Business and technical tracks

AI Practitioner to AI Architect

Five in-house programmes in two tracks: business staff with no coding, and IT, data and engineering staff.

A facilitator demonstrates an AI agent to a full room of participants
A participant builds a custom Gemini agent on a laptop

Business track, no coding

Levels 2 and 3: AI Practitioner and AI Specialist

Level 2 · Business track · No code

AI Practitioner

Hand your repeatable tasks to AI assistants and agents, and check what they deliver.

2 days
6 modules
Requires AI Foundation (L1) or equivalent

From RM 12,500 15 hours of training

What your team will learn

Module 1

Clean and qualify data

2 hours

Fix duplicates and gaps in your own data, and agree which documents AI may use.

Module 2

Analyse with AI

3 hours

Ask plain-language questions of your data in Copilot in Excel, Gemini in Sheets or Power BI, and check the answers before they reach your manager.

Module 3

Assistants for your own work

3 hours

Pick a chatbot, assistant or agent for each task, and set up an assistant for your recurring work in Custom GPTs, Claude Projects or Gemini Gems.

Module 4

Build your first agent

3 hours

Shortlist tasks worth automating, then build a simple no-code workflow in Power Automate or Zapier and a first agent in Copilot (Agent Builder or Copilot Studio) or Gemini (Gems).

Module 5

Delegate to desktop agents

2 hours

Hand a multi-step task to an agent that works across your files and apps: Claude Cowork, Microsoft 365 Copilot Cowork or OpenAI Codex (it also works on documents, data and reports). Set the scope, review its plan and approve the result. We use the tool your company has; if you have none, the trainer demonstrates on Mereka’s account.

Module 6

Check your agent’s work

2 hours

Review what your agent did, use a troubleshooting checklist, and follow team rules on approved tools and data.

What participants walk away with

A cleaned dataset and a list of approved data sources for your team
A data report with a clear recommendation, checked before it is shared
An assistant for your own recurring work
A first no-code automation, an agent that handles one task and a ranked list of what to automate next
One real task completed end to end by a desktop agent, with its plan and result reviewed
An agent checklist and team rules for approved tools and data

Start here if

  • You use AI weekly but results vary
  • You repeat the same reports or document tasks every week
  • You have finished AI Foundation (L1) or an equivalent course

Who it’s for

Analysts and HR, finance and operations staff

Work with reports, data and documents every week.

What participants need

  • A laptop and a work account for the AI tool your company uses
  • A paid plan of that tool, or Microsoft 365 Copilot, to build agents
  • Power Automate or Zapier for automations
  • We check licences in the discovery call
Licence details

Building an agent needs Copilot Studio or Agent Builder in Microsoft 365 Copilot, or a paid plan of your chosen tool. The desktop-agent module needs a paid Claude plan, ChatGPT with Codex, or Microsoft 365 Copilot with Cowork switched on by IT (Copilot Cowork needs a Microsoft 365 Copilot licence and is billed per task). If your company has none of these, the trainer demonstrates on Mereka’s account.

Tools we use

Microsoft 365 Copilot (Copilot in Excel, Agent Builder, Copilot Studio), ChatGPT (Custom GPTs), Claude (Projects), Gemini (Gems, Gemini in Sheets), Power BI, Power Automate or Zapier, and for desktop agents Claude Cowork, Microsoft 365 Copilot Cowork or OpenAI Codex.

Next Up: AI Specialist (L3) on the business track, or AI Integrator (L4) for IT staff. Sideways: an AI Tool Deep Dive on your company’s tool.

Level 3 · Business track · No code

AI Specialist

Build the AI tools your team relies on, roll them out, and keep them accurate. No code.

2 days
5 modules
Requires AI Practitioner (L2) or equivalent

From RM 14,500 15 hours of training

What your team will learn

Module 1

Organise your team’s knowledge

3 hours

Choose and tidy the documents your team’s assistant will use, name an owner for each, and decide what stays out.

Module 2

Instructions your team can rely on

2 hours

Write the assistant’s role, limits and fallback answers, and test them on the questions your team actually asks.

Module 3

Build your team assistant

3 hours

Build the assistant in Copilot Studio or your chosen platform, check its answers against your documents, and add voice notes and images.

Module 4

Automate a team process

4 hours

Build a multi-step automation with approval steps and alerts, and roll it out to your team.

Module 5

Launch, adoption and upkeep

3 hours

Train colleagues, track usage and feedback, and run a monthly check so the tool stays accurate.

What participants walk away with

A team knowledge source with named owners
Tested team instructions
A team assistant checked against your documents
A live team automation with an approval step
A rollout plan and a monthly check routine

Start here if

  • Colleagues ask you to share your AI setups
  • Your team repeats a process that could be automated
  • You will own a team AI tool after training

Who it’s for

Digital and operations leads

Own the AI tools their department uses.

