Where humans meet AI agents.
The orchestration practice for regulated industries.
AI agents are entering the workforce as a new category of worker — onboarded, governed, integrated alongside your team. The technology is here. The integration is the work — and that is where we sit alongside your teams.
$3T
in projected productivity gains from agentic AI, based on research across 17M+ companies
99%
of companies plan to deploy AI agents in production. Only 11% have done it.
70%
of European banks already use agentic AI — most without robust governance
€35M / 7%
maximum EU AI Act fine for non-compliance — enforced August 2026
SOURCES — KPMG & World Economic Forum, 2026 · EY Global Financial Services Regulatory Outlook, 2026 · Cambridge Judge Business School Global AI in Financial Services Report, 2026 · EU AI Act Regulation (EU) 2024/1689
When to call us
Agentic transformation brings new challenges. Here are the four we see most often.
Senior teams across European fintech, banking, and insurance are working through these questions right now. If one of them is yours, you are in the right room.
challenge 01
Translating ambition into momentum.
When the strategy is announced but the operating model has not yet moved.
The platforms are bought. The roles are filled. The strategy is on the table. Six months in, the agents are still in pilot mode and the day-to-day has not changed. The work between we have a strategy and we are running differently is where most agentic transformations stall — and where the integration sits.
You may recognise this when
the board is asking for outcomes the programme has not yet produced, agentic decisions keep being deferred, and the distance between the announcement and the operating reality is becoming visible to your peers.
challenge 02
Aligning stakeholders who are all doing their job well.
When the room is full of right answers — and the decision still cannot move.
The CRO is right to want strong oversight. The CTO is right to want productivity. The Chief Data Officer is right to flag data readiness. Each function is doing exactly what it is supposed to do — and the agentic decision still cannot move forward. This is not a failure of any one function. It is an integration challenge.
You may recognise this when
a critical agentic decision is stalled across functions, and you need someone who can hold every perspective at once without belonging to any of them.
challenge 03
Bringing live agents into how the work gets done.
When the technical deployment lands — and the workflow has not yet caught up.
The integration is operating. The agents are live. But adoption is lower than the business case assumed, and the agents are being treated as tools rather than as part of the team. The work to integrate an AI workforce alongside a human one is different from the work to build it — and that is the next chapter.
You may recognise this when
adoption is below target, workarounds have settled in, and the ROI conversation is starting to feel uncomfortable to have at the board.
challenge 04 / priority for 2026
Launching when getting it right the first time matters most.
When every dimension has to align from day one.
A new AI agent product going to market under EU AI Act Annex III. A first regulated agentic workforce. A bank-fintech partnership where agents operate on both sides. A new European market entry where DORA and the AI Act apply at once. These are the launches where every dimension — legal, technical, commercial, people, data — has to be designed in from day one, not retrofitted.
You may recognise this when
you have an AI Act deadline, a launch where every dimension must hold from the start, or a partnership where integration cannot be left to discover itself.
If one of these four feels familiar, you are in good company.
A thirty-minute conversation. No commitment.
How we work
One discipline. Five dimensions. Three phases. Built for what your teams now face.
The frameworks that built modern organisations — ADKAR, Kotter, McKinsey 7-S — remain useful starting points. They were designed for a world where every actor was human, every dimension was managed in sequence, and regulation was a checkpoint at the end.
​
Three of those assumptions no longer hold.
​
Your workforce now includes AI agents acting alongside your people. Legal, technical, commercial, people, and data decisions all move at once. And regulation is a design parameter on day one, not a final review. We built our methodology for that reality — to sit alongside the frameworks your teams already use, not to replace them.
First — the five dimensions
Every engagement holds all five at once. Never one at a time.
i.
Legal & regulatory
EU AI Act, DORA, GDPR, MiCA become operating architecture — not paperwork.
ii.
Technical
Agent capability, integration, observability and what your stack can carry.
III.
Commercial
Business case, productivity, partnership structures, revenue logic.
IV.
People & organisational
Human workforce alongside agentic workforce — decision rights, culture, ethics.
v.
