Let's Connect
AI training for employees

AI Training for Employees

Help your team use the AI tools you've approved. We build the training around their jobs and experience. See a sample lesson, then talk to us about what your team needs.

Claude
Claude
ChatGPT
Gemini
M365 Copilot
GitHub Copilot
Public resources

Start with something you can inspect

Use the example and working files before deciding what help you need.

1
The input, request, first draft and corrected answer are available to inspect.
3
A deadline, an information rule and an unsupported recording promise.
0
The sample and PDF are public. Booking a meeting is optional.
Ajaia

AI training built around your work

1
Start with the taskChoose a real type of work and a useful finished result.
2
Use suitable inputsPrepare invented or approved material. Keep private records out of public exercises.
3
Let people tryGive learners time to attempt the task before showing an answer.
4
Check what changedCompare the output with its sources and correct the mistake.
5
Support another attemptUse changed facts, a reviewer and an agreed follow-up task.
The work

A convincing draft can still be wrong

Speed alone does not show whether an AI-assisted task is useful. People need to recognise missing facts, unsupported commitments and information they should not use.

Check what is getting in the way
×
A deadline copied from the wrong part of the notes
×
A draft asking for customer information it should not use
×
A recording promised when none has been arranged
×
Practice that never reaches another piece of work
A useful lesson gives people a chance to find and explain a mistake.
Made-up invitation exercise
Example
Input
Review
3
Corrections to find
The dates and session details are invented.
Deadline
1
Information rule
1
Recording promise
1
Source notes
5
Written task
1
Changed-facts attempt
1
Compare every claim with the notes
Then try the changed dates
A useful sequence

From a demonstration to checked work

The sample shows the teaching approach. It does not prescribe a standard session length, price or learner capacity.

01
Choose the work
Agree the learners, recurring tasks, starting skills and approved tools.
02
Prepare the example
Set out the input, a useful output and the checks people need to make.
03
Demonstrate and try
Show a small step, then let learners attempt the task themselves.
04
Review and correct
Ask what the source supports. Correct the output before relying on it.
05
Follow up on another task
Use changed material and a manager or work reviewer to check what people can repeat.
What to check

What employees need to practise

A program can focus on one tool or several approved tools. The tasks and support depend on the people doing the work.

Working material

A task people recognise

A clear starting note or document
A short employee request
An imperfect AI answer worth checking
A corrected result and explanation
Another attempt with changed facts
Drafting
Turn agreed notes into an email or update. Preserve dates, owners and unresolved questions.
Comparison
Compare document versions and show which source supports each reported change.
Review
Check an answer against its input and the usual work standard.
Follow-up
Give managers a small task, a checklist and a way to route questions.
How to use it

A short review checklist

Keep the teaching instructions outside the employee’s request. Ask learners to explain the correction.

01
Check dates and commitments
Match dates, times and deadlines to the notes.
02
Keep the scope clear
Check whether joining is optional, how places work and what preparation is expected.
03
Use permitted information
Keep customer emails, staff records and other private work out of this made-up exercise.
04
Remove unsupported promises
Do not promise a recording or another session unless the notes establish it.
05
Try the changed facts
The session moves to 19 November and sign-up now closes on 10 November. Update the email. Before sending it, check that the other details still match the notes.
Useful material

Read, download or discuss

Public samples are available without an email address. We agree any Ajaia engagement after a conversation.

Share the sample agenda, email exercise, review checklist and follow-up questions with colleagues.

What you can use
  • A sample session plan
  • Starting notes
  • An ordinary request
  • A flawed AI draft
  • Corrections
  • Manager follow-up
  • Scope questions

Read the notes and draft before checking the worked answer. No live AI account is needed.

What you can use
  • Read the source
  • Find a wrong deadline
  • Check information rules
  • Remove an unsupported promise
  • Explain the correction
  • Try changed facts
  • Check a new attempt

Explain the jobs people do, the approved tools and the work you want them to practise. We agree the program after qualification.

What you can use
  • Audience and starting skills
  • Work tasks
  • Approved tools
  • Access arrangements
  • Preparation
  • Delivery and support
  • What managers will review
Worked example

Check an AI-written invitation

Made-up training example. These notes and dates are invented. This is not an event you can register for.

Written example
The starting notes
Invented exampleReview the sourceKeep gaps visible
We're holding an optional AI practice session on 12 November 2026, from 14:00 to 15:30 UTC. It's online. Sign up by 5 November. There are 24 places, then a waiting list. We haven't arranged another session yet. We'll use made-up examples. Don't bring customer emails, staff records or other private work information. The session won't be recorded. We'll send the exercise and checklist afterward. Please read the exercise first. It should take about ten minutes. Access instructions will follow registration.
Open the sample
Written example
The employee’s request
Invented exampleReview the sourceKeep gaps visible
Can you turn these notes into a short invitation for managers to share with their teams? Keep it friendly and include the sign-up deadline.
Open the sample
Written example
A first draft
Invented exampleReview the sourceKeep gaps visible
Hi everyone, We're running an optional AI practice session on 12 November, from 14:00 to 15:30 UTC. There are 24 places, and registration closes on 12 November. Please read the exercise beforehand. It takes about ten minutes. Bring a recent customer email to practise with. We'll share the recording afterward.
Open the sample
Written example
What needs correcting
Invented exampleReview the sourceKeep gaps visible
Registration closes on 12 November: The notes say 5 November. Someone following this draft could miss the deadline. Bring a recent customer email: The session uses made-up examples. Customer and staff records aren't allowed. We'll share the recording: The session won't be recorded. The exercise and checklist will be sent afterward.
Open the sample
Check the result

Keep the next action clear

These are worked examples and suggested checks, not measured customer outcomes.

Questions

Before you get in touch

Useful details about the sample, responsibilities and a possible engagement.

Who is this training for?
It's for employees learning to use AI in their jobs, and the people arranging that training. We ask about their roles, experience and approved tools before suggesting a plan.
Do we have to buy a standard package?
No. We start with a conversation about what your team needs. A proposal then sets out the sessions, trainer, preparation, materials and support. The agenda here is an example.
Which tools will people use?
We agree that with you. The plan needs to reflect the tools and features people can access, their licenses and your rules for information. The public example uses made-up notes and can be read without an AI account.
What happens in a session?
A trainer works through a task, then people try it themselves. They check the answer against the material they started with and fix anything that is wrong. The tasks and starting level depend on the group.
What happens afterward?
We agree that in the proposal. It could include a manager reviewing a later task or a follow-up session to work through questions. We keep attendance, task quality and reported use separate when discussing results.
Can we see the material before talking to you?
Yes. Read the example and download the approval packet here. Neither needs an email address or a meeting. Customer references are available upon request. You don't need to buy an assessment before discussing training.
Next step

Talk through the work your team needs to learn

Bring a short description of the learners, the tools they can use and the tasks you want to improve. We agree the trainer, preparation, delivery, support and scope after a conversation.

Customer references available upon request. Agree the scope, facilitator, timing, support and price before delivery.
The work
Name the task
The tools
Check access
The people
Agree roles
The review
Keep an owner