For HR, L&D and AI transformation leads

An AI readiness assessment that shows who can use AI well

AISA is the data intelligence layer for the human side of AI transformation.

AI is not a self-driving car. It needs a good driver. The AISA AI Fluency Score™ measures how well each employee actually works with AI, from what they do rather than what they say about themselves. Results roll up by team and function, so you see who is ready, where the gaps are and where training will pay off.

Leave your work email and the 90-second dashboard tour starts straight away.

  • Evidence behind every score
  • Hosted in the EU
  • Never used to train AI

AISA AI Fluency Index

August 2026 edition · 1,800 professionals

Benchmark

Average AISA AI Fluency Score™

48/100

The average sits in the Developing tier (28–59). Where does your workforce sit?

3,000+

assessments to date

98%

same or adjacent tier on retake

18.5 pts

average self-overrating (n = 274)

Used by global enterprises with workforces up to 60K

Why AISA

AISA is the most sophisticated, validated and credible way to measure AI fluency in organizations

AISA's AI readiness assessment is a structured behavioural interview, the method research ranks first for predicting job performance, with every score drawn from evidence and given against a published rubric. Each reason below is one you can check.

  • 18.5pts

    A behavioural interview, not a survey

    Self-ratings run 18.5 points high. AISA scores what people have actually done.

  • Skill shown in action

    Live exercises show how each person really works with AI.

  • 98%

    Objective and repeatable

    98% land in the same or an adjacent tier on retake.

  • 93%

    Grounded in recognised science

    Covers 93% of Anthropic’s AI fluency markers and all 25 US Department of Labor sub-competencies.

  • ISO 10667

    Built to testing standards

    Designed to ISO 10667 and the AERA/APA/NCME Standards for Educational and Psychological Testing.

  • Evidence behind every score

    Every score traces to what the person said and did, and is checked before release.

  • Integrity-checked

    Every session is checked for answers that are not the person’s own.

  • 3,000+

    Proven at scale

    3,000+ assessments behind the scale and the benchmark.

  • Always up to date

    Refreshed every week with the latest AI models and tools.

The reliability data, the standards and the research behind the method are in the AISA methodology.

The people side

Organizational AI readiness depends on your people

A typical AI readiness framework covers five areas: strategy, data, infrastructure, culture and people. The first four can be bought, built or written down. The fifth decides whether any of it pays off, because AI only creates value when someone uses it well.

Prosci, the change management research firm, puts it plainly: the success of AI initiatives “depends far more on whether the organization’s people are prepared to adopt, trust, and use it than on the AI technologies themselves.”

It is also the hardest to see. Most AI readiness checklists rely on self-report, and self-report flatters: across 274 professionals, people overestimated their own AI skill by 18.5 points on average.

AISA measures organisational AI readiness on the people side from evidence: what each person has actually done with AI, and what they show in live exercises. The result is a measure of workforce AI readiness you can act on: who can use AI well, where each team's gaps are, and where training will move furthest.

The experience

What the AI readiness assessment is like for your people

Each person talks with Aisa, an AI interviewer, for 20 to 40 minutes in their own time. Aisa starts from the work they actually do, follows up on what they say, and asks them to show the skill in short live exercises. The questions adapt to every role, and every answer is scored against the same published rubric.

9:41
AISA
AISAAI Fluency Assessment
19:36
Solid Report
BE
Aisa: Tell me about the last piece of work where AI did real heavy lifting for you.
Employee: I used it to draft our quarterly client report from the raw sales data.
Aisa: Before it went out, what did you check, and how?
Employee: I traced every figure back to the source sheet. It had invented one growth number, so I fixed it and told the team.
Aisa: Live exerciseHere’s one to try. Your intern used AI for tomorrow’s board deck and hits send at 3pm. You open it, and something catches your eye…
Share your thoughts...

Evidence in the report

Critical Thinking: traced figures to the source and caught an invented number

An illustrative example. Every real conversation follows the person's own work.
  1. Real work, not hypotheticals

    It starts from what each person actually does with AI in their own role, so a marketer and an engineer get different questions on the same scale.