Data & intelligence
AI agents as participants in the data layer — governed, traceable, accountable.
Then — the three phases
A continuous cycle.
Not a linear project.
phase 01
Assess
Map the full picture across all five dimensions before anything moves.
Stakeholder topology
Five-dimension diagnostic
Agentic readiness scoring
phase 02
RESOLVE
Hold the space where alignment becomes possible and conflict becomes decision.
Alignment architecture
Cross-dimension translation
Structured decision design
phase 03
COMMAND
Drive execution until outcomes are real — not just designed or agreed.
Operating rhythm
Outcome intelligence
Adaptive steering
AI agents are the new digital workforce. The IT department of every company is going to be the HR department of AI agents in the future.
Jensen Huang · CEO, NVIDIA
CES 2025 keynote · January 2025
Field Notes
Three patterns we are seeing in European fintech and banking right now.
What we are seeing across the work — drawn from our engagements, the published research, and the conversations we are part of with senior leaders in the sector.
i.
The 99/11 gap — and where the value accrues to those who close it well.
Every European bank, insurer, and fintech leadership team we have spoken with this year is planning to deploy AI agents in production. Few have done it at scale yet. The technology works. The gap is integration: data foundations, governance architecture, accountability chains, workforce design — the structural pieces that have to be in place before the productivity case lands.
​
The organisations that move first on the integration work will be the ones who realise the productivity gains. The structural pieces are easier to design in early than to retrofit later.
"99% of companies plan to put agents into production but only 11% have done so due to implementation challenges related to data, governance and security."
KPMG & World Economic ForumHow AI agents can become strategic partners for business · January 2026
ii.
European banks are running agents — and looking for the right governance architecture.
Around 70% of banks across Europe now use agentic AI to some degree. Many are actively building the governance frameworks that will support agentic decision-making at scale. The question is not whether the agents will be deployed — they already are — but how the architecture of oversight, traceability, and accountability gets designed alongside them.
​
The August 2026 EU AI Act deadline is the formal moment. The teams designing their governance architecture now have more room to get it right than those waiting for the deadline to act.
"More than 70% of banking firms are using agentic AI to some degree, but there is a general lack of robust governance frameworks."
EYGlobal Financial Services Regulatory Outlook · 2026
iii.
Architectural oversight is becoming the question senior teams want to answer well.
Across the financial services research, the loss of human oversight is now cited as a top-three risk by 51% of senior leaders. The standard human-in-the-loop response is being rethought — manual oversight is hard to scale across thousands of agent actions per day, and senior teams are looking for architectural solutions that hold up.
​
The thinking is converging on accountability designed into the system: traceable decision chains, escalation policies, fallback protocols, named owners for each agent class. This is the work that distinguishes the firms moving from pilot to production well.
"51% of respondents cite the loss of human oversight as the third highest AI risk overall. The rapid deployment of agentic AI compounds cyber vulnerabilities, rendering manual oversight increasingly ineffective."
Cambridge Judge Business School
Global AI in Financial Services Report · 2026
What we do
Three ways to engage.
One discipline.
Start here
The Agentic Diagnostic
A clear picture of where your programme is — and where the integration work sits next.
A fixed-scope assessment across all five dimensions. Three working sessions with your team. A written diagnostic and a 90-day roadmap you keep. The recommended way to start a conversation with us.
Scoped delivery
The Build Sprint
Focused engagement on one defined piece of the puzzle.
Designing and shipping a specific agentic capability — a governance architecture, a partnership framework, a regulated AI product readiness. Outcome-defined, time-boxed, scoped fee.
Embedded
The Embedded Partnership
A senior orchestrator alongside your leadership team.
​
For agentic transformation that needs to be held continuously rather than delivered in a project. We sit alongside the CEO, CTO, CRO and board through the cycles — and stay until the work is done.
The Lab
A private network of senior leaders shaping how Europe integrates AI agents into the workforce.
A space for senior leaders integrating AI agents into the workforce to think alongside their peers. Closed-door. Off the record. Working sessions, not panels. Quarterly gatherings in Brussels, Amsterdam, Luxembourg or Paris.
i.