Process improvement staff

Turn repeatable work into automations.

What participants need

  • A laptop and a work account for the AI tool your company uses
  • Maker access in Copilot Studio, Power Automate or Zapier, or your chosen platform
  • An IT contact who can grant access and approve the rollout to your team

Next Up: AI Integrator (L4) if you move into IT. Sideways: AI Transformation for a company-wide rollout.

Technical track

Levels 4 and 5: AI Integrator and AI Architect, plus AI Developer

Level 4 · Technical track · Code and low-code

AI Integrator

Connect AI securely to the systems your business runs on, in code and in n8n.

3 days
7 modules
Requires scripting and API experience, or AI Specialist (L3)

From RM 24,000 24 hours of training

What your team will learn

Module 1

Connect AI with MCP

3 hours

MCP is a standard way to connect AI to data and apps. Choose the MCP servers that fit your use case and data sensitivity, connect them to your approved AI assistant and build one small MCP server over an internal API.

Module 2

Security for builders

3 hours

Set up sign-in, permissions and audit logs, keep keys and passwords out of code and prompts and defend against prompt injection (hidden instructions in a document that try to take over the AI).

Module 3

Integration workflows in n8n

6 hours

n8n is a visual automation tool. Chain business systems, models and MCP servers in n8n, with retries and failure alerts.

Module 4

Call AI from code

3 hours

Call model APIs from a script, ask for structured outputs your systems can read and manage cost and rate limits.

Module 5

Retrieval (RAG) in code

4 hours

Chunk, index and search company documents in code, return cited answers, and measure retrieval accuracy.

Module 6

Self-hosted models and agents

3 hours

Run an open-source model or agent on internal servers, limit what it can reach, and weigh self-hosting against the cloud for PDPA and cost.

Module 7

Automated tests and logging

2 hours

Run test sets automatically after every change, and log failures, cost and response times.

What participants walk away with

MCP connections to your day-to-day tools, and one MCP server you built
Security settings, audit logs and an approval process for new connections
An n8n workflow that links three or more applications, with retries and failure alerts
A script that calls an AI model and returns structured output
Retrieval in code that returns cited answers, with measured accuracy
Automated tests and logs, and a clear rule for when to self-host
Integration templates for common patterns, tailored to your stack
n8n workflow blueprints you can export and adapt to your own systems

Start here if

  • You can write a short script
  • You have called an API
  • You look after company systems or data

Who it’s for

IT teams

Connect AI to the systems the business already runs.

Technical operations staff

Run and support automations day to day.

What participants need

  • A test environment agreed with your IT team two weeks before day 1
  • API keys for the AI services you use
  • Nothing connects to your live systems during training
Setup details

The test environment is a test server or virtual machine, API keys for the AI services you use, a test internal API and access to n8n. We send the checklist.

Tools we use

MCP, n8n, model APIs and open-source models on your own servers.

Next Up: AI Architect (L5). Sideways: AI Developer, for staff who code daily.

Level 5 · Technical track · For technical leads

AI Architect

Set how AI systems are designed, tested and signed off across your organisation.

2 days
5 modules
Requires AI Integrator (L4) or equivalent

From RM 24,000 16 hours of training

What your team will learn

Module 1

Design AI systems

4 hours

Design system prompts, tool calls and document retrieval as one system, on no-code or developer platforms, then test each part and the whole.

Module 2

Data and memory design

3 hours

Decide which data sources, permissions and memory each AI system gets, and tell data problems from model problems.

Module 3

Quality and guardrail standards

3 hours

Set the tests, input and output filters, and cost limits every AI system must pass before launch.

Module 4

Compliance and security

3 hours

Handle data with consent, minimisation, audit trails and data residency (which country your data is stored in), defend against prompt injection (hidden instructions in a document that try to take over the AI) and data leaks and review systems before launch.

Module 5

Prompt, retrieval or fine-tune?

3 hours

Choose between better prompts, better retrieval or fine-tuning (further training a model on your own examples), write a one-page brief, and assess vendor proposals.

What participants walk away with

A system design covering prompts, tools and retrieval, tested part by part
A data and memory plan for each AI system, with permissions
Quality and guardrail standards every system must pass before launch
Data handling and security controls that meet PDPA and your own policies
A review and sign-off process for AI systems before they go live
A one-page fine-tuning brief, or the evidence that you do not need one
Architecture templates for common AI system patterns
Review templates for system design, test plans and governance

Start here if

  • You approve AI systems before they go live
  • Several teams are building AI tools in different ways
  • You have finished AI Integrator (L4), or lead engineers who have

Who it’s for

Solution architects and senior engineers

Design how AI connects to company data and tools.