  2. Follows up like an expert interviewer

    Vague answers get probed and specifics earn credit. Confident talk without specifics does not score highly.

  3. Every score traces to evidence

    Each skill score links to what the person said and did, so anyone reading the report can see exactly why.

  4. Skill shown, not described

    Short live exercises: catch the flaws in an AI draft, improve a weak first answer, choose the right tool, re-plan when a step breaks.

  • 20–40 min

    in each person’s own time

  • Any language

    typed or spoken

  • Weekly

    refreshed with the latest AI tools

  • 89.5%

    of people finish it

In their words, at the end of an assessment

  • This was probably my best AI conversation.

    Digital Marketing Strategist

  • It was helpful to reflect on how I use AI as a QA partner while still keeping human judgment at the center.

    QA Manager

  • It was a good experience to go through what I did and what I actually know while answering — this was a very good conversation.

    AI Engineer Trainee

Same person, two instruments

AI readiness survey, quiz or assessment: what each one can see

Take one finance manager. An AI readiness survey and AISA's AI readiness assessment both report on her. The survey signs her off as ready. The assessment finds where she is putting client data at risk, and what to train first.

AI readiness survey

Self-rated
How confident are you using AI at work?
4 out of 5
How often do you use AI tools?
Every day
Have you completed AI training?
Yes
ResultAI-ready

Self-ratings run 18.5 points above demonstrated skill on average.

AISA AI readiness assessment

Evidence-based

54

AISA AI Fluency Score™ · Developing

  • Strength · Workflow & Application

    Breaks month-end reporting into steps and has AI draft each one.

  • Risk · Safety & Responsibility

    Pastes client contracts into a free chatbot to summarise them.

Train first: safe data handling, before AI is rolled out further in her team.

Every line links to what she said and did in the conversation.

An illustrative example.

An AI readiness survey asks people how confident they feel. A quiz checks what they remember. An internal test and a manager's read each see part of the picture. Here is what each method can tell you about the people side of AI readiness.

How the AISA Score, a self-rating survey, a quiz, an internal test and a manager’s read compare on the people side of AI readiness
What you needAISA ScoreSelf-rating surveyQuiz or multiple choiceInternal testA manager’s read
Shows what people do with AI, not what they believeYesNoNoPartlyPartly
Evidence behind every scoreYesNoPartlyPartlyNo
Adapts to each person’s roleYesNoNoPartlyYes
One scale across teams and over timeYesPartlyYesPartlyNo
Benchmarked against an outside populationYesNoPartlyNoNo

The standard

The AISA AI Fluency Score™, the standard AI readiness score

Lenders read a credit score. Driving examiners issue a licence. Both work because everyone is measured the same way, on a scale people trust. The AISA AI Fluency Score™ does that job for AI: it is the driving test for working with AI, and the AISA certificate is the licence. As an AI readiness score it runs from 0 to 100, every score is given against the same published rubric, and it rests on 3,000+ assessments.

Two in three professionals score below Proficient. 1.3% reach Expert.

AISA AI Fluency Index, August 2026 edition
  • 20.2% Emerging
  • 46.7% Developing
  • 24.7% Proficient
  • 7.1% Advanced
  • 1.3% Expert

Six questions to ask any AI readiness vendor, with our answers

  • Will a score hold if the same person takes it again?

    98%

    land in the same or an adjacent tier on retake

    Scores correlate at r = 0.79 between attempts.

  • Is it mapped to recognised AI fluency research?

    93%

    of Anthropic’s AI fluency markers covered

    The rubric also covers all 25 US Department of Labor AI literacy sub-competencies.

  • How much real data sits behind the scale?

    3,000+

    assessments, refined over about 50 iterations

    Proven at scale, and ready for organisation-wide decisions.

  • Will our people actually finish it?

    89.5%

    completion rate

    People finish it, and leave with a report and a certificate.

  • Is the method backed by research on job performance?

    r = .42

    structured interviews, the top predictor of job performance

    From Sackett et al.’s 2022 re-analysis. AISA is a structured behavioural interview, run by AI.

  • Which standards is it built to?

    ISO 10667

    the assessment-delivery standard the method is designed to

    Alongside the AERA/APA/NCME Standards for Educational and Psychological Testing.

