AI Agent Training

Build your own AI agents.
We'll show you how.

A 6-week hands-on program. Go from zero to a working AI agent built on your data, your processes, and your workflows — so your people stop depending on external vendors for AI capability.

First cohort starts June 18, 2026

6 weeks. Hands-on. Real results.

Each week builds on the last. You learn the concepts, apply them to your own workflow, and get support when you're stuck.

Live sessions

Weekly 90-minute workshop. Interactive and hands-on.

Homework

Practical assignments that apply each week's concepts to work you already own.

Support

Direct access to instructors between sessions when you hit a wall.

Cohort-based

Learn alongside others solving similar problems. 8–15 participants per cohort.

Example syllabus

What a cohort covers, week by week.

This is the open-cohort curriculum. Teams on the private track get a syllabus tailored around their tools, workflows, and business constraints after a 30-minute discovery call.

Week 1

Claude Code

  • What AI agents are (and aren't) — cutting through the hype
  • Getting started with Claude Code — setup, configuration, first interactions
  • Working with an AI coding agent effectively
  • Prompt patterns that get reliable results
Homework

Set up Claude Code and complete a real task from your workflow using it.

Week 2

Skills & subagents

  • Extending Claude Code with custom skills and subagents
  • Connecting agents to your real data sources (APIs, databases, documents)
  • Building reusable skills for your team's specific workflows
Homework

Build a custom skill that connects to a data source from your workflow.

Week 3

Agents with OpenClaw

  • From skills to autonomous agents — designing agents that handle multi-step workflows
  • Building and configuring agents with OpenClaw
  • Designing agents for specific tasks: ticket triage, product research, competitor monitoring, reporting
Homework

Build an agent that performs a real task in your workflow end-to-end.

Week 4

Quality & reliability

  • Prompt engineering for consistent, reliable agent outputs
  • Testing and evaluating agent performance
  • Human-in-the-loop patterns — approval gates and escalation
  • Error handling and fallback strategies
Homework

Add quality gates, evaluation, and error handling to your agent.

Week 5

Memory & context

  • Giving agents persistent memory across sessions
  • Knowledge base integration — agents that learn from your data
  • Context management for long-running workflows
Homework

Add memory and knowledge base access to your agent.

Week 6

Capstone

  • Production deployment, monitoring, advanced orchestration, or multi-agent coordination — finalised with the cohort
  • Present your working agent to the cohort
Homework

A fully functioning agent running on your real workflow.

By the end

You'll walk away with:

  • A fully functioning AI agent running on your real workflow
  • The skills to build more — you understand the architecture, not just the steps
  • Confidence to extend and maintain your agents independently
  • Templates and starter code you can reuse for future agents

What people build in this program.

Every participant builds an agent scoped to their actual work. Common starting points:

Product Research Agent

Monitors competitor websites, aggregates product updates, and delivers weekly summaries to Slack.

Ticket Triage Agent

Reads new Jira / Linear tickets, classifies priority and type, routes to the right team, and flags blockers.

Documentation Agent

Watches code changes and automatically updates relevant documentation in Confluence.

Meeting Prep Agent

Pulls context from tickets, docs, and Slack threads to generate briefing notes before meetings.

Customer Feedback Agent

Aggregates support tickets and feedback, surfaces trends, and delivers insights to the product team.

Weekly Report Agent

Pulls activity from Jira, GitHub, and analytics tools to assemble a weekly status report for leadership.

These are starting points. Participants scope their agent to their own workflow during Week 1.

Investment

Two ways in: join an open cohort with the standard syllabus, or run a private cohort tailored to your team after a discovery call.

Cohort (Group)
$2,000/ person

Best for individuals, or sending a few people from your team. Standard syllabus, shared cohort.

  • 8–15 participants per cohort
  • 6 weekly live sessions (90 min each)
  • Homework with instructor feedback
  • Direct support between sessions
  • All templates and starter code
Team (Private)
$10,000· up to 10 people

Best for teams that want shared internal AI capability. Starts with a 30-minute discovery call so we can tailor the syllabus to your team's needs, tools, and business constraints around AI.

  • Private cohort for your team only
  • 30-minute discovery call to tailor the syllabus to your team's needs, tools, and business constraints around AI
  • Customised syllabus scoped to your workflows
  • 6 weekly live sessions (90 min each)
  • Homework with instructor feedback
  • Direct support between sessions
  • All templates and starter code

Sending more than 10? Contact us for team pricing.

Not sure which fits?

Services

We build and run the agent squad for you — wired into Jira, Linear, or GitHub Projects. Shipping in 10 days.

See services

Training

We teach your people to build and run agents themselves. 6 weeks, hands-on, a working agent on their real workflow.

You're reading this

Common questions

Before you enrol, here's what most people want to know.

What technical background do participants need?

You just need to have played with or used Claude or OpenAI in any setting — ChatGPT, Claude.ai, an API tinker, a work experiment, anything. You don't need to be a developer. We start from the fundamentals and move up.

What tools and platforms do we use?

Major AI APIs (Claude, OpenAI) plus the tools your team already uses — Jira, Linear, GitHub, Confluence, Slack. The skills transfer across platforms.

Is this live or self-paced?

Live. Each cohort runs on a fixed 6-week schedule with weekly sessions. The hands-on format and instructor support are what make this work — it's not a video course.

What if someone falls behind on homework?

Reach out. That's what the between-session support is for. We'd rather help you get unstuck than have you skip a week.

Can we take training if we already use AI StackWorks managed services?

Absolutely. Some teams use both — managed agents for core workflows, and train internal people to build supplementary agents for their specific needs.

When is the first cohort?

The first cohort starts June 18, 2026. Register your interest to reserve a spot — we'll confirm the full schedule and onboarding details once your place is secured.

First cohort starts June 18, 2026

Register your interest

Drop your details and we'll follow up with next steps for the first cohort on June 18, 2026.

We'll only use your details to follow up about the training.