Is Your Team Ready for AI Tools?
Buying licences is the easy part. Whether staff use them well, safely and in a way anyone can measure depends on seven things this scorecard checks. The result maps to one of three training plans, so you know what to run first.
Answer for the team you are planning to train, not for its keenest member. If it is the whole business, answer for the typical staff member.
AI Training Needs Scorecard
Answer for the typical member of the team, not the keenest.
A practical self-assessment, not a formal skills audit. Items on data handling are general guidance and not legal advice. Nothing entered is recorded or transmitted.
What the Three Plans Mean
Each band maps to a training plan. The plans are cumulative: a team on the foundations plan will also need the structured plan later.
Foundations plan
Start with the rules and the basics. A short acceptable use position everyone reads, a half-day hands-on session on one approved tool, and a named person to ask. Do this before any wider rollout.
Structured rollout plan
The basics are in place but uneven. Role-based sessions built around each team’s real tasks, a champion in each team, a shared prompt library, and a simple measure of who is using what.
Advanced practice plan
Usage is broad and governed. The gap is depth: workflow-level automation, evaluation of output quality, and measured outcomes so training budget goes where it changes results.
How the Score Works
Eight questions, each scored zero to three where three is most ready. The total is shown as a share of 24 and the band thresholds are fixed.
Score each area
Current usage, policy awareness, prompt skills, data handling understanding, tool access, champions, measurement and leadership use. Zero is least ready, three is most.
Add them up
All areas are weighted equally. A team can score well on usage and poorly on policy, which is the most common and most risky pattern.
Compare to the bands
Under 40 per cent is the foundations plan. From 40 to under 70 per cent is the structured rollout plan. 70 per cent and above is the advanced practice plan.
Check your lowest answers
Whatever the band, the questions where you scored zero or one are the sessions to run first. The band sets the plan, the low scores set the order.
The Areas That Decide Readiness
Four of the eight areas explain most training failures. Each one has a specific fix that does not require a large program.
Policy awareness before skills
Training people to use AI tools well before they know what they may not put into them raises risk rather than lowering it. Staff who paste customer records into a public tool are usually the most enthusiastic users. The acceptable use position comes first, in plain language, and it is the opening ten minutes of every session.
- One page: what is approved, what must never be entered, who to ask
- Cover personal information handling under the Privacy Act 1988
- Repeat it at the start of every hands-on session
- Check understanding with three questions, not a signature
Data handling understanding
Most staff do not know which tools train on their input, where data is processed, or which categories of information are sensitive. That is not their fault; nobody told them. A single session that names the categories, shows the approved tools and explains why the rules exist changes behaviour more than any policy document.
- Name the categories: customer, personal, financial, health, confidential
- Show which approved tools handle which categories
- Explain the difference between a consumer account and a business account
- Give a way to check when unsure that is faster than guessing
Prompt skills built on real tasks
Generic prompting courses do not transfer. Training that works starts from the team’s own tasks: the weekly report, the customer reply, the quote, the meeting summary. Each person leaves with two or three prompts they will use tomorrow, saved in a shared library.
- Collect five real tasks per team before the session
- Build the prompts live, on the team’s own examples
- Save them to a shared library with an owner
- Follow up in two weeks: what is being used, what is not
Champions and access
A champion in each team, with a few hours a month allocated, sustains usage after the trainer leaves. Access matters just as much: if the approved tool takes a week to get a licence for, staff will use the unapproved one they already have.
- One champion per team, chosen for patience rather than seniority
- Allocate time; unfunded champion roles fade in a month
- Same-day access to the approved tool for anyone who asks
- Champions meet monthly and feed the prompt library
Next Steps
AI Governance Checklist
Settle the acceptable use position before the first session.
Open the checklist →AI Maturity Scorecard
See how people and adoption sit against your other capability dimensions.
Score maturity →Frequently Asked Questions
Should we train everyone or start with one team?
Start with one team that has a clear, repetitive task and a willing manager, and use it to build the material for everyone else. Business-wide rollouts before anyone has worked out which prompts and tools suit which roles produce a lot of sessions and little change. One team done well gives you a prompt library, a champion and a measured result to show the rest of the business.
What if staff are already using AI tools on their own?
Then you are on the structured rollout plan at best and the foundations plan if there is no policy, regardless of how skilled the individuals are. Unsanctioned usage is a sign of demand, which is good, and of risk, which is not. The first session should find out what people are using, without blame, and give them an approved path that is at least as convenient. Skills training comes after that.
How long does the foundations plan take?
For a team of up to about twenty people, a written acceptable use position, one half-day hands-on session on an approved tool, and a named person to ask can be in place within a few weeks. The constraint is usually getting the policy decision made and the tool licences approved, not the training itself. Larger organisations take longer because of the number of teams, not because the content is different.
How do we measure whether training worked?
Pick one or two measures before the session and check them a month later. Usage measures: how many people used the approved tool this week, and how many prompts are in the shared library. Outcome measures: the time a specific task takes before and after, measured the same way both times. Satisfaction surveys straight after a session tell you whether people enjoyed it, not whether it changed anything.
Do managers need different training?
Yes, and it is often the most useful session to run. Managers set the tone on whether AI use is expected or tolerated, they approve the time champions need, and they are the ones who decide whether freed hours become anything. A short session on what the tools can and cannot do, how to judge output, and how to measure a task before and after is usually more valuable than teaching managers to prompt.
Is my result recorded?
No. The scorecard runs in your browser and nothing is transmitted. The questions ask about team practice, not about any confidential detail. Screenshot the result for a planning discussion and use the tool as often as you like.
Sources and further reading
- The Privacy Act 1988 (Office of the Australian Information Commissioner)
Know Which Plan You Are On?
Tell us your band, the team size and the tasks they spend most time on. We will suggest a first session, what it should cover, and how to measure whether it worked.