Free Diagnostics • Cited Sources
Nineteen instruments that measure where an organisation actually stands on AI: governance, risk, capability, spend, and legal exposure. Each is scored against a published standard or a named study, and each ends in a written assessment.
Built by Adrian Dunkley: scientist, entrepreneur, and award-winning AI builder. Every diagnostic here started as a question a client asked.
Choose a diagnostic. Most take between five and fifteen minutes.
Answer for how things run. Not for what the policy documents describe.
Open your assessment. Enter your name and email to see the scored breakdown.
Take the one-pager. A summary sized to go straight into a board pack.
The Diagnostics
Each diagnostic scores a defined set of dimensions and returns a written assessment with its findings ranked. The questions, the scoring bands, and the figures behind them come from published research or from a named standard: ISO/IEC 42001, the NIST AI Risk Management Framework, the EU AI Act, MIT's AI Risk Repository. Every source is cited on the tool page, so you can check the reasoning rather than take the score on trust.
Scores a country's case for building its own AI capability across four dimensions: strategic need, fiscal and infrastructure capacity, talent and institutions, and the alternatives available. Returns one of four strategy paths, from Partner First to Sovereign Ready, with the cost benchmarks behind each.
Open the Tool →Weighs data sensitivity, usage volume, in-house capability, and competitive differentiation to establish how much of the AI stack a company should own. Returns a recommendation from Use the APIs through to Sovereign Stack, with the reasoning a board will ask for.
Open the Tool →Places an individual on the four-persona scale Microsoft and LinkedIn identified across 31,000 workers in 31 countries: Skeptic, Novice, Explorer, Power User. Scores understanding, frequency, depth, and habit change, then names the single behaviour that moves you up a level.
Open the Tool →Applies the seven risk domains of MIT's AI Risk Repository, a taxonomy built from a review of 74 existing frameworks, to your own organisation. You rate likelihood and impact for each; the tool plots the matrix and returns a rating and a next step per domain.
Open the Tool →Anthropic's research documented leading assistants changing correct answers under user pressure and mirroring users' stated beliefs. This self-assessment measures the other half of that problem: how much of the agreement you invite and accept, across validation seeking, appetite for challenge, verification habits, and reliance.
Open the Tool →Thirty questions across six dimensions, from how models work to what Article 4 of the EU AI Act now requires of employers. Every question carries a written explanation, and the report breaks the score down by dimension so a team can see where its training gap sits.
Open the Tool →Scores an organisation against the five-level maturity scale used by Gartner and by MIT Sloan Management Review's research with BCG, across five pillars: strategy, data, talent, governance, and delivered value. Names the pillar holding you back, and places you on the scale where most companies stall between experiment and production.
Open the Tool →The ILO's global index puts one in four jobs in some band of generative AI exposure, with transformation rather than elimination the likelier outcome. This scores your own role across task mix, digital output, the part of the job that depends on presence and accountable judgement, and your employer's actual pace of adoption.
Open the Tool →IBM's 2025 breach study ties unapproved AI use to one in five breaches and US$670,000 in added cost per incident. This scores the four things an organisation controls: policy, technical controls, how staff behave in practice, and whether a problem would be detected at all.
Open the Tool →Scores an organisation against the four functions of the NIST AI Risk Management Framework: Govern, Map, Measure, and Manage. Twelve questions rank all four and give each a concrete next step, so you know which function to fund first.
Open the Tool →Applies operational-risk method to AI. Twelve inputs covering size, AI footprint, and controls, calibrated by sector and country, produce a median estimate of the capital exposed to ungoverned AI use, the annual likelihood of a material loss event, and the drivers behind the figure.
Open the Calculator →Deloitte's study of 695 directors across 56 countries found AI absent from 31 percent of board agendas, and discussed at every meeting by only 17 percent. This audits the four things a board controls: how AI reaches the agenda, whether directors can challenge what they hear, how AI spend is approved, and whether AI risk surfaces before it becomes an incident.
