How to Prepare Your Team for Its First AI Workshop

Choose an observable learning goal tied to work the team already understands. For a first session, that could mean turning approved notes into a clear internal update and explaining which parts still need checking. A goal such as “learn everything about AI” is difficult to practice or evaluate.

Write one sentence before choosing tools: “By the end of this session, participants should be able to draft an update from a supplied brief, identify unsupported details and explain who reviews it.” This is a suggested learning objective, not a promise about a particular workshop's outcome.

• Explore Latitudes' group workshops and team training

Include the people who perform the task, someone who knows what an accurate result requires and the person responsible for the process. In a small business, one person may fill several roles. The important point is that a polished-looking draft has someone qualified to review it.

Ask participants what they have tried and where they feel uncertain. Someone using AI daily and someone opening a tool for the first time may need different support. Agree on a pace, explain unfamiliar terms and provide a written version of the exercise so everyone can follow it.

• Identify who supplies the source information.
• Name the person who can judge accuracy and suitability.
• Ask about access or accessibility needs before the session.
• Give participants a way to raise questions without sharing private records.

Bring one short, approved example with enough detail to review the output. A fictional event notice, a generic internal update or a made-up customer question can work well. Avoid starting with a complicated process that depends on several systems.

Keep names, account details, confidential documents and private customer records out of practice material. Removing a name does not always remove identifying context. When in doubt, build a fictional example from scratch. If the session may be recorded or transcribed, agree on that separately before it starts.

• Review practical questions about responsible AI use

You need an agreed practice setup, but you do not need to buy several tools. Confirm whether participants will use an approved account, watch a demonstration or work through a written example. Make those arrangements before the session so account setup does not take over the learning time.

If people will use a tool directly, check its current terms, workspace settings and the access your organization permits. Do not assume a personal account and a business workspace have identical controls. Keep the session within the approved setup and avoid asking participants to connect business systems just to complete an introductory exercise.

• Compare tools by the task you want to practice

A simple structure is to explain the task, demonstrate a request, let participants practice and compare the output with the source. Reserve time for reviewing mistakes. Watching a convincing demonstration is different from being able to repeat the process and recognize a problem.

The following 60-minute agenda is an illustrative planning option. It is not a stated duration, package or booking offer from Latitudes. Adjust the scope and timing to participants' experience and the work involved.

• First 10 minutes: agree on the task, learning goal and information boundaries.
• Next 10 minutes: demonstrate one request using the approved example.
• Next 15 minutes: let participants try the exercise or review a supplied response.
• Next 15 minutes: compare results with the brief and discuss corrections.
• Final 10 minutes: record questions and choose one follow-up practice task.

Put the original brief beside the result. Ask participants to identify what is correct, what was added without support and what is missing. Readability matters, but an easy-to-read statement can still be wrong.

For the fictional update above, check whether the draft invented a calendar date or named a coordinator. Then review tone and completeness. Ask each participant to explain a correction rather than simply accepting an edited version. That makes the review process part of the learning.

• Can each factual statement be traced to the supplied notes?
• Did the draft add promises, deadlines or policies?
• Does the wording preserve the original meaning?
• What needs a person's approval before the output is used?

Training should cover how to use a tool and how to judge whether its output belongs in the work. The OECD's 2025 report, Generative AI and the SME Workforce, discusses targeted training and internal guidelines as measures for preparing staff to use generative AI responsibly. Its survey covered seven countries, so its findings should not be treated as a forecast for your own business.

NIST's voluntary AI Risk Management Framework Playbook also offers actions for defining responsibilities and evaluating AI risks. These resources can help you prepare questions. They do not certify a workshop, prove legal compliance or establish that a particular tool is suitable for every task.

• Read the OECD's 2025 report on generative AI and the SME workforce
• Explore the NIST AI RMF Playbook

Choose one practice task, one reviewer and a date to discuss what happened. Keep the task within the information boundaries used in the session. Note the corrections participants made and the questions they could not answer.

Use that record to decide whether the next step is more training, a clearer brief or a separate workflow discussion. A workshop does not by itself authorize connecting systems or changing a production process. Latitudes offers workflow consulting and hands-on implementation as distinct services whose scope and responsibilities are agreed before work begins.

• Discuss a suitable next step with Latitudes

Share a short preparation note so participants know what to expect. A useful note answers the following questions without requiring anyone to upload private material.

• What task will we practice, and why was it chosen?
• Who is attending, and what experience do they bring?
• Which fictional or approved sample will we use?
• What tools, accounts or demonstration arrangements are permitted?
• Who will check the result, and what will they check?
• What follow-up task would help us apply the learning?