Your AI Readiness Score: What Each Band Means
The AI Readiness Index gives you a number. This page tells you what to do with it.
# Your AI Readiness Score: What Each Band Means
The AI Readiness Index scores an organization from 0 to 100 across eight dimensions, and that score falls into one of four bands: Ad Hoc, Watched, Gated, or Governed, each pointing to a different next step.
The assessment takes about four minutes. Answer for how your team actually works today, regardless of what a policy says it should do. Full credit on a question means the practice is written down; habit alone doesn't count.
What does an Ad Hoc score mean? (0-39)
Your team is using AI. The rules are vibes.
Tools are in play but nobody has documented which actions need a human yes, which outputs get reviewed, or who can revoke access. Whatever is safe right now is accidental.
What's likely true at this level: AI tools may hold someone's real password instead of a scoped credential. Outputs reach clients without a human checking the claims. New workflows go live on real data without a test run first.
Your next step: Start with the three fixes your results page ranked by impact. These are usually one-line rules. Read What Is AI Readiness? for the full picture of what "ready" looks like. If you want to build your first governed agent with a group, Agent Lab sessions are free.
What does a Watched score mean? (40-64)
Humans review AI output, but informally. Nothing is written down.
Teams in this band are doing the right things by instinct. Someone checks the draft before it sends. Someone tests new tools before rolling them out. The risk: those habits survive until one busy week, one departure, or one "I thought someone else was checking that" moment, and then they quietly stop working.
In practice: review habits exist, but they live in one person's head. Some access is scoped, some isn't. The thresholds and filters that determine what counts as good output live in conversation.
Where to go from here: Write the rules down. For every tool your team uses, document which actions it takes alone and which wait for a human. Then write the judgment criteria for your highest-stakes workflow. Keep it to one page.
What does a Gated score mean? (65-84)
The yes-gates exist and they're documented. Trust is graduated. Gaps remain.
Teams scoring Gated have done the hard part: they decided what requires human approval, wrote it down, and practice it. What remains is usually in the lower-weighted dimensions. Reversibility plans may be incomplete. Ownership of prompts and credentials may not be on paper. Not everyone who touches the tools has been trained on the rules.
Day to day: your highest-traffic workflows have written governance. Newer or lower-priority tools may not. One or two dimensions on your results page scored below 50%, and those are the specific gaps.
Close the gap: Fix the dimensions your results flagged. If Reversibility scored low, turn on versioning and write the undo path for each tool. If Ownership & Skills lagged, consolidate credential ownership and train everyone who touches the tools. Cover what's actually exposed, beyond what's already documented.
What does a Governed score mean? (85-100)
Gates, grounding, scoped access, and ownership are documented, practiced, and checked. Your team moves fast because the rails are real.
Draft-vs-send is a rule, not a habit. Credentials are scoped and revocable. Test-before-production is the default. Everyone who touches AI tools knows the rules, including people who didn't write them.
At this stage: your remaining gaps are narrow. Your results page may show one or two questions scored "Partly" instead of "Yes, and it's written down." Those are the last seams.
From here: Document your operating rules so a new hire or a new team can learn them in a single session. If you're building new agents, the governance structure you have is the foundation. AI Leadership Frameworks covers how to extend governance across departments.
The eight dimensions behind your score
The Index measures eight areas, each weighted by how much damage a gap in that area can cause:
- 01
The Yes Gate
(20%): Does a human approve before AI touches the outside world?
- 02
Truth in Every Draft
(15%): Are outputs grounded in verified facts, never invented?
- 03
Scoped Access
(15%): Do tools hold minimum credentials, revocable in minutes?
- 04
Graduated Trust
(12%): Do new workflows run watched before they run scheduled?
- 05
Stop and Ask
(10%): Does the system pause when something is unclear?
- 06
Written Criteria
(10%): Are goals, thresholds, and filters documented?
- 07
Reversibility
(8%): Can you undo what went wrong?
- 08
Ownership & Skills
(10%): Does your organization own its AI assets, and is everyone trained?
The Yes Gate carries the heaviest weight. A stopped action causes zero damage. A wrong action that already sent is a different problem.
Your results page shows each dimension's individual score. The low-scoring dimensions are where to focus.
What to do with your score
Retake the AI Readiness Index as your team writes rules and closes gaps. The score reflects how your organization operates today. It moves when your practices change.
If you want to work through your fix list with your team in the room, BNEDai runs Readiness Working Sessions. Half a day, your team, your actual tool stack. You leave with written operating rules. Email hello@bnedleadership.com to book.
Frequently asked questions
Build something that actually runs your workflow.
A focused, free 60-minute live session with Jacqueline. You build alongside her, on your own real task, and leave with an agent that is already running.