The AI Readiness Score.
Forty-eight yes-or-no calls against the published AI Readiness Audit. Four minutes for a decisive operator, an honest 0 to 100 score, and the exact gaps that sink AI initiatives before they reach production.
Forty-eight calls. One honest number.
Answer for your organization as it runs today, not as the roadmap says it will. Yes means you could show it to an auditor this week. Not sure counts as no, because a control you cannot verify is a control you do not have.
Free result, no email needed. Answers stay in your browser unless you request the emailed gap report.
Keyboard: 1 yes, 2 no, 3 not sure, arrows to move, Enter for the next pillar. Progress autosaves in your browser.
Six dimensions. The ones the failure data points at.
The scorer runs the 48 checks of the published AI Readiness Audit, a framework built from primary research on why AI projects actually fail: problem selection, data readiness, evaluation discipline, and organizational wiring, not model quality. Each dimension links to its section of the full audit.
How the scoring works.
Most readiness quizzes are a marketing funnel wearing a lab coat. This one shows its arithmetic.
All 48 checks weigh equally.
Each dimension is scored as the percentage of its eight checks you meet, and the overall score is the average of the six dimension scores. No hidden weighting, no proprietary index.
Not sure counts as no.
This is an audit. A control you cannot verify is a control you do not have, and marking it unmet tells you exactly what to go verify first.
Bands are fixed ranges, not percentiles.
Forming (0 to 39), Developing (40 to 64), Operational (65 to 84), and Compounding (85 to 100) describe the state of your checks on this framework. No benchmark dataset sits behind this tool, so it never claims you beat some percentage of companies.
Gaps rank in the framework's own order.
The audit instructs readers to score strategy first because a weak answer there invalidates everything downstream. Your gap list follows that same order: pillar by pillar, check by check.
Every check traces to a cited source.
The 48 checks come from the published AI Readiness Audit, built on RAND, Gartner, MIT NANDA, IBM, Cisco, DORA, McKinsey, and practitioner postmortems. Every statistic was re-verified against its primary source in June 2026, and the sources ship with the report page.
Your answers stay in your browser.
Progress autosaves to localStorage on your device. Answers are only sent to BearPlex if you request the emailed gap report, and then only to build the report we send you.
The four readiness bands
Most of the framework's checks are unmet. The foundations the audit starts with, problem definition and data readiness, come before any serious AI build.
Real foundations exist, but enough checks are unmet that the failure modes the research documents, stalled pilots and silent quality decay, are still live risks.
The majority of checks are met across all six dimensions. The remaining gaps are specific and fixable, and closing them is mostly disciplined engineering work.
Nearly every check is met. Organizations at this range have the wiring the research associates with AI initiatives that survive contact with production.
Questions about the scorer.
Straight answers about how the score is built, what the bands mean, and what happens with your data.
No. There is no benchmark dataset behind this tool, so it never claims you score better than some percentage of companies. The four bands (Forming, Developing, Operational, Compounding) are fixed score ranges on this framework's own 48 checks, and each band's meaning describes the state of your checks, nothing else.
Because this is an audit, not a quiz. If you cannot verify that a control, practice, or owner exists, it does not exist for planning purposes. Scoring not sure as unmet keeps the result honest and tells you exactly what to go verify first.
From the BearPlex AI Readiness Audit, a free published framework built from primary research: RAND's AI failure root-cause work, Gartner and S&P Global abandonment data, MIT NANDA's GenAI Divide study, IBM's breach cost data, DORA's delivery research, Cisco's AI Readiness Index, and McKinsey's organizational rewiring research, among others. Every statistic was re-verified against its primary source in June 2026, and every source is cited on the report page.
No. Your overall score, the readiness band, all six dimension scores, and your three highest-impact gaps are free on screen with no email required. The email is only for the full gap report: every unmet check with the audit's evidence sentences, sent to your inbox and unlocked on screen.
Your answers are saved in your own browser (localStorage) so a refresh never loses progress. They are only sent to BearPlex if you request the full gap report by email, and then only to score the report we send you. We do not share your email, and there is no newsletter attached to this tool.
Read the full AI Readiness Audit.
All 48 checks with their evidence and all 33 primary sources, free and ungated. The scorer is the four-minute version; the audit is the one you take into the boardroom.
Bring the problem.We bring the discipline.
Tell us which world your problem lives in, or let the diagnostic find out. The first conversation is with an engineer, not an account manager.
NDA-first process · SOC 2 Type II audit in progress · GDPR compliant