Mergetic independently checks every AI decision before it takes effect — and keeps the proof.
An evaluation tells you your AI is risky. But once it's live and acting on real customers, who stops the bad output before it lands? Three structural gaps that linear AI pipelines can't close.
When the model that generates an output is also the one that approves it, there's no authority that can overrule it. Correlated judgment can't catch correlated failure.
A bad decision propagates through a workflow before anyone notices. Post-hoc monitoring tells you what already went wrong; it can't stop the output acting.
EU AI Act Articles 12 and 14 presume an independently verifiable record of why an output was allowed. A model narrating its own reasoning doesn't meet that bar.
The supervisory pressure is no longer abstract — and it is Irish. The Central Bank of Ireland is now Market Surveillance Authority for high-risk AI in financial services, and its 2026 Outlook is blunt: firms own every AI outcome — including from third-party systems — and boards are expected to evidence governance, controls and effective challenge. Decision by decision. Not policy documents — evidence.
Four steps, every time: 1 your AI proposes an answer · 2 Mergetic scores it independently · 3 a separate governance authority decides — release, send back, or block · 4 every step is recorded, replayable. Think of it as a control tower for your AI: nothing acts without clearance. Prefer to watch? Open the interactive demo →
The controller proposes.
Governance disposes.
Score clears the threshold and governance approves — the output is released.
Borderline score — routed for a check before it can act.
Below threshold — the output is sent back to be reproduced.
Governance vetoes — the output is held and escalated to human oversight.
Every figure below is captured by permanent telemetry during execution — not estimated after. The full evidence base — four baselines of measured fact — lives in the Evidence Centre.
Measured across the Stage-2 governance benchmark (312 governed decisions), research configuration. Structural characteristics shown; absolute cost and timing figures are shared under NDA as the Tier-2 operational profile.
Runtime governance is easiest to understand through the person who has to answer for the AI. Pick a domain — the problem, what Mergetic does, and the outcome.
The AI made its recommendation, the loan officer accepted it, and the record shows only the outcome — not the reasoning, not the inputs it weighted, not who reviewed it. The CPC 2025 requires her to demonstrate the AI operated within defined rules and that a human was accountable. She can demonstrate neither.
Every AI decision passes through the governance layer first. Every input is logged. Every output is evaluated against the CPC 2025 RulePack. Every decision carries a confidence rating, a risk flag, and a complete record of every reasoning step.
Siobhán shows the regulator a complete decision record within the hour — what data the AI considered, what weight it gave each factor, what score the output achieved, and the human reviewer who confirmed it before it was applied.
Measured, not assumed. Every decision scored, routed and recorded with its full reasoning trail. Walk a financial-services case that permits — and a high-risk case where governance vetoes the controller.
Judge separated from generator
Per-decision binding veto
Hash-chained, tamper-evident
Any model, any cloud
Mergetic's architecture embeds governance capabilities that map directly to the EU AI Act's core obligations for high-risk AI systems.
Timeline: under the Digital Omnibus, formally adopted in June 2026, high-risk obligations — including risk management (Art. 9), traceability (Art. 12), human oversight (Art. 14) and post-market monitoring (Art. 72) — apply from 2 December 2027 for stand-alone Annex III systems, and 2 August 2028 for AI embedded in regulated products. Article 50 transparency obligations still apply from 2 August 2026, penalty exposure is unchanged, and supervisors already expect evidence of governed AI today. Deferred — not diluted.
Continuous evaluation and scoring at every pipeline stage. Composite risk metrics trigger automatic escalation or blocking before an output can propagate.
Full execution logs captured in the Memory Spine — every reasoning path, every synthesis decision, every governance override, auditable end-to-end.
Configurable escalation triggers route edge cases to human reviewers. The system supports human decision-making authority — it does not replace it.
Mergetic provides architectural capabilities that support regulatory alignment. This does not constitute legal advice or certification. Organisations should consult qualified legal counsel for compliance assessment.
Two products, bought separately, sharing one architecture: the Reliability Benchmark tells you where your AI’s risk is; Mergetic governs it in production — a binding veto before each decision acts. Both are available now.
We make AI accountable before it acts. Mergetic is live — early customers work directly with the founder and shape the roadmap. Be among the first.
Mergetic is licensed annually by decision exposure: sized to the number, consequence and scale of the AI workflows it governs — never by seats, tokens or model calls. Pricing is scoped in the walkthrough. A fixed-fee Governed Workflow Pilot is available, credited toward your first annual licence.
We use your details only to respond, and never share them. See our Privacy Notice. Prefer email? john@mergetic.com