AI Workshop Prep Sheet for Your Team

Why fill this in before the workshop?

The weakest AI workshops start by demoing general tools on examples that look nothing like anyone's work in the room. Participants leave impressed, go back to their desks and do not know where to begin. A useful workshop is built on the team's real tasks and raw material, and ends with an output in every participant's hands that they use the next day.

This sheet gathers what any trainer needs to design a workshop that fits your team: who the participants are and their level, which tasks eat their time, what material can be worked on, which constraints must be respected, and how you will know the workshop worked. Fill it in yourself as the team lead, helped by a short participant survey (at the end of the sheet).

Three rules make a workshop useful: participants work on their real material, not generic examples; every session ends with a tangible output; and skill matters more than the tool, because tools change and method stays.

How to use the sheet

Two weeks before: send the participant survey (end of the sheet) to everyone attending and give them three days.
Ten days before: fill in sections 1 to 7 using the survey answers. It takes about half an hour.
One week before: send the completed sheet to the trainer with anonymised samples of material, and agree the outputs.
Three days before: check accounts, devices and internet (section 8).
One month after: return to section 7 and measure what was achieved.

Section 1: the team

Team or department name, and what it does in brief:

Number of participants and their roles (e.g. 3 marketing, 2 customer service, 1 admin):

Main working language (MSA, dialect, English, French) and the language needed for outputs:

What made you ask for the workshop now? (a problem, an opportunity, a management request):

Section 2: skill baseline

Ask each participant to rate themselves on the statements below from 1 to 4 (1 = never done it, 2 = tried it once or twice, 3 = do it sometimes with confidence, 4 = do it regularly and could explain it to others), then summarise the results here:

Statement	Number choosing 1–2	Number choosing 3–4
I write instructions that set the role, context, task, constraints and output format.
I improve the output with follow-up requests instead of accepting the first result.
I check numbers, names and references the tool gives against an original source.
I know what must not be pasted into a public tool, and I anonymise data before use.
I summarise a long document at different levels depending on the reader.
I produce a simple image or design consistent with the organisation's identity.
I use the tool to clean or analyse a spreadsheet and check the result by hand.

Which tools does the team use today, who uses them, and on which accounts (personal or work)?

Is the gap between levels large? Who are the more experienced people who could help colleagues during practice?

Section 3: three real tasks

Choose three tasks the team repeats, that take time and where AI could help. These tasks become the material for the exercises. If you have an AI use map (resource 13), take them from there.

Task one

The task, who does it and how often per month:

How it is done today, step by step, and how long it takes:

What is the ideal output? Describe it precisely (type, length, language, tone, for whom):

Task two

The task, who does it and how often per month:

How it is done today, step by step, and how long it takes:

What is the ideal output? Describe it precisely:

Task three

The task, who does it and how often per month:

How it is done today, step by step, and how long it takes:

What is the ideal output? Describe it precisely:

Section 4: available material

The strongest exercises are built on real material: a document, a spreadsheet, product photos, meeting notes, customer messages. Decide what can be brought and what needs anonymising before the workshop.

Which material can be used as is (published or not sensitive)?

Which material needs anonymising (removing names, numbers and identifying details)? Who will do it?

Which material must not be used in the workshop at all?

Section 5: constraints

Time available (number of hours, full day or separate sessions, best timing):

Devices and internet (does every participant have a laptop? connection quality in the room?):

Tools and accounts (approved tools, existing subscriptions, tools banned by internal decision):

Data and privacy rules to respect (internal policy, regulated sector, client requirements):

The team's worries or reservations (job loss, mistakes, time, cost):

Section 6: desired outputs

What should every participant hold at the end of the workshop? Choose tangible outputs, not "a better understanding of AI".

Individual output for each participant (e.g. a library of 5 ready instructions for their daily tasks):

Shared team output (e.g. an approved reply template, a draft of usage guidelines, a pilot plan for one task):

What you do not want from the workshop (topics out of scope):

Section 7: success criteria

Write criteria you can measure straight after the workshop and one month later.