Ask other vendors these six. Then bring them, and your own, to our demo.

AISA is a member of UKAI, the UK's AI trade association. It publishes the quarterly AISA AI Fluency Index and 147 in-depth AISApedia guides, and was on stage in October 2026 at the Data Science and AI Summit London and Digital Leaders' AI Week. Read the methodology and the third-party reviews of AISA.

Measure · Improve · Prove

What the AI readiness assessment tool shows you

Every conversation becomes data you can act on. Individual results roll up into team, function and organisation views as each assessment completes, and each view answers one of the questions a board asks about AI.

  • 79%of employees say the training they get isn’t fully personalised.

    Docebo, The AI Readiness Gap 2026

    Measure

    Set a baseline before AI transformation programmes and training: every employee, team and function on the same 0–100 scale.

  • 85%of employees say their AI training doesn’t help them use AI in their role.

    Docebo, The AI Readiness Gap 2026

    Improve

    Aim the training budget, and the type of training, with individual, team and function analytics. Segment your people, find the AI sceptics, and use the AI Enthusiasm Score to see who is most likely to learn.

  • 72%of enterprise leaders now formally measure their return on AI.

    Wharton and GBK Collective, 2025

    Prove

    Re-assess after training and show the board and your AI transformation team how far AI fluency has moved. Measuring AI adoption by logins shows who opened the tool; the AISA Score shows who uses it well.

Where does our organisation stand?

One live AI fluency number for the organisation, next to the industry average and an AI-fluent organisation benchmark.

Real dashboard · demo data

The AI Fluency Index gauge next to the industry average and an AI-fluent organisation benchmark

Questions about a rollout? Walk through the dashboard with us.

Prefer to look first? Take the 90-second dashboard tour on AISA for Teams.

From person to board

Employee AI readiness, from each person to the board

Employee AI readiness is where every number starts. Each person gets their own result, and the results roll up automatically, so AI talent readiness is visible at every level of the organisation. The same roll-up powers our AI fluency assessment for teams.

  1. Level 1

    Person

    An AISA AI Fluency Score™ from 0 to 100, five dimension scores, an AI persona and a personal growth plan.

  2. Level 2

    Team

    Team average, spread, strongest dimension and the Achilles heel, updated as each assessment completes.

  3. Level 3

    Function

    Functions side by side, so the training budget goes where the gaps are.

  4. Level 4

    Organisation

    One AI fluency score for the whole organisation, benchmarked and tracked over time.

Each person assessed keeps their full report, private to them, and a verifiable certificate to add to LinkedIn, so the assessment gives back while it measures. See a sample report.

Security

Data security for an enterprise AI readiness assessment

  • Hosted in the EU

    Assessment data is hosted in the EU (AWS eu-west-1), and the application runs in Dublin, next to it.

  • Never sold, never used for training

    Data is used only to produce the report. It is never sold, shared with third parties or used to train AI models.

  • Deleted within 30 days

    People can ask for deletion at any time, and AISA deletes within 30 days.

  • Your records stay yours

    An organisation’s records are visible only to its own owners and admins.

  • Audited suppliers

    Built on suppliers with independently audited certifications, including SOC 2, ISO 27001 and ISO/IEC 42001.

  • Integrity-checked

    Every session is checked for answers that are not the person’s own.

The framework

An AI readiness framework for people, in five dimensions

Across the EU, driver training follows the GDE matrix: control of the vehicle, mastery of traffic, the goals and context of each journey, and the driver as a person. AISA reads AI the same way. Its five dimensions cover the machine, the road, the vehicle, the journey and the person, with 11 skills scored underneath them against the published AISA rubric.

  1. 01 · Handling the machine

    Prompting & Communication

    How people brief AI, give it the context it needs and work an answer through several rounds.

  2. 02 · Reading the road

    Critical Thinking

    Whether they check what AI gives back, and know where it tends to go wrong.

  3. 03 · Knowing the vehicle

    Technical Understanding

    Enough about how AI works to choose the right tool for the task and know its limits.

  4. 04 · The journey

    Workflow & Application

    How far AI is built into real work, with bigger tasks broken down so AI can help.

