L&D teams are planning 2027 with a familiar problem. AI tools have spread faster than the skills to use them well.
This list sets out twelve priorities, grounded in 2026 evidence. It starts with free tools you can use this week, because a good plan starts with data about your own people.
Start with free tools
Before you commission anything, gather evidence. These free tools each take ten to twenty minutes.
- The Manager Capability Heatmap maps a group of managers against fifteen behaviours.
- The AI Leadership Readiness Scorecard asks leaders twelve questions on how ready they are to lead with AI.
- The Shadow AI Exposure Check scores how exposed a team is to hidden AI use.
- The AI Team Charter helps a team agree how it will work with AI.
What the 2026 evidence says
The CIPD Good Work Index 2026, based on 5,074 UK workers, found that 76% took part in some form of training in the past year. Yet two-thirds of employees who use generative AI monthly have had no AI training at all.
SHRM surveyed 5,875 US workers in March and April 2026. Nearly 40% said their workplace had provided workshops on practical, day-to-day AI skills; 37% had an introduction to AI tools and nearly a third had coaching. Workers also reported about four hours a week fixing AI outputs.
In the Wharton and GBK Collective study, fielded in July 2026, 78% of leaders said AI training is adequate. Only 56% of employees agreed.
The twelve priorities
Adoption: help people take on the change
- Equip managers first. Gallup found that 78% of employees whose manager clearly supports AI use are frequent users, against 44% of others. Manager support is the strongest lever L&D can pull.
- Teach managers to model AI use. In a Microsoft study of 1,800 workers, manager modelling was linked to AI value 17 points higher and critical thinking 22 points higher.
- Close the communication gap. SHRM found only 1 in 3 individual contributors heard about AI before it arrived in their work. Brief managers so they can brief teams.
- Make experimentation safe. Microsoft found only 13% of AI users are rewarded for reinventing work with AI even when results are not met. Build safe practice into every programme.
Adaptation: help people work with AI and agents
- Train checking, not only prompting. Four hours a week fixing output is a skills issue. Teach people how to review AI work against a clear standard.
- Strengthen critical thinking. 71% of SHRM respondents said critical thinking matters more now. Treat it as a core skill, not an extra.
- Help teams agree how they work. A team charter turns individual habits into shared rules for tools, checking and errors.
- Prepare managers for agents. Wharton found 67% of departments use agents. Managers need to decide where people approve each step and where lighter oversight is enough.
Acceleration: make learning keep pace
- Make AI learning continuous. If 69% of SHRM's AI users rate their experience as none, beginner or intermediate, a single course will not close the gap.
- Reshape roles, not just skills. McKinsey's 2026 report cites its earlier research finding that around 75% of current roles will need reshaping as AI becomes embedded.
- Link learning to outcomes. Wharton found 35% of leaders tracking at least one growth measure reported strongly positive returns, against 13% of those tracking efficiency alone.
- Invest in human skills. The CIPD recommends that organisations provide training in generative AI and address training gaps where they appear.
Build this capability in your organisation
Growth Performance designs bespoke programmes that turn insight into measurable behaviour change.
Read the evidence carefully
Most of these sources are surveys of self-reported views. Microsoft sells AI tools, Gallup and McKinsey sell advice, and GBK Collective designed and ran the Wharton study. Their findings are associations, not proof of cause.
Field dates matter too. McKinsey's survey ran from June to September 2025, while the others ran in 2026. Use the most recent figures where they exist.
Where the sources agree, act. Where they disagree, test on a small scale first.
Sequencing the year
Twelve priorities are too many to start at once. Spread them across the year so each one has time to take hold.
In the first quarter, focus on managers. Brief them early, help them model AI use and give them a way to answer questions.
In the second quarter, move to teams. Help each team agree a charter and a checking standard, and build critical thinking into existing courses.
In the second half, focus on continuity and outcomes. Replace one-off courses with regular practice, and agree how you will measure growth, not only time saved.
Questions for your next planning meeting
- Which of our managers clearly support AI use, and how do we know?
- How many of our people have had AI training, as opposed to training of any kind?
- How much time do our teams spend fixing AI output?
- Which outcomes will tell us that our 2027 investment worked?
How to plan your 2027 programme
Start with your free tool data. Group the twelve priorities under the three conditions, and pick one or two from each.
For the organisational side, see the Acceleration condition. Our page for L&D and HR teams sets out how we work with people functions.
For team-based work, our team events for 2027 can carry several of these priorities in one day.
Try this now
- Send the Manager Capability Heatmap to one group of managers this week.
- Pick one priority from each of the three conditions for your 2027 plan.
- Share this list with your HR or people leadership team.
