AI training on your team’s work

Updated 31.10.2026

Practice on real tasks

Participants use agreed documents and spreadsheets to draft answers, then check them against the source.

A useful exercise starts with material your team already knows. For example, participants can group customer feedback by theme, ask for the original quotes and check whether a complaint really recurs. They also learn to spot an answer that sounds certain but has no support in the source.

1. Your task

Summarise customer feedback

2. Work with AI

Ask for themes and sources

3. Review the result

Check facts against the source

Instructions your team can share

Participants keep checked instructions, and AI champions help colleagues use them.

The team agrees on a review method before turning a successful prompt into a shared template. Together, participants decide which information may go into a chosen tool, who checks the answer and what to do when the source is incomplete. That way, a colleague can repeat the task after the session.

Guided team sessions
Your tools and real data
Help with difficult cases
Reusable working templates

Choose an implementation project

We identify recurring tasks, the data they need and how to check the result before proposing software.

A recurring task becomes a project candidate when the input is available, someone owns the process and the result can be checked. We record those conditions alongside cases that need only a better instruction. The next step may be a pilot, an integration or no new software at all.

Explore

Find opportunities in your work

Test

Test methods with your team

Shape

Define data, owner and outcome

Start a project

Choose what to implement

Plan your team’s training

Tell us what your employees work on so we can shape a program around it.

Let’s talk

How we work with your team

Majento is based in Tashkent and works with companies in Central Asia, Russia, Belarus and other markets in English and Russian. Process reviews, training and solution testing can take place remotely. We agree on in-person sessions with your team.

Before starting, we confirm the participants, meeting schedule and working materials. For development, we involve your IT team to agree on integrations, data access and acceptance checks.

Meet the team and find our contacts on the company page.

We teach teams to use AI in their everyday work. In sessions, employees work with their own documents and spreadsheets and learn to check what a model gives them. The programme is built around your processes, with no more than 15 people in a group. By the end, we identify projects worth implementing and the people who will help colleagues use AI. Sessions run in English or Russian.

Who the training is for

Executives practise requesting summaries, checking the basis for conclusions and choosing tasks for review. Marketing and sales teams work on campaign materials, customer questions and content. Operations teams use documents, reports and repeatable procedures. We set the programme after seeing the participants' tasks.

What the team practises

Giving AI a task with a goal, context, source material and output format.
Writing prompts for repeatable work with examples and review criteria.
Checking quotations, calculations, omissions and invented details.
Deciding what company data may go to a model and what must be masked.
Assigning an AI agent a task, permission and approval step.

Tool selection follows company infrastructure and policy. Work data does not go to an external service without an agreed procedure.

Four stages of the engagement

Stage	What the team does
1. Align the team	Work through real examples and agree what a useful answer must contain
2. Find opportunities	Record repeated work, bottlenecks, risks and places where a rule is enough
3. Shape a project	Define an owner, input data, systems, access rules, expected output and acceptance criteria
4. Choose the next step	Choose a pilot, an integration, a new instruction or a decision to leave the process alone

Training can stop at the third or fourth stage. The decision belongs to the team that owns the process.

A practical exercise

Instead of asking “analyse this document”, a participant can write:

Compare these two delivery descriptions. Find differences in quantities, deadlines and payment terms. Quote both documents for each discrepancy. State when information is missing. Do not fill gaps in contractual terms with assumptions.

The participant checks every quotation against the source documents. The template sets the review, but the employee remains responsible for the decision.

How to choose a task for the first session

Choose work the participant already knows how to check, such as comparing two versions of a document or drafting a reply from approved material. Before the session, the employee records the steps they take today, where they find the source data and how they decide that the result is usable. We practise on a de-identified example. If sensitive details cannot be removed safely, we use training data with the same structure.

During the session, we compare the current method and the AI-assisted result on the same example. Quotations, numbers and omissions get a separate check. The employee keeps a useful instruction together with its review steps. A failed answer is useful too when the team can identify where it went wrong. Our guide to a first team session covers preparation, practice and review in more detail.

Choosing a practical task

The materials below cover preparation and practical walkthroughs on work tasks. Pick your section:

What the team takes away

At the start we agree which outputs to record. They can include checked prompt templates, instructions for selected workflows, a register of cases worth testing and a short list of candidates for implementation. If the skill needs to be assessed, the employee completes a real work task independently against agreed criteria.

Reusable Agent Skills instructions

An Agent Skill is a saved instruction: input, steps, expected result, and check. A colleague receives the instruction and applies it to other approved data. Whether automatic connection is possible depends on the chosen environment.

More: How an Agent Skill works.

What the move from a lesson to a workflow looks like

An AI Masters training guide (in Russian) describes a similar sequence: map repeatable tasks, pilot a few processes, create working templates and measure the result. AI Masters is an education project by Serge Shima, a Majento co-founder, and the figures in the guide belong to that training programme.

The sequence is what matters to us: identify repeatable work, test it on real material, document the review method and decide whether separate software is needed. When a skill or instruction is enough, we do not force a development project.

How a training case becomes software

We move a case into an implementation project when it repeats, has usable data, has a responsible owner and can be checked against examples. The project brief records the systems involved, the actions that require approval, the fallback when a source is missing and the criteria for accepting the first version.

The full route is walked through in a training example built on fictional data. AI implementation explains the next stage.

Prices are in USD, excluding VAT.

Formats and starting prices

Team AI training - from $2,600.
One-to-one mentoring - from $1,500.
A team AI hackathon - from $4,800.

These are minimum prices. We build the programme around your business tasks and processes. We agree on the specific programme, format and final price before work begins.

Frequently asked questions

How many people can join a group

No more than 15. This leaves time for participants' questions, practical exercises and reviewing answers. If more employees need training, we discuss separate groups.

Can sessions take place in person

Yes. We prefer in-person sessions. In Tashkent, we can hold sessions at the client's office; we agree on the venue and schedule when preparing the programme. Online sessions are also available.

How long does training take

The AI Masters team programmes include 1-2 day intensives and 3-4 day programmes. These are examples of formats, not a fixed duration for every company. We set the number of sessions and the calendar after discussing your team's tasks.

How is the programme prepared

We review employees' tasks and business processes, then choose exercises, materials and ways to check the result. Before the sessions, we agree on which data may be used. There is no fixed preparation lead time: we agree on it for the specific programme.

What does the team take away

We identify AI champions - employees who can help colleagues work with AI. Together, we identify projects worth implementing. For the selected tasks, we document useful instructions, boundaries and a way to check the result.

Is training included in an implementation project

Yes. During implementation, we train the first group of employees and AI champions. The team learns to use the solution on its own tasks and check the result. Training can also be purchased separately.

Can training be purchased separately

Yes. A team can complete training without ordering software. If practice reveals a task for an agent or integration, we assess that task as a separate project.

The training draws on programmes from AI Masters, an education project by Serge Shima, a Majento co-founder. The AI Masters team training programme describes the topics and session format, and the AI Masters knowledge cards help with review after the sessions. When the team's work points to a project, Majento can build and implement the solution.