The Participant pathway, end to end.
The Reg 10 Accelerator is a Participant sandbox — not an ACB scheme. regulation10.ae produces the evidence stack the Accelerator Testing Team expects, with §1 and §2 covered line by line.
Participant pathway
Four steps from onboarding to Output Report
The Accelerator Framework treats the Participant as the testee and the Accelerator Testing Team (ATT) as the reviewer. There are four checkpoints between onboarding and the Output Report. regulation10.ae produces a deterministic artefact at every one of them.
Step 01
Eligibility and abstract
We capture the Participant abstract — entity, system, purpose, processing categories — through the onboarding wizard and render it on the public AI System Register. That alone clears Accelerator §1 G1.2 and G1.5.
Step 02
Design dossier
Modules 1.1 (risk classification), 1.1b (Article 5 prohibition screener), 1.2 (DPIA) and 1.3 (transparency notice EN+AR) generate the redactable design document set required by §1 G1.3 — citation-pinned, content-hashed, regenerable.
Step 03
Assessment-model declaration
DPIA §1 names the lens — Open Loop, OECD AI Principles, EU AI Act Article 27, or ISO/IEC 42001 — and lets us cleanly express the Bespoke-criteria pathway under §1 G1.6.
Step 04
Output report
Module 4.1 evidence pack plus Module 3.3 WORM audit trail compose the single signed Output Report the ATT and Commissioner expect under §1 G1.10. Re-runnable in under ten minutes if facts change mid-testing-window.
Coverage matrix
§1 Governance and §2 Substantive criteria — line by line
Mapping of the DIFC Reg 10 Accelerator Framework requirements to the regulation10.ae surface that satisfies them. Source of truth: the internal coverage-verification annex dated 12 May 2026, pinned to commit 44fe046.
Frame
No ACB accreditation needed to enter.
The ACB scheme and the Accelerator are different instruments
DIFC Reg 10.3.3 vs Accelerator Framework §1
The DIFC Reg 10 ACB Framework defines an Accredited Certification Body as an entity with the ability to certify a Certification Applicant's System for High Risk Processing. Part 1 of that framework is what an ACB applicant has to prove — Independence, Expertise, Written Undertakings, Established Procedures, Conflicts of Interest, and so on.
The Accelerator Framework is different. It is a sandbox for Participants — developers, deployers and operators — to test their own systems for privacy-by-design compliance. The Accelerator Testing Team is the reviewer, not an ACB. The framework explicitly lets Participants pick the assessment model (Open Loop / OECD / Humane Intelligence / Bespoke) and can be run as a bolt-on to UAE Reg Lab, UK ICO or EU AI Act Article 57 sandboxes.
Nowhere in the Accelerator Framework does a Participant need to be — or use — an ACB. ACB certification is what an entity does for commercial deployment under Reg 10.3.3. It is not what an entity does to enter the sandbox. The entire Part 1 of the ACB Framework (Independence, Expertise, Written Undertakings, Established Procedures, Transparency, Conflicts of Interest, Communications with the Commissioner, Certification Application Review, Certification Programme Requirements Review) is orthogonal to Accelerator participation.
The only place the ACB criteria intersect with the Accelerator is on the tester side, in abbreviated form, as part of ATT qualification. That is the Commissioner's selection problem, not the Participant's.
Bespoke-criteria readiness · v2.4 — 24 May 2026
Modules 2.1 (Bias) and 2.2 (Threat Model) shipped in v2.4.
Accelerator Framework §2 reserves the right of the Accelerator Testing Team or Commissioner to impose Bespoke criteria — for example quantitative bias metrics or STRIDE threat models. Modules 2.1 (Bias / Fairness) and 2.2 (Threat Model) shipped in v2.4 (24 May 2026), so the Bespoke pathway is met out of the box. Module 2.3 red-team report remains bundled into the Reg 10 Sprint engagement; Module 3.1 drift monitoring is provisioned on first Enterprise tenant signup.
Ready to file?
Walk into your Accelerator Testing Team session with the §1 / §2 matrix in hand.
A 30-minute readiness call maps your system facts onto the matrix, flags any §2 row that needs a Bespoke-criteria contingency, and gives you a one-page submission summary.