Veritasys

AI governance & assurance for the UK public sector

Trust in every algorithm you deploy.

Veritasys helps councils and public bodies turn AI pilots into governed, transparent and defensible services, with evidence your board, your auditors and your residents can follow.

Choose where to start See how it works
  • ATRS
  • UK GDPR · Arts 22A–22D
  • ISO/IEC 42001
  • AI Playbook
AI assurance suppliers in the UK, 2024 estimate
524
Source ↗
Principles in the AI Playbook for the UK Government
10
Source ↗
Working days to come back to you with a scoped proposal
2 days
How it works →
From purchase order to board-ready report
4 weeks
How it works →

Where most clients start

AI Governance Health Check

Four weeks · up to 5 AI systems · one clear plan

  • AI system inventory (up to 5 systems or use cases)
  • Risk register scored on likelihood and impact
  • Control mapping: ATRS, UK GDPR and ICO guidance, ISO/IEC 42001 Annex A, the AI Playbook
  • Deep-dive on your highest-risk system
  • One ATRS-ready draft record
  • Board-ready summary and a 90-day action plan
Start a conversation What's included

Why now

The rules for public-sector AI are arriving at once.

  1. Transparency

    ATRS records are mandatory for central government

    Central government departments and in-scope arm's-length bodies must publish Algorithmic Transparency Recording Standard records for in-scope tools. Councils are publishing them voluntarily, and members and residents are starting to ask for them.

    Read the scope policy

  2. Automated decisions

    New safeguards for significant automated decisions

    The Data (Use and Access) Act 2025 replaced UK GDPR Article 22 with Articles 22A to 22D: safeguards for significant decisions taken solely by automated means, including the right to human intervention.

    Section 80 on legislation.gov.uk

  3. Regulator guidance

    A statutory ICO code on AI and automated decision-making

    Regulations require the Information Commissioner to prepare a code of practice on AI and automated decision-making. Organisations that map their AI use now will be ready when it lands.

    SI 2026/425

Approach

Five steps. Four weeks. One clear answer.

  1. Week 0

    Agree

    Scope, data terms (OFFICIAL only), named contacts and a purchase order. An agreed scope, no surprises.

  2. Week 1

    Discover

    Kick-off, document request and up to six interviews. We build the AI system inventory with your teams.

  3. Week 2

    Assess

    Each system is mapped against ATRS, UK GDPR and ICO expectations, ISO/IEC 42001 Annex A and the AI Playbook. Risks are scored.

  4. Week 3

    Verify

    A deep-dive on your highest-risk system: decision flow, human review, supplier terms and evidence.

  5. Week 4

    Report

    Findings with evidence and ratings, an ATRS-ready draft record and a 90-day plan, played back to your board.

Our stance

Assurance, not alarm.

We are practitioners, not a law firm. We work alongside your information governance, digital and legal teams to make AI use visible, proportionate and accountable.

  • Tool-agnostic: we assess what you use, we do not sell software
  • Evidence first: every finding traces to a document, an interview or a test
  • Plain English: written for members and residents as well as specialists

Start with one conversation

Is your AI ready for scrutiny?

Tell us where you are and what you need. We will come back within 2 working days with a scoped proposal. No commitment.

Choose where to start