  5. 05 · The person

    Safety & Responsibility

    What data goes in, what risks come out, and who is affected downstream.

Checklist

A workforce AI readiness checklist

Thirteen questions to answer about your people before you invest in AI training, each tied to the framework above. Self-report gives a confident answer to every one of them. Evidence gives an accurate one, and AISA collects that evidence person by person and team by team.

Handling the machine

Prompting & Communication

  • Do people give AI the context it needs, rather than one-line requests?
  • Do they work an answer through several rounds instead of taking the first one?

Reading the road

Critical Thinking

  • Do they check AI output before it reaches a client, a report or a decision?
  • Do they know where AI tends to be wrong?

Knowing the vehicle

Technical Understanding

  • Do they understand enough about how AI works to pick the right tool for a task?
  • Do they keep up as the tools change?

The journey

Workflow & Application

  • Is AI built into real workflows, or saved for one-off questions?
  • Do people break bigger tasks into steps AI can help with?

The person

Safety & Responsibility

  • Do people know which company data can go into AI tools?
  • Do they think about who is affected by what AI produces?

Across the organisation

Leadership view

  • Do you know which teams and functions are ahead, and which are behind?
  • Do you know who your AI champions and your AI sceptics are?
  • Do you have a baseline before training, and a way to re-measure after it?

Readiness vs maturity

AI readiness vs AI maturity assessment

The two answer different questions. An AI maturity assessment, usually built on an AI maturity model, rates the organisation's systems: strategy, governance, data and infrastructure. An AI readiness assessment on the people side asks whether your workforce can use AI well, and AISA answers it from evidence.

AI maturity assessment compared with AISA's AI readiness assessment
AspectAI maturity assessmentAI readiness assessment (AISA)
Looks atThe organisation: strategy, governance, data and infrastructureThe people: how well each person works with AI
MethodStakeholder interviews, documentation review, a maturity modelA 20 to 40 minute conversational assessment per person, scored from evidence
OutputA maturity level and strategic recommendationsAn AISA AI Fluency Score™ for every person, team and function, with training priorities
Best forCIO and CTO strategic planningHR, L&D, Chief People Officers and AI transformation leads

The two are complementary: organisational maturity without individual capability is strategy without execution. For a view of the wider market, see AI readiness assessment tools compared. If your readiness work includes AI literacy for the EU AI Act, see the AI literacy assessment.

Get started

How to run an AI readiness assessment across your organization

You can start with one team and grow from there. If a single department goes first, a team AI assessment gives you a baseline and a template for the rest of the rollout.

  1. 1

    A demo, with your questions

    We walk you through the dashboard with demo data and talk through how a rollout would work for you.

  2. 2

    A pilot team

    Write down your own read of each person first. Then invite the team by email or a team link, and compare the results with your notes.

  3. 3

    The whole organization

    Every admin works from one dashboard. Re-assess after training to show what moved.

Before the next AI training budget is signed off, see who actually needs it.

In 30 minutes we walk you through the dashboard with demo data, show you the full report each person receives, with the evidence behind every score, and talk through how a rollout would work for you.

AI readiness assessment FAQ

What does an AI readiness assessment measure?

On the people side, an AI readiness assessment measures how well your workforce can actually use AI at work. AISA scores 11 skills across five dimensions (Prompting & Communication, Critical Thinking, Technical Understanding, Workflow & Application, and Safety & Responsibility) in a 20 to 40 minute conversation with exercises. Every score is backed by evidence from what the person said and did, and rolls up into team, function and organisation results.

How is this different from an AI maturity assessment?

An AI maturity assessment rates the organisation’s systems: strategy, governance, data and infrastructure. AISA’s AI readiness assessment measures the people who have to use AI, person by person, and rolls the results up to team, function and organisation level. The two work together: organisational maturity without individual capability is strategy without execution.

How long does the AI readiness assessment take?

Each assessment takes 20 to 40 minutes per person. It is a conversation with Aisa, an AI interviewer, taken in each person’s own time, so a whole team can complete it over days or weeks without blocking calendars. Your dashboard updates as each assessment completes.