Security and compliance engineers

Review AI systems before and after launch.

What participants need

  • Access to the AI platforms your company is considering
  • A sample of the documents or data your designs will use, in a test environment

Tools we use

The AI platforms your company is evaluating, with sample data in a test environment. Data sources: SharePoint, Google Drive and databases.

Next Sideways: Strategic AI Leadership, to brief senior leaders on what you have set.

Standalone · technical track

AI Developer

For people who write code every day. Use AI coding assistants well, review and test what they write, and agree team rules so code and secrets stay safe.

2 days
4 modules
Requires daily coding work; no AI Foundation needed

From RM 24,000 16 hours of training

What your team will learn

Module 1

Coding assistants in your workflow

5 hours

GitHub Copilot, Claude Code, Cursor or the tool your company has approved: when to ask, how to give context and how to keep changes small.

Module 2

Review and test AI-written code

5 hours

Read AI-written code like a colleague’s pull request, write the tests first and catch the mistakes AI makes most often.

Module 3

Keep code and secrets safe

3 hours

Which code and data can go into which tools, keeping keys out of prompts and repositories and checking licences of suggested code.

Module 4

Team rules for AI in your code

3 hours

Agree how your team uses AI: review rules, what must be tested, what is never pasted and how to share good prompts.

What participants walk away with

A coding assistant set up the same way across the team
A review checklist for AI-written code
Team rules for AI in your code, ready to adopt

Who it’s for

Tech leads

Set the team rules for AI in your code and review what it writes.

QA and DevOps staff

Test AI-written code and keep keys and secrets out of prompts.

What participants need

  • The coding assistant your company has approved, on a paid business plan
  • A practice repository
Licence details

Any approved coding assistant works: GitHub Copilot, Claude Code, Cursor or another. We plan the session around the tools you have approved.

What’s included
  • 16 hours of hands-on training on your own stack, in a practice repository
  • A review checklist and team rules your team can adopt the same week
  • Certificate of completion
  • Monthly alumni sessions for three months after training

Every level

What's included at every level

Hands-on practice

Participants build and test their own assistants, automations or systems during the training.

Certificate of completion

A certificate of completion for every participant who finishes the programme.

Impact survey and alumni sessions

One month after training we survey participants and send HR the results. Monthly alumni sessions run for three months.

Prompt library and competency report

AI Practitioner and AI Specialist: a prompt library built around your team’s tasks and a before-and-after competency report.

Add-ons

Add a 30-day follow-up

Training is day one. Over the next month your team uses AI on real work, then shows leadership what changed.

Included

Impact survey and alumni sessions

No extra cost

  • We email participants the impact survey one month after training, and HR gets the results
  • Monthly alumni sessions for three months
Add-on

30-day follow-up session

RM 2,500per session, up to 25 people

  • Two hours online, about a month after training
  • Teams present what they built and how they now use AI
  • We invite your managers and leadership, so they see the results for themselves
  • The impact survey is filled in during the session, and HR gets the results within 5 working days
Add-on

Follow-up session with drop-in clinics

RM 5,000up to 25 people

  • Everything in the follow-up session
  • Plus two 1-hour online clinics in weeks 2 and 3, for your team only
  • People bring the work they are building and get help when they get stuck
  • Your team arrives at the follow-up session ready to present

The same for every level and for AI Developer. Prices exclude 8% SST. HRD Corp claimable.

Proof

How we know it works

What participants told us straight after training (May to October 2026). We report the same surveys for your cohort.

86% Expect to save at least 2 hours a week with AI
87% Plan to use AI at least once a week
4.4/5 Average rating for the trainer’s knowledge
4.2 hrs Saved per person per week, self-reported, about RM 510 a month After foundation-level programmes, at RM 28.41 an hour

What participants built

  • 26no-code AI chatbots built by participants in a one-day hackathon
  • 7 monthsafter training, a manufacturer confirmed in writing that its teams still use the Sheets, Gemini and Forms templates they built
Case study

How our engineers used AI to build a team analytics dashboard

From spreadsheets to a dashboard of our engineering activity, refreshed every 6 hours.

Get started

Book a free discovery call

  1. Step 1

    We learn your goals

    In 30 minutes: your team, your tools and what you want AI to change.

  2. Step 2

    We recommend and quote

    The right programme and format, in a tailored proposal.

  3. Step 3

    We handle HRD Corp

    We prepare the HRD Corp documents and claim from your levy, so there is no deposit. Paying another way? We set it up with you.

Prefer to message? Email ailab@mereka.io or WhatsApp +60 11-3011 6472

HRD Corp claimable

AI Practitioner to AI Architect

Book a free discovery call