Resources
Use these practical guides, canvases and checklists in real product, AI and go-to-market conversations.
Each resource is deliberately lightweight. Take it into a working session, adapt the language to your context and capture the decisions that follow.
Template
Review what an AI agent can see, do and change before it gets real permissions, and decide which actions need a human in front of them.
Use it when: Use this before connecting an agent to production systems, customer data or external communication channels.
Open the toolChecklist
Check whether your evidence would stand up to an independent AI assurance review before you commission one.
Use it when: Use this before scoping an assurance engagement, so you know what claim you want assured and what evidence you can provide.
Open the toolChecklist
Find every place your product generates or manipulates content, and check that marking and disclosure actually reach the person who sees it.
Use it when: Use this to build a transparency inventory for EU AI Act Article 50, or before launching a new generative feature.
Open the toolTemplate
A worksheet for assessing how an AI system could affect people and society, structured so it can be revisited as the system changes.
Use it when: Use this for a first impact assessment, then reopen it when a trigger event occurs.
Open the toolTemplate
A short runbook for the first hours after an AI feature causes harm or behaves unexpectedly in production.
Use it when: Fill this in before launch so that when something goes wrong, nobody is working out ownership in the moment.
Open the toolTemplate
Decide what to watch once an AI system is live, what level of change triggers a review, and who has authority to act.
Use it when: Use this before launch, alongside the release decision, and review it every quarter.
Open the toolChecklist
A stage-by-stage security check for AI systems, organised around the five lifecycle phases used in the UK AI cyber security code and ETSI EN 304 223.
Use it when: Use this at design review and again before release, then revisit it whenever the model, prompts or data sources change.
Open the toolChecklist
Map each of your AI systems against the EU AI Act dates as amended by the 2026 Digital Omnibus, and decide what to start now rather than in 2027.
Use it when: Use this when you need to explain to leadership which AI Act obligations apply to which systems, and when.
Open the toolChecklist
A checklist for screening whether an AI system triggers a DPIA and what safeguards apply to automated decision-making under UK GDPR and the Data (Use and Access) Act 2025.
Use it when: Use this whenever an AI system's output could influence a decision about a specific person.
Open the toolGuide
A master index of the evidence a defensible AI governance decision actually needs, from system description through to a final approval recommendation.
Use it when: Use this to check an evidence pack is complete before a system goes live, or before an audit or regulatory question arrives.
Open the toolTemplate
A working canvas for identifying who could be harmed by an AI system, what type of harm is realistic, and how severe it could be.
Use it when: Use this before testing begins, so the test plan is built around the harms that actually matter for this system.
Open the toolChecklist
A checklist for spotting the legal risks around an AI system that sit outside data protection: intellectual property, discrimination, consumer protection and liability.
Use it when: Use this before deeper legal review, to make sure the right questions are on the table from the start.
Open the toolTemplate
A starting structure for describing what an AI system actually is, before any risk or compliance conversation begins.
Use it when: Use this before a risk review, a procurement decision or a governance sign-off, so everyone is working from the same description of the system.
Open the toolChecklist
A checklist for reviewing an AI vendor or licensing agreement before buying or building an AI system, including RAG and agentic AI tools.
Use it when: Use this before signing a vendor agreement or greenlighting a build-vs-buy decision.
Open the toolChecklist
A checklist for working out where an AI system sits under the EU AI Act's risk tiers, and whether provider or deployer obligations apply.
Use it when: Use this early, since the classification decides which obligations and deadlines apply.
Open the toolChecklist
A self-assessment structured around NIST AI RMF's four functions: Govern, Map, Measure, Manage.
Use it when: Use this to spot the biggest gaps in AI governance before a fuller review.
Open the toolTemplate
Work through a real workflow step by step and decide, deliberately, what AI should and shouldn't own.
Use it when: Use when redesigning a workflow and deciding what AI should own versus a person.
Open the toolTemplate
Prepare for a hard trade-off conversation so it lands as a decision, not a conflict.
Use it when: Use before a conversation where you need to say no, or deliver news a stakeholder won't like.
Open the toolChecklist
A quick diagnostic for catching delivery problems before they show up in the release.
Use it when: Use partway through a sprint, or at a retro, when something feels off but it's hard to name.
Open the toolTemplate
Frame an AI opportunity around workflow, value, trust and the cost of getting it wrong.
Use it when: Use when an AI idea sounds promising but the product case is still vague.
Open the toolTemplate
A starter set of reusable prompts for research, drafting, review and everyday automation.
Use it when: Use when you keep re-writing similar prompts and want a reusable starting point.
Open the toolChecklist
A practical gate for AI features: failure modes, evaluation, oversight and rollback, checked before launch.
Use it when: Use before an AI feature goes live, not after something goes wrong.
Open the toolTool
Match your strengths to real IT and digital roles, and find a concrete next step.
Use it when: Use when you know you want a digital career but aren't sure which path fits.
Open the toolChecklist
Pressure-test audience, positioning, proof and channel choices before launch activity expands.
Use it when: Use before committing budget, channels or a public launch date.
Open the toolChecklist
A working checklist for cutting scope back to what actually needs testing first.
Use it when: Use when a “minimum” version keeps growing back toward the full plan.
Open the toolTemplate
Twenty-one prompts for clarifying the problem, value and smallest credible next move.
Use it when: Use before a discovery interview or early product working session.
Open the toolTemplate
Stress-test a roadmap against outcomes, evidence and the trade-offs it's quietly making.
Use it when: Use before sharing a roadmap with stakeholders, or when priorities start to drift.
Open the toolTemplate
Prioritise testing effort by risk instead of spreading it evenly and thin.
Use it when: Use when testing time is limited and effort needs to go where risk is highest.
Open the toolStart a conversation
Useful first calls usually start with one unclear decision, a deadline and a team that needs a practical next move.
Tell us about it