Can I measure AI readiness before and after training?

Yes. Every assessment is scored against the same published rubric, so a re-assessment after training is a like-for-like comparison, person by person and team by team. Scores hold up on retest: 98% of people who retake AISA land in the same or an adjacent tier (r = 0.79 between attempts). The AISA AI Fluency Index adds an outside benchmark: an average of 48/100 across 1,800 professionals in the August 2026 edition.

If every conversation is different, can scores be compared?

Yes. Questions adapt to each person, but every score is given against the same criteria and the same published rubric, and every score is tied to quoted evidence. Two people who had different conversations are measured on the same scale.

Does a conversation favour people who are articulate?

No. Scores rest on what people have actually done and shown, including in hands-on exercises, not on how polished they sound. The scoring credits people whose demonstrated skill runs ahead of how they describe it, and people can answer by typing or by voice dictation in their own language.

Can someone get AI to answer for them?

Every session is integrity-checked for answers that are not the person’s own, and follow-up questions test whether people can explain and build on what they said. Flags appear in the report, and no certificate is issued for flagged work.

What happens to the score when AI models change?

A score means the same thing over time. Every score is anchored to quoted evidence and to a published rubric that defines each level. The conversation is refreshed every week with new AI developments, while the rubric and scoring criteria stay fixed.

How reliable is a conversation-based score?

98% of people who retake AISA land in the same or an adjacent tier, and scores correlate at r = 0.79 between attempts. The method is designed to the AERA/APA/NCME Standards for Educational and Psychological Testing and to ISO 10667, the international standard for assessment service delivery.

Are there third-party reviews of AISA?

Yes. Third-party reviewers have examined AISA’s method and taken the assessment themselves. TechBullion names AISA best overall for measuring AI fluency, FoundrList finds it the most rigorous and evidence-based approach, and Mehmet Yalcin, Director of Products at The Guardian, tested it with a product team. Every review is linked at aisa.to/reviews.

Is one conversation enough?

The conversation covers all five dimensions and includes exercises in which people show their approach. It is also repeatable, so organisations measure again after training and see what changed, team by team.

Is it only for technical roles?

No. It measures how people work with AI in any role, from finance and operations to engineering, and adjusts its technical depth to the person. Assessments taken in other languages score as high as those taken in English.

Why not build an AI readiness test in-house?

A basic AI literacy test is quick to build. A behavioural, conversational assessment with a score that is accurate, reliable and trusted is a much bigger job. AISA has been refined over about 50 iterations and 3,000+ assessments, on a published rubric mapped to Anthropic’s AI fluency research and the US Department of Labor’s AI Literacy Framework, with integrity checks on every session and an organisation dashboard that is ready now. Every hour spent perfecting a tool is an hour not spent on your people.

We have already tested our people internally. Do we start again?

No need to throw that work away. An internal test shows what people know; the AISA Score adds what they can do, on a benchmarked scale. In the demo we look at what you have already measured and plan the rollout around it.

How do we know the scores match our people?

Check them against your own read. Before a pilot team takes the assessment, write down where you expect each person to land, then compare their results with your notes. Every score links to what the person said and did, so any difference becomes a conversation about specific work. Scores also hold on retest: 98% of people who retake AISA land in the same or an adjacent tier.

Keep exploring

The Science Behind AISA

Metropolitan PoliceHarvard UniversityCrowdboticsE.S.E.

In 2026, Anthropic published the AI Fluency Index — the largest empirical study of AI fluency to date, analysing nearly 10,000 conversations. AISA covers 93% of the behaviours Anthropic identified as markers of AI fluency and goes even deeper with 4 additional criteria. The U.S. Department of Labor's AI Literacy Framework (TEN 07-25) defines what every worker needs to know about AI — AISA covers 100% of its 25 sub-competencies.Read our analysis: Anthropic's AI Fluency Study & AISA · DOL AI Literacy Framework & AISA

AISA's framework is developed by a team with deep roots in tech, behavioural science, and AI product leadership — the rubric is informed by backgrounds spanning the Metropolitan Police, Harvard, Crowdbotics (Silicon Valley), and the European School of Economics.