The short answer: Train staff on AI by starting with the work, not the tool. Pick three tasks each team repeats every week, teach one prompt method on those tasks using your own documents, set clear data rules before the first session, and measure time saved after 30 days. In Kenya, team training costs from Ksh 13,500 per person.
For most Kenyan employers, the question is no longer whether staff will use AI. Staff already do, on their phones, often on personal accounts. The real decision is whether they use it well and safely, or badly and in secret. This guide is for HR managers, training leads and business owners who want to get that right.
- What to teach first: one prompt method on three real tasks per team
- Formats: free demonstration, team programme, private one-to-one, or seats in a public class
- Team rate: Ksh 13,500 per person for 10 or more people
- Private one-to-one: Ksh 30,000 for eight hours
- Before day one: a one-page AI use policy and a "never paste" list
- How to judge it: a 30-day measurement sheet, filled in by each person
Why start with the work, not the tool?#
Because staff remember what saved them time on Monday, not a tour of features. Training that starts from a tool ("this is ChatGPT, this button does this") is forgotten within a week. Training that starts from a task ("here is how you write the weekly client update in ten minutes") gets used.
The habit is already there. In Microsoft and LinkedIn's 2024 Work Trend Index, 75% of knowledge workers said they use AI at work, and 78% of those users bring their own AI tools rather than ones their employer provides (global survey, May 2024, checked October 2026). Training does not create the habit. It points the habit at the right work and adds the right rules.
Before you book anything, ask each team lead one question: "Which three tasks does your team repeat every week that involve reading, writing or summarising?" Typical answers in a Kenyan office:
- Finance: explaining variances, drafting payment reminders, summarising supplier statements
- HR: job adverts, interview questions, first drafts of policies and letters
- Sales: follow-up messages, proposals, call notes turned into next steps
- Operations: weekly reports, meeting minutes, standard customer replies
- Management: board papers, tender responses, summaries of long reports
Those tasks become the syllabus. Everything else is optional.
Which format suits your team: free demonstration, team programme or private sessions?#
Start with a free demonstration if leaders are unsure, book a team programme once at least 10 people need the skill, and use private one-to-one sessions for owners and executives who cannot sit in a group.
- Free demonstration: for the decision makers. A trainer shows AI working on your kind of tasks so you can judge whether a programme is worth it. No cost.
- Team programme: for 10 or more staff. Ksh 13,500 per person, built on your own documents and tasks.
- Private one-to-one: for an owner, director or senior manager. Ksh 30,000 for eight hours, at their pace, on their own work.
- Seats in a public class: for fewer than 10 people. The two-hour AI Kickstart (Ksh 500, Wednesdays 7:00 to 9:00 pm on Google Meet), AI Essentials (Ksh 5,500, one Saturday morning, at Delta Annex or on Google Meet) or the AI Productivity Programme (Ksh 13,500, two Saturdays).
How the free demonstration works in practice: a short session for the people who will approve the budget, using tasks from your own business, followed by a written proposal.
An honest note on in-house training. If you have one confident AI user on staff, they can run lunchtime sessions on prompts. That works for a start. It tends to stall on three things: they lack time to prepare, they have no syllabus, and nobody sets the data rules. An external trainer is worth paying for when you need all staff to reach the same level by a set date.
What should a good team AI programme include?#
A good programme is built around your documents and ends with each person using AI on their own tasks. Ask for these five things in writing before you pay any trainer, including us.
- A discovery call. The trainer learns your teams, your tasks and the tools you already pay for, such as Microsoft 365 or Google Workspace.
- An approved syllabus. You see and sign off the plan before the first session. No surprises on the day.
- Your own documents. Staff practise on your real templates, reports and customer messages, with personal details removed, not on generic examples.
- A follow-up clinic. A session after the training where staff bring what worked, what failed and what they are stuck on. Most of the learning sticks here.
- An AI use policy template. A one-page policy you adapt and sign, so the rules are clear from day one.
Our team programmes include all five. We do not record sessions. Instead, every prompt used in class is sent in writing afterwards, so staff can copy it straight into their own work.
The core skill in our courses is building an AI assistant for a repeated job: a Claude Project that holds the instructions, tone and reference files for that task. A sales officer gets an assistant for proposals. An HR officer gets one for job adverts. Once it is set up, the task takes minutes instead of an afternoon.
How much does AI training for employees cost in Kenya?#
Team training with Cloudwise costs Ksh 13,500 per person for 10 or more people. A team of 12, for example, costs Ksh 162,000. Private one-to-one training is Ksh 30,000 for eight hours.
