio, not as a function. That is a small phrase and a large shift. It means the innovation office is judged on how well it explains and defends the shape of the portfolio, not on the number of pilots run. It means that adoption governance sits inside the innovation office, not adjacent to it. And it means the peer benchmarking layer is treated as a permanent input, not an occasional deliverable.

In practice, three habits appear repeatedly.

First, a short quarterly cycle in which the innovation office presents a portfolio view, showing what is being scouted, what is in evaluation, what is being adopted, and what has been rejected and why. The rejected column matters. It is what a supervisor will ask for when the horizontal workshops begin.

Second, structured peer exchanges that are small, off the record, and repeat. Not conference panels, not vendor-hosted roundtables, but forums where three to eight comparable institutions can compare notes on the same live decision under an agreed rule set. In our own work at The Connector, this is what the Peer Forum and Roundtable formats exist for, and it is telling that the demand for them has moved from marketing-driven curiosity to governance-driven necessity in the last twelve months.

Third, a discipline around innovator discovery. Not more vendor demos, but fewer and better ones, framed by a specific portfolio question rather than a general 'what is new'. Our Discovery Innovation Meeting format is built around that discipline: an innovation lead brings a defined question, we bring a curated set of two to four innovators whose current state genuinely maps to it, and the meeting exits with either a next step or a documented 'no', both of which count as governance artefacts.

None of these habits requires a new team. They require a slightly different mandate for the team that already exists, and a slightly different reporting cadence into risk, transformation and the executive committee.

Why this matters right now

Three deadlines converge in the next eighteen months. The EU AI Act annex III obligations for high-risk systems are already live from 2 August 2026. DORA is in its second full supervisory year, and the ECB is now explicitly reading digital and AI governance as a medium-term priority. FIDA implementation, if the trilogue holds its current path, will require most banks to have their data-sharing governance defined by 2027.

Innovation offices that walk into 2027 with the same mandate they had in 2024 will be asked questions their structure cannot answer. That is not a hypothetical. Several of the banks we work with have already had informal supervisory conversations about how AI adoption decisions are governed inside the innovation function, and how peer inputs feed into those decisions. The direction of travel is clear, and it is not slowing.

There is also a talent dimension. Innovation leads who are trapped in defensive governance queues are leaving. The ones who stay are the ones whose banks have given them a portfolio mandate, a peer benchmarking budget, and a defensible cadence for both. That is not a soft point. It shows up in retention data across the top ten European banking groups already.

Closing thought

The uncomfortable question for every innovation lead reading this is not 'how do we do more'. It is 'what would we say if a supervisor asked how the last four AI adoption decisions were governed, and how peer inputs shaped them'. If that answer takes more than one page, the structure needs work. If it takes less than one page but rests on a single named individual, the structure needs work.

Innovation governance in late 2026 is not a rebrand. It is a quiet reset of what the innovation office is supposed to do, who it reports to, and how it justifies its portfolio. The banks that treat it as a structural exercise, not a communications exercise, will be the ones whose innovation function is still recognisable in 2028.

For the innovation leads asking themselves how to build the peer and discovery layer of that structure without adding headcount, this is precisely the ground we cover in the Peer Forum, Discovery Innovation Meeting and Roundtable formats, and in Finance X Magazine.

That is not the point of the article. The point is that governance has caught up with innovation, and the innovation office that treats that as an opportunity rather than a constraint is the one that keeps its mandate.