Skip to main content
Automated Governance Systems — AGSDotted globe beside the letters A G S, the AGS brand mark.AGS
COMPANY

The AGS ConstitutionEquitable, accessible, responsible and ethical AI

The founding commitments that govern how Automated Governance Systems Ltd designs, builds, sells and supports AI-enabled products and services.

Version 1.0 · Dated 16 November 2026

Preamble

Automated Governance Systems Ltd (AGS) exists to make governance an engine for performance: practical, automated and measurable. We build software and services that use artificial intelligence (AI) to help people understand their work, improve it and keep the evidence behind every decision.

AI can widen opportunity or deepen harm. It can make complex work clearer, or hide accountability behind automation. This Constitution records the choices AGS has made about which of those outcomes we work towards. It rests on four pillars: equitable AI, accessible AI, responsible AI and ethical AI. Each pillar supports the others, and none may be traded away to strengthen another.

Pillar I

Equitable AI

AI should share its benefits fairly and must not shift its costs onto the people least able to bear them.

Our commitments

  1. We consider everyone affected by an AI-supported outcome, not only the customer or the person using the product.
  2. We look for bias, unequal error rates and unfair effects in the data, rules and models our products rely on, and we tell customers about the limits we know of.
  3. We measure improvement by its value to people as well as to the organisation, and we do not treat efficiency gained at the expense of fairness as an improvement.
  4. We price clearly and VAT-inclusively, publish what is included, and offer free learning and trial routes so that cost alone does not decide who can learn about responsible AI.
  5. We do not use AI to make or recommend consequential decisions about employment, education, housing, credit, insurance, benefits, healthcare, immigration, legal rights or access to essential services. Those decisions belong to accountable people.

Pillar II

Accessible AI

AI should be usable and understandable by everyone it affects, whatever their ability, language, device or level of technical knowledge.

Our commitments

  1. We design to the Web Content Accessibility Guidelines (WCAG) 2.2 at level AA, test against them, and publish known gaps and how to get help on our Accessibility page.
  2. We write in plain language first, explain acronyms when we use them, and provide captions, transcripts and readable alternatives for audio and video.
  3. We offer accessibility tools such as text sizing, contrast, reading support and read-aloud, and we never make a person's chosen settings harder to keep.
  4. We support multilingual navigation and keep the meaning of important information consistent across languages and formats.
  5. We explain what an AI feature does, what it does not do and when a person should check its output, close to the point where the output is used.

Pillar III

Responsible AI

AI should support human judgement, with clear ownership, proportionate controls and evidence that can be checked.

Our commitments

  1. Every AI-supported outcome that matters has a named, accountable person who can understand, challenge, amend, override or stop it.
  2. Consequential actions, such as releasing a change, approving a supervisory command or publishing on behalf of AGS, require approval by a second person who is not the requester.
  3. We record significant actions, decisions, evidence, ownership and timing in audit trails that are protected from silent editing.
  4. We recommend automation only where it has a beneficial purpose, suitable data, proportionate controls, accountable ownership and a meaningful way for people to intervene. Being technically possible is not enough.
  5. We manage AI risk using recognised frameworks, including the NIST AI Risk Management Framework, ISO/IEC 42001 and ISO 31000. We do not claim certification against any standard unless an independent body has certified it.

Pillar IV

Ethical AI

AI should be honest, respect people's dignity, privacy and autonomy, and never be used to manipulate.

Our commitments

  1. Our AI assistants, including HELIX and our live avatars, say plainly that they are AI. They never claim to be human, clinically qualified or an independent authority.
  2. We separate known facts from assumptions, risks and recommendations, and we never describe anything as compliant, lawful, secure, fair, certified or risk-free without verified evidence.
  3. We collect and use only the personal information we need, handle it in line with UK data protection law, and explain our practices in our Privacy Policy.
  4. We do not use shaming, fear, false urgency, dark patterns or dependency-building techniques in our products, sales or communications.
  5. If someone indicates they may be in immediate danger, our assistants direct them to emergency or crisis services rather than continuing the conversation. AGS products are not a therapy, medical, legal or emergency service.
  6. We credit the work of others accurately, respect licences, and never present third-party material as our own.

Scope

This Constitution applies to every AGS product and service, including the Automated Process Intelligence Platform (APIP), SCADA for AI, AGS Conductor, AGS Automation, Delivery Hub, HELIX, the AGS Academy, the Knowledge Portal and Asset-Based Consulting. It also applies to AGS directors, staff, contractors and partners when they act for AGS.

Where this Constitution and a product feature disagree, the Constitution takes priority and the feature is treated as needing correction.

Human oversight and accountability

AI in AGS products provides information, structure, guidance and recommendations. People make the decisions that matter. AGS directors are accountable for this Constitution and for making sure its commitments are reflected in how products are designed and released.

Changes to live products that affect pricing, access, safety or customer commitments go through a recorded change request and are approved by someone other than the person who raised it.

Evidence and transparency

We show customers what our products record, where AI has been used, and the confidence limits and sources behind consequential recommendations. AI-generated content is labelled so it is not mistaken for a verified record.

We will tell customers about material changes to how our AI features work, and we will not hide limitations to make a sale.

Raising concerns

Anyone may raise a concern that AGS, or an AGS product, has not met this Constitution. Write to enquiry@automatedgovernancesystems.com with the subject 'Constitution'. We will acknowledge the concern, review it with a person who was not involved in the matter, and record the outcome.

Raising a concern in good faith will never count against a customer, learner, partner or member of staff.

Review and amendment

AGS reviews this Constitution at least once a year, and sooner when law, recognised guidance or our products change significantly. Amendments are proposed through a change request, approved by an AGS director who did not propose them, and recorded in the version history.

Amendments may strengthen or clarify the four pillars. They may not remove one.

This Constitution sets out AGS's commitments and values. It is not a certification, an audit opinion or legal advice, and it does not replace our Terms of Use, Privacy Policy, Accessibility Policy or any contract. Where a commitment depends on law or a standard, check the current authoritative source.

Version history

  • Version 1.0 · 16 November 2026

    First edition, establishing the four pillars of equitable, accessible, responsible and ethical AI.

Guidance this Constitution draws on

Read our About page, Accessibility Policy, Privacy Policy and Terms of Use. To raise a concern, email enquiry@automatedgovernancesystems.com.