Building AI you can trust
AI is changing how organisations decide, operate and compete. But its real value depends on how well it's governed. We help organisations use AI that is safe, transparent and accountable, so the people responsible for it stay firmly in control.
Join usOur mission
Confidence in every AI decision
AI is the defining technology of our time. Yet it only delivers when people can trust it: when its decisions can be explained, its performance can be measured, and its risks can be managed. Our mission is to give organisations that confidence, so they can scale AI to solve real problems without losing oversight. We do this through three commitments.
We're building the governance layer that lets organisations put AI to work with confidence: powerful, transparent and accountable.
Putting teams in control
We give teams one place to register, monitor and manage every AI model they run. Every change is recorded, and every model is tied to a clear owner and business goal.
Making AI understandable
We make AI decisions clear to everyone, not just data scientists. Teams can see why a model produced an outcome, spot when something looks wrong, and feed back to improve it.
Holding AI to account
We make AI traceable and evidence-ready, so every decision can be reviewed and every audit or regulator request can be answered with confidence. We do this in partnership with the EW-AiRM™ framework, embedding enterprise-wide AI risk management directly into the MA3STRO stack.
Our coalition
MA3STRO
Born as a Horizon Europe bid. Kept as the coalition that builds the agentic AI governance stack.
Multi-Agent Enabled Systems Through Robust Orchestration
MA3STRO is a European coalition of partners, co-led by AI & Partners, building a governance stack for agentic AI. It began in March 2026 as a consortium answering a Horizon Europe call on next-generation AI agents. The partners chose to keep working together after the bid, because the problem it set out to solve has only become more urgent: AI agents can only be trusted at scale if every action they take can be monitored, validated and audited.
Why it was set up
- 01
The call asked for it
Horizon Europe asked for autonomous, robust and reliable AI agents, with monitoring and validation built in.
- 02
The thesis behind it
Multi-Agent Enabled Systems Through Robust Orchestration. Agents only scale if every action is auditable.
- 03
Why it outlived the bid
It is a real proposition, not just a grant application. With the EU AI Act now applying, the need for auditable agents is here today.
How it came together
Mar 2026
Consortium formed for the Horizon Europe call on next-generation AI agents (HORIZON-CL4-2026-05)
Apr 2026
MAESTRO proposal submitted: a 36-month research and innovation action across five work packages
Jul 2026
The partners decide to continue beyond the funding round
Jul 2026 onwards
Renamed MA3STRO. Same partners, now a standing coalition
Oct 2026
Governance, embedded. MA3STRO partners with De-Risking Solutions Ltd to embed the EW-AiRM™ framework into the MA3STRO governance stack.
In partnership with EW-AiRM™
Governed by design
In October 2026, MA3STRO entered a master service agreement with De-Risking Solutions Ltd to embed EW-AiRM™, the Enterprise-Wide AI Risk Management framework created by Prof. Markus Krebsz, into the MA3STRO governance stack. Every agent orchestrated through MA3STRO can now be assessed, filtered and monitored against a single enterprise-wide risk architecture.
Orchestration needs a score
A maestro does not improvise; they conduct from a score everyone can read. In MA3STRO, that score is EW-AiRM™: an enterprise-wide AI risk management framework spanning six strategic pillars across the strategic, operational and resilience layers, grounded in the MIT AI Risk Repository's risk taxonomy.
From principles to pre-deployment
EW-AiRM™'s HAiPECR filter gives MA3STRO a structured pre-deployment gate, so agents are screened before they act, not audited after they fail.
A named human, always
At the heart of the framework sits accountability that regulators can touch: every AI system traceable to a named, accountable human. Within MA3STRO, this closes the loop the EU AI Act now demands of agentic systems.
EW-AiRM™ governance
MA3STRO orchestration
AI agents