Quarterly working sessions
Four closed-door gatherings per year. 20-30 senior leaders. Chatham House rules. A peer conversation about the work, among the people doing it.
ii.
Members-only research
Reference architectures, anonymised case studies, peer benchmarks — shared between the members and not published outside.
iii.
The Orchestrator Programme
For the senior practitioners among our members who want to develop the methodology in their own practice. First formal cohort 2027.
Why this matters now
The technology has arrived.
The integration is the work ahead.
A bank deploys an AI agent for credit decisioning. The model team builds it. Legal flags it as Annex III under the EU AI Act. The CRO sets the oversight requirements. The COO scopes the productivity case. The Chief Data Officer maps the data dependencies. The board sets the go-live timeline. Each function does what it is supposed to do — and the integration of all five answers becomes the next piece of work that needs an owner.
This is what agentic transformation looks like in practice in 2026. The technology is mature. The economics are compelling. The regulation is real. And the integration — the orchestration of legal, technical, commercial, people, and data into one coherent operating reality — is the work where we sit alongside your teams.
i.
The agentic gap is real and measured.
99% of organisations plan to deploy AI agents in production. 11% have done so at scale. The work between those two numbers — governance, data, workforce, accountability — is where the productivity case lands.
Source: KPMG / WEF, 2026
ii.
European banks are already operating with agentic AI.
70% of European banks now use agentic AI to some degree. Many are actively building the governance frameworks that will support agentic decision-making at scale, ahead of the August 2026 EU AI Act deadline.
Source: EY Global Regulatory Outlook, 2026
iii.
Architectural oversight is now the question senior teams are working on.
51% of senior financial services leaders cite the architecture of human oversight as a top-three priority. Manual oversight is hard to scale; architectural oversight is where the field is moving.
Source: Cambridge Judge Business School, 2026
IV.
The regulatory framework is hard-codified.
€35M or 7% of global turnover is the maximum EU AI Act exposure. DORA penalties run parallel. The first enforcement actions are expected by Q4 2026 — which makes the next twelve months the time to design the architecture properly.
Source: EU AI Act Regulation (EU) 2024/1689
Frequently asked
What you might be wondering.
We are the orchestration practice for European regulated industries integrating AI agents into the workforce. We work alongside banks, fintechs, insurers and regulated technology firms to hold legal, technical, commercial, people, and data dimensions together while deploying agentic AI — across EU AI Act readiness, DORA implementation, AI governance architecture, and workforce integration. Most engagements begin with the Agentic Diagnostic.
Senior leaders — Chief AI Officers, Chief Risk Officers, Chief Data Officers, COOs and CEOs — in European fintechs, banks, insurers and regulated technology firms. Engagements are typically with organisations of 50 to 5,000 employees that have AI agents in production or in active deployment and need cross-dimensional orchestration.
A fixed-scope, two-week assessment across all five dimensions of agentic transformation. The output is a written diagnostic, a five-dimension map of your situation, and a 90-day roadmap. It is the recommended first engagement for any organisation we have not worked with before. See full details.
The high-risk provisions of the EU AI Act apply from 2 August 2026 — including credit scoring, fraud detection, biometric verification, and insurance pricing under Annex III. We integrate AI Act readiness into your operating architecture rather than treating it as a compliance project. We map your obligations across Articles 9, 10, 14, and 43 and translate them into how your systems, data, and people work together.
The Lab is a members-only network of senior leaders in European regulated industries integrating AI agents into the workforce. Quarterly closed-door working sessions in Brussels, Amsterdam, Luxembourg or Paris. Chatham House rules. Members-only research. The Orchestrator Programme — our certification track for senior practitioners — runs alongside the leader network. Membership is by application.
otterdam, Netherlands. We operate across the Benelux region and the wider European Union — primarily France, Germany, and the Nordics. We work in English and French; Dutch as a working language.
AI agents are joining your workforce. We can help you bring them in well.
Two doors. Pick the one that fits where you are.
Membership is by application. A short conversation to assess fit for the network and the quarterly gatherings.