For smaller groups, sending staff to public classes is usually cheaper:
- AI Kickstart: Ksh 500 per person, two hours live on Google Meet. Good for a first taste.
- AI Essentials: Ksh 5,500 per person, one Saturday, 9:00 am to 1:00 pm. Each person builds their own AI assistant.
- AI Productivity Programme: Ksh 13,500 per person, two Saturdays, eight hours. Adds a live connected workflow and a 30-day plan, with a certificate of completion.
Public classes are paid in full by M-Pesa when booked. For a full comparison of prices across the market, read our guide to how much AI training costs in Kenya.
When you compare quotes, compare the cost per person who is still using AI 30 days later, not the cost per seat. A cheap one-hour talk that changes nothing costs more than it looks.
What must staff know about data protection before day one?#
Before anyone pastes company work into an AI tool, they need to know what must never go in. Set the rules first, then train.
Under the Data Protection Act, 2019, your business is responsible for the personal data of customers and staff, whichever tool processes it. In practice, three steps cover most of the risk:
- A "never paste" list. ID numbers, KRA PINs, M-Pesa statements, bank details, customer names with phone numbers, salary and health information, passwords, and anything under a confidentiality agreement.
- Strip the identity, keep the shape. Replace real details with labels such as "Client A" or "[amount]". The AI's answer is just as good.
- Company accounts, not personal ones. Decide which tools are approved and which accounts staff use for work.
Our free AI acceptable use policy template puts all of this on one page you can adapt and sign before training starts. For the detail on each tool's privacy settings, read is it safe to use ChatGPT at work under Kenya's Data Protection Act.
How do you measure results after 30 days?#
Ask each person to track three tasks for 30 days: how long each took before training, how long it takes now, and how often they used AI for it. Then check the quality of the output with their manager.
A simple measurement sheet, one row per task:
- Task: for example, "weekly client update" or "first draft of a job advert"
- Minutes before training: the person's own estimate, written down before the first session
- Minutes now: measured in weeks two and four
- Times used per week: how often AI was used for this task
- Quality check: the manager's view: better, same or worse than before
- Problem or question: anything that went wrong, to bring to the follow-up clinic
For example, at Pwani Freight (an invented freight forwarder we use in class), the operations team might track customs document summaries, client shipment updates and quotation follow-ups. Each of those is a weekly, repeated, mostly written task: exactly the kind AI helps with.
At day 30, total the minutes saved per week across the team and compare it with the cost of the training. Also look for the quieter signals: fewer late reports, faster replies to customers, managers editing less. If a task shows no improvement, drop it and pick another. Not every task suits AI.
What should you ask any AI trainer before you pay?#
Ask what your staff will be able to do on day 30, and ask to see the syllabus. A good trainer answers both in writing.
- What will each person be able to do after the training? Look for named tasks, not "understand AI".
- Will we practise on our own documents? If the answer is no, the training is generic.
- Which tools do you teach, and why? The answer should fit the tools your company uses or will buy.
- How do you handle data protection? The trainer should bring rules and a policy, not leave it to you.
- What happens after the session? A follow-up clinic, written prompts and a way to ask questions beat a one-off day.
- Can we speak to a past client? Read reviews from people we have trained and ask any trainer for the same.
Be wary of anyone who promises a "certified" AI qualification after a single day, or who cannot say which tool they teach on.
Frequently asked questions#
How many staff can attend one class?#
The team rate of Ksh 13,500 per person starts at 10 people. The group needs to be small enough that every person builds something on their own work while a trainer checks each screen. A larger organisation can be split into groups by department, so each group works on its own tasks.
Can you train our staff on Microsoft Copilot?#
We teach on Claude and compare ChatGPT, Gemini, Copilot and Meta AI in class, so staff learn where each tool fits. The prompt method and the data rules work the same way in Copilot. If your company has standardised on Microsoft 365 Copilot, raise it on the discovery call.
Do you train at our office?#
Public classes run at our training room at 4th Floor, Delta Annex, Delta Corner, Westlands, Nairobi, or live on Google Meet. For team programmes we agree the venue on the discovery call.
Cloudwise Limited, an ICT company in Westlands, Nairobi, runs hands-on AI training on Claude for individuals and teams. If you are ready to plan training for your staff, see how our corporate AI training in Kenya works, from the discovery call to the 30-day follow-up.
Training a team of 10 or more? Ksh 13,500 per person, built on your own documents. Book a 30-minute call at cloudwise.co.ke/ai-training/teams.
Written by
Gavin



Comments
Be the first to comment.
Leave a comment