An AI readiness assessment comes down to whether your organization can answer "yes" to a short list of plain questions across five areas: strategy, tooling, adoption, enablement and governance. Below are those questions, with what a "no" usually means and what to do about it. You can run through it in about thirty minutes with the people who actually know the answers.
This is written for L&D leaders, CLOs, HR directors and operators who already have AI tools in place and want to know why they are not showing up in the work. It is a working checklist from my own practice, not a standardized benchmark, so treat the scoring as a conversation aid.
How to use it
Answer each question yes or no. Be strict: "sort of" is a no. Then count the "no" answers per section. A section with two or more "no" answers is where to begin. If you cannot answer a question because nobody knows, that counts as a "no" too, and it is usually the most useful finding.
Do not do this alone. Strategy answers come from an executive, tooling answers from IT, adoption answers from usage data and managers, and the rest from L&D and HR.
1. Strategy
- Is there a written statement of what AI is supposed to improve, with a number or outcome attached? A "no" means tools were bought for activity, not for a result. You will have trouble judging success or choosing what to stop.
- Does a named executive own the outcome? A "no" means decisions stall or get made by whoever is loudest. Name one person.
- Can you say which three to five workflows or roles matter most? A "no" means effort is spread thin across everyone, which usually means it lands on no one.
- Is there a budget line, even a small one, for training and support and not just licenses? A "no" means enablement will be treated as optional.
2. Tooling
- Do you have a current list of every AI tool in use, including AI features inside software you already own? A "no" means you are making decisions without a map. Start with a simple spreadsheet.
- Does each tool have a stated purpose and an owner? A "no" often shows up as overlapping tools that do the same job, with nobody accountable for either.
- Have you checked for redundancy and for gaps? A "no" means you may be paying twice for one capability while a real need goes unmet.
- Can people get to approved tools easily, with the right access and settings? A "no" pushes people toward personal accounts and unapproved tools.
3. Adoption
- Do you know who is using which tool, by role and team, and not just total logins? A "no" means you are measuring access, not use. Two teams with the same license count can look nothing alike.
- Can you name one or two tasks per role where AI is part of the normal workflow today? A "no" means use is occasional experimentation, not practice.
- Do managers know what good AI use looks like on their team, and do they ask about it? A "no" is a common reason adoption stalls: people take their cue from their manager.
- Do you know where people tried a tool and quit, and why? A "no" means you are missing the friction that explains the numbers.
4. Enablement
- Has everyone who uses AI at work had training that covers your rules and your tools, beyond a general tutorial? A "no" means people are learning from videos made for strangers, not for your work.
- Is there role-specific guidance, such as examples, prompts or checklists for real tasks, kept where people work? A "no" means every person reinvents the same approach.
- Do you have a way to tell whether behavior changed after training, such as a work sample or a manager observation? A "no" means completion is your only evidence.
- Is there a place to ask questions and a person who answers? A "no" means problems stay private and recurring errors never get fixed.
5. Governance
- Is there a written acceptable-use policy that employees have actually seen? A "no" leaves each person to guess. The AI governance framework post lays out a structure for one.
- Is it clear what data may and may not go into each tool? A "no" is the highest-risk answer on this list. Fix it first.
- Do you treat uses that influence decisions about people (hiring, performance, pay) differently from drafting and summarizing? A "no" means your most sensitive uses are getting the least care.
- Is there a scheduled review and a path for reporting a problem? A "no" means the policy will age quietly.
Reading your results
- Mostly "no" in strategy: settle the outcome and the owner before buying or training anything else.
- Mostly "no" in tooling: do the inventory and cut or consolidate.
- Mostly "no" in adoption: you need usage data by role, and a conversation with managers.
- Mostly "no" in enablement: this is the usual gap where tools are in place but work has not changed. Build role-specific guidance and a way to see behavior change.
- Mostly "no" in governance: write the data-handling page this week.
Regulatory and policy requirements differ by jurisdiction and industry. This checklist is not legal advice.
If you want a second set of eyes
The checklist shows you where the gaps are. It does not tell you which to close first or how to sequence them. If you are deciding between hiring a strategy consultant and a smaller, fixed-scope option, this comparison lays out when each fits.
Next step
To go further on AI enablement, see The AI Enablement Standard. New posts on this topic are announced in the newsletter.