Open the Tool →IBM's survey of 2,000 CEOs found only 25 percent of AI initiatives returned the ROI expected of them. This scores the executive rather than the technology: personal fluency, strategic judgement, how well you enable your team, and whether you lead responsible use or assume someone else does.
Open the Tool →The twelve artefacts an auditor, insurer, or enterprise customer asks to see, drawn from ISO/IEC 42001 and the NIST AI RMF: the policy, the owner, the inventory, the controls, the training, and the evidence. Each scores from does not exist to in place and maintained, so you leave with a list of what to produce.
Open the Tool →The SEC and the FTC have both brought enforcement actions against firms marketing AI they did not have, and Gartner estimates only about 130 of the thousands of vendors claiming agentic AI actually offer it. This separates a genuine AI firm from a rebranded IT generalist, flags the language that gives a thin vendor away, and returns the questions to put to them before you sign.
Open the Tool →BCG's rule puts 70 percent of AI value in people and process, which is the part training is meant to change. This scores whether a trainer has built and deployed AI in a real organisation or has only taught it, weighs the evidence behind their claims, and tells you whether to proceed before you commit the budget.
Open the Tool →Gartner expects more than 40 percent of agentic AI projects to be cancelled by the end of 2027, on cost, unclear value, and weak controls. This scores the four foundations agents stand on: documented processes, data and systems that can support autonomous action, supervision you can actually exercise, and a value case that survives a finance review.
Open the Tool →The 25 deliverables of a working AI governance programme, laid out on an interactive 3D board. Rate each from Not Started to Complete for a score built from both your maturity and your completed lines, and a clear view of which deliverables are still missing.
Play the Board →Sizes an AI investment from the outcome rather than the vendor quote. Enter the savings, new revenue, and retention being targeted; the calculator returns the investment that target implies, split into direct AI cost, change management, and a risk provision, with time to value, payback, and return over two, three, or five years. Built on BCG's 10-20-70 rule and IDC's return-on-AI figures.
Open the Calculator →The Regulation runs to 113 articles, and only a few reach any one organisation. This establishes which: whether the law applies to you at all, whether you are a provider or a deployer, and whether your system is prohibited, high risk, or lighter touch. It then audits every duty that follows and returns ranked gaps with the article behind each, the deadlines that apply after the July 2026 Digital Omnibus, and your penalty exposure.
Open the Audit →Sixteen scenarios, most of them deliberately strange, because odd situations stop people reciting the policy they think they are meant to quote. Scores eight dimensions of working practice with AI, from literacy and verification habits to how much of your own thinking you have quietly handed over. Returns a score out of 100, one of nine animal archetypes, and a certificate carrying a code anyone can check.
Find Your Animal →Log weigh-ins and left and right measurements zone by zone, watch the change land on a male or female figure or a photo you upload, and read a dashboard of stats, movement and trend. Pick any date, switch between pounds and kilograms, and download a clean PDF report. Everything stays in your browser.
Open the Tracker →Resource Hubs
Beyond the diagnostics: two dedicated hubs with everything an educator or a parent needs in one place. Each holds an eight-module framework course, working guidelines, a copy-ready prompt library, the research worth reading, and straight recommendations on what to adopt and what to refuse.
The eight-module AI Framework for Schools and Teachers, eight classroom guidelines, a teacher prompt library covering differentiation, rubrics, feedback, and parent communication, seven vetted research sources, and clear adopt-pilot-refuse recommendations.
Open the Schools Hub →The eight-module AI Framework for Parents and Kids, eight house rules that survive contact with a clever twelve-year-old, an age-band ladder from under 8 to 16 plus, a parent prompt library, and allow-watch-refuse recommendations.
Open the Families Hub →Free Download
Worksheet • Dentistry
A one-page worksheet to map your practice's AI opportunities. Turn the book sideways to read. Photocopy or scan, then fill in the boxes with your team. Boxes 01, 04, 05, 06, and 09 are the minimum to complete before you commit to any tool.
Start with the AI Boss Test, the flagship 40-question assessment of your AI leadership across 8 dimensions. Then use the tools above to go deep on the gaps.