Structure the facts
Separate organisation-specific facts from reusable policy templates so information has a clear home.
Policies, data & agent governance
A shared foundation for privacy, cookie, consent and data-handling information, with a research direction towards accountable AI agents.
Live policy platform · agent governance in developmentProduct strategy · Structured content · Multi-tenant publishing · Governance R&D
01 / The problem
Websites change faster than their policies. Suppliers, integrations and data-handling choices move on, while the information explaining them is often maintained manually in several places.
Trust Center grew from our Consent ePolicy work. Running that service exposed the manual effort involved and the gaps between what third-party integrations did and what a website said they did. We wanted a more connected foundation.
02 / The work
Separate organisation-specific facts from reusable policy templates so information has a clear home.
Deliver legal and privacy hubs on each organisation’s own domain, using the same underlying platform.
Develop a path from observed data-flow changes to a reviewed decision, an agent control and an approved policy update.
03 / The system
The publishing foundation separates shared policy structures from the facts belonging to each organisation. The research extends that connection towards the behaviour of agents and integrations.
The business, its contacts, services and processing choices.
Reusable templates and clauses, with organisation-specific content.
A legal and privacy hub on the organisation’s own domain.
Administration and review of the information being published.
Cloudflare Workers and a shared database support the publishing service. The proposed monitoring-to-control loop is a further research programme.
04 / Research & development
Structure the facts behind public notices and investigate how a changed supplier, tool or data flow could trigger a review of the affected requirements.
A policy update needs a traceable source and an approval decision. Publishing machinery alone cannot establish that the underlying practice is compliant.
Develop a bounded demonstration that detects a changed data flow, assesses regional requirements, controls the agent’s response and updates approved policies with an evidence trail.
Observed traffic, provider statements and verified processing locations are different kinds of evidence. Unknowns and conflicting requirements must remain visible.
Plan repeatable scenarios for changed permissions, unavailable tools, uncertain destinations and adversarial content. Compare task completion with the decisions to allow, hold or block an action.
Security hardening and evaluation are part of the work. The test is whether a control reduces risk while still allowing useful work to proceed.
05 / Where it stands
The policy-publishing platform operates today. It provides the foundation for our next research question: how to connect changing system behaviour with permissions and evidence that people can assess.
The existing policy platform gives this research a practical starting point. The next demonstration is deliberately bounded to one website and one connected agent workflow before broader use is claimed.
Security hardening, scanner readiness, regional data-flow assessment and the agent-risk demonstrator are ongoing development. These are not presented as validated controls or a guarantee of legal compliance.
Our track record
Client delivery. Products of our own. The everyday work of making them run. Each gives us a different perspective on what makes technology useful.
Explore our delivery track recordOur team’s work connects brands, creators and educators with their audiences, through digital platforms and the systems behind them.
We live with the products we create. Supporting them beyond launch informs the decisions we make on the next one.
Recurring problems become reusable technology. The questions still unanswered become the starting point for our research.