At the end of the workshop (e.g. every participant produced their individual output and can explain the checking steps):

After one month (e.g. two tasks from section 3 are done with the tool plus review, with time measured before and after):

Who follows up after the workshop, and how (follow-up meeting, question group, a 30-day pilot):

Section 8: logistics checklist

Accounts for the required tools are created for every participant at least three days before.
Every participant knows to bring a charged laptop; a phone alone is not enough for all exercises.
The internet is tested in the room, with a fallback (a hotspot) if needed.
Raw material is anonymised and saved in a shared folder with clear file names.
The screen or projector is tested and text is readable from the back of the room.
The allowed-data rules are written on a sheet or slide shown at the start.
A short evaluation form is ready for the last five minutes.

Participant survey (send to everyone attending)

Pre-workshop survey

Name (optional): [ ] | Role: [ ]
1. Which tools have you used for work in the past month? [ ]
2. Rate yourself from 1 to 4 on:
   - writing clear instructions for the tool [ ]
   - improving the output with follow-up requests [ ]
   - checking numbers and references [ ]
   - knowing what must not be pasted [ ]
3. Name two tasks you repeat that take longer than they should: [ ] / [ ]
4. What material could you bring to work on (with names removed)? [ ]
5. What exactly do you want to leave the workshop with? [ ]
6. What worries you about using AI in your work? [ ]

A filled example: a development association

Illustrative example

"Sanabel Development Association" is a fictional association in Tripoli whose coordinator asked for a workshop for a team of eight. These are his main answers:

Team: 3 project coordinators, 2 communications staff, 2 finance and admin, and a director. Working language is Arabic; donor reports are in English.
Motivation: periodic donor reports take a week every quarter, and the social pages need regular content.
Baseline: six of eight chose 1–2 for "checking numbers" and "what must not be pasted". Two people in communications use a chat tool on personal accounts and could help colleagues.
The three tasks: (1) turning field notes into narrative report paragraphs in English, twice a month; (2) writing posts about activities, about 12 a month; (3) summarising partner meeting minutes into decisions and tasks, 4 a month.
Material: field notes from a finished activity with beneficiaries' names removed, general activity photos showing no children's faces, two old sets of minutes. Not allowed: beneficiary lists and their data.
Constraints: two days in the same week, four hours each. Two laptops short (their owners will pair with a colleague). No paid subscriptions at present.
Worries: that the tool invents figures in reports, and that beneficiary data is exposed.
Outputs: for each participant, a library of five instructions for their tasks; for the team, a narrative report template and a one-page draft of usage guidelines.
Success criteria: at the end of the workshop, every participant has produced a report paragraph from real notes and checked every figure in it. After one month, the monthly report is written with the tool plus review, and its time is measured against the previous month.

With these answers, a trainer can build a "number hunter" exercise on the association's own notes, give more time to anonymising and checking, and leave image generation for the final part.

Common mistakes

Mistake: asking for a workshop "about AI" with no specific tasks. Fix: three real tasks are enough to steer the whole workshop.
Mistake: bringing material with personal data "because it is internal". Fix: anonymise before the workshop, by a named person.
Mistake: assuming everyone is at the same level. Fix: the short survey reveals the gaps and allows exercises at a basic and a challenge level.
Mistake: vague outputs such as "raising awareness". Fix: a tangible output in every participant's hands.
Mistake: leaving accounts and devices for the morning of the workshop. Fix: prepare three days ahead; half an hour lost creating accounts breaks the day's rhythm.
Mistake: no follow-up after the workshop. Fix: a one-month success criterion and a person responsible for measuring it.

Completion checklist

I sent the participant survey and collected the answers.
I summarised the skill baseline and identified the more experienced people.
I wrote three real tasks with how they are done today and the ideal output.
I set out which material is allowed, what needs anonymising and by whom, and what is off limits.
I wrote the constraints: time, devices, tools, data rules, worries.
Individual and shared outputs are tangible.
Success criteria are written for the end of the workshop and for one month later, with a follow-up owner.
I sent the sheet to the trainer at least a week before.