AI Readiness Assessment: 5 Steps to Evaluate Your Company

An AI readiness assessment tells you whether your company can adopt AI productively and safely, and exactly what to fix first if the answer is "not yet." You can run a useful assessment in two weeks with a survey, a handful of interviews, and a scoring rubric across five areas: data, current usage, skills, process clarity, and risk management readiness. This tutorial gives you the full process.

The output you're aiming for is one page: a score per area, the two biggest gaps, and the first three actions.

What You Need for an AI Readiness Assessment

• Two weeks of calendar time, part-time, for one owner
• A short all-hands survey tool
• 30-minute interviews with five to eight people: a leader, IT or security, and heavy users across departments

Step 1: Survey Employees About AI Use and Skills

Send a five-minute anonymous survey asking: which AI tools do you use for work, sanctioned or otherwise; what tasks do you use them for; what stops you from using AI more; and how confident are you in your ability to use these tools well. Guarantee amnesty for honest answers about unsanctioned use, in writing.

The survey does double duty. It measures readiness, and it surfaces your shadow AI inventory, which you'll need to develop an AI use policy.

Step 2: Interview Leaders, IT, and Everyday AI Users

Interviews allow you to dig deeper into why and how your team uses AI. Ask each person: where does your team lose the most time to repetitive work; where has AI already helped or misfired; what data would you worry about feeding into a tool; and who would you ask today if you had an AI question. That last one is crucial. If nobody can name the person responsible for AI governance, ownership is your first gap.

Step 3: Score 5 Areas of AI Readiness

Use this editorial scoring rubric consistently across all five areas. It is a practical discussion aid that can identify weaknesses and opportunities. Higher scores mean greater readiness.

Score each area from 1-5:

1 — Nothing in place: no agreed process, owner, or evidence.

2 — Ad hoc: some individuals follow a practice, but it is inconsistent.

3 — Defined: a process and owner exist, but coverage or adoption has gaps.

4 — Operating: the process is documented, used consistently, and supported by recent examples.

5 — Reviewed: the process is operating, documented, and regularly checked and improved.

Record one piece of evidence beside each score.

Data readiness

Do you know where your sensitive data lives, who can access it, and what your obligations are for it? Could an employee tell which data is safe to put in an AI tool? Undefined sensitive data can make safe AI use harder to assess and govern.

Current usage

Is AI use visible, inventoried, and happening through approved channels? A company with heavy shadow use scores low on governance and high on AI appetite, and both facts matter.

Skills

Can your people get good output from AI tools, judge when output is wrong, and know what these tools do badly? Overconfidence without skepticism should score worse than honest inexperience.

Process clarity.

Are your recurring processes documented enough that someone could automate them? AI workflow automation amplifies the processes you have. Fuzzy processes automate into fuzzy results.

Risk management readiness

Score how clearly your company identifies AI risks, assigns reviewers, and documents safeguards. Record risk exposure separately as low, medium, or high based on the data and decisions involved. High exposure increases the safeguards needed.

Step 4: Interpret Your AI Readiness Scores

Patterns matter more than totals:

• High appetite, low governance (common): Lead with an acceptable use policy and approved tools.
• High governance, low skills: if rules exist and nobody's confident, lead with training on prompt engineering and low-risk pilot projects.
• Low process clarity: document two or three key processes before automating anything. Otherwise, pilots might fail and negatively impact implementation.
• High risk exposure, anything else low: Go slowly. Get sensitive data rules and named reviews in place before expanding use at all.

Step 5: Create a One-Page AI Readiness Action Plan

State the five readiness scores, the separate risk exposure level, the two widest gaps, and three concrete actions with owners and deadlines. The assessment exists to move toward doing the right work faster, so resist the urge to create a 10-page report.

Share the readout with everyone who took the survey. People who gave honest answers about shadow use deserve to know they’ve been heard. Visible follow-through is a great first step toward using AI at work without losing employee trust.

Example: A One-Page AI Readiness Assessment

Illustrative assessment for a fictional marketing team:

Readiness scores:

Data: 3 / 5 - Data categories defined; guidance not yet tested with users.

Current usage: 2 / 5 - Some tools disclosed; no complete inventory.

Skills: 3 / 5 - Users can draft and check output; practice varies.

Process clarity: 4 / 5 - Content workflow documented and used consistently.

Risk management: 2 / 5 - Reviews happen informally; accountability is unclear.

Assessment:

Risk exposure: Medium — the proposed pilot will draft public marketing content; customer records and consequential decisions are still outside its scope.

Two priority gaps: AI use is not consistently inventoried, and review responsibilities are not documented.

Action 1: Operations lead completes a tool and use-case inventory within one week.

Action 2: IT/security lead documents approved tools, permitted data, and escalation rules within two weeks.

Action 3: Content lead runs a small public-content pilot with named reviewers within three weeks, after actions 1 and 2 are complete.

Share the readout with participants, record what changed, and reassess the affected areas after the pilot.

Troubleshooting your AI readiness assessment

Survey response is low. Have a founder or executive send it personally with a deadline, and keep the time needed to complete the survey under five minutes.

Everyone self-reports high skill. Add one practical prompt exercise to interviews and assess the results.

Scores are low everywhere. Pick one department with a friendly leader and run the whole playbook there first. A small proven win can prevent a stalled plan.

What Your Completed AI Readiness Assessment Should Include

In a matter of days or weeks, you will have a scored baseline, a shadow AI inventory, named gaps, and a three-action plan. Re-run the assessment every six to 12 months and after any major tool rollout. Use the scores and supporting evidence to track your AI governance program.

This article is general information, not legal advice. Have qualified counsel review your policy before adoption.

For a rundown of practical AI techniques you can apply at work and the latest on AI news, subscribe to The Rundown newsletter.