Two Delivery Models, One Class Catalogue

Training comes in two shapes. Smaller organisations get a customized delivery built on business cases collected before anything is taught: a one-week survey of the tasks that recur every week, a scoring pass that ranks them for AI suitability, and a 90-minute workshop that selects the top five. The guides written afterwards teach the tool by finishing those five real tasks, so the first result is work that was already on the list. Organisations that want a named title book a specialist class from the catalogue instead.

The commercial frame is the same across every title: a maximum of 12 participants, labs on GitHub Codespaces or a DevContainer so nothing has to be installed on corporate machines, and every class runnable privately in-house. The catalogue is ten titles deep and each one has a detail page below. Agentic software engineering runs on Claude Code or GitHub Copilot, the business-user track on Claude Cowork or Copilot Studio, and the architect track on Microsoft Foundry with the Microsoft Agent Framework.

Private delivery runs against the organisation’s own repositories, tenant and deployment targets. The harness configuration, skills, agents and pipelines written during the class stay in use afterwards, instead of a sample project the team has to translate on Monday.

Format follows the audience rather than the other way round. At the simple end, written teachings with hands-on labs worked through at the participant’s own pace. At the other end, fully agentic classes with generated video, avatar presenters, and an AI that watches each learner, spots where they went wrong, and corrects them on demand. Every format runs on the customer’s own scenarios.

Official Microsoft curriculum and Enterprise Skills Initiative classes are booked through a Learning Partner and delivered by a Microsoft Certified Trainer, certified since 2000, with thirty years on the Microsoft platform behind the material. That is where an existing training entitlement or an ESI budget can be spent. The material is licensable as well as deliverable: class repositories carry presenter scripts, generated decks and verified labs, so a customer’s own trainers can teach it under a written licence.

Expert customizable project-based Training

In-depth hands-on classes

Agentic Software Engineering using Claude Code

Aug, 2026 - Last Updated: Aug, 2026 4 Days
Claude Code Agentic Software Engineering Intermediate

Master agentic software engineering with Claude Code across 14 progressive modules covering fundamentals, harness configuration, skills creation, MCP server design, headless CLI automation, planning and spec-driven development, sub-agents, agent teams, harness optimization, security hardening, plugin distribution, agentic DevOps, Claude Managed Agents, and the Claude Code SDK. Build production-grade multi-agent systems that run in parallel, enforce governance through hooks, and deliver at team scale. The class also covers everything required to pass the Claude Certified Architect: Foundations certification.

Explore Agentic Software Engineering using Claude Code

Claude Cowork: From Personal AI to Autonomous Agent Teams

May, 2026 - Last Updated: May, 2026 5 Days
Claude Claude Cowork Agentic AI MCP Microsoft 365 AI Architects Business Owner Beginner

Master Claude Cowork end-to-end across 11 modules: day-one foundations, skills, connectors, MCP, projects, design and brand voice, browser and desktop automation, M365 integration, harness and plugins, agent teams, second brain, and enterprise scaling.

Explore Claude Cowork: From Personal AI to Autonomous Agent Teams

Agentic Business Processes with Microsoft Foundry, Agent Framework & Copilot Cowork

May, 2026 - Last Updated: May, 2026 5 Days
Microsoft Foundry Microsoft Agent Framework Copilot Cowork Agentic AI Agentic Architects AI Engineers Advanced

Five-day pro-code masterclass in Agentic AI engineering with Microsoft Foundry, the Microsoft Agent Framework, and Microsoft 365 Copilot Cowork. Covers Foundry SDK 2.0, Agent Identity, MCP production patterns, Durable Agents, and Cowork integration.

Explore Agentic Business Processes with Microsoft Foundry, Agent Framework & Copilot Cowork

Agentic Software Engineering using GitHub Copilot

May, 2026 - Last Updated: May, 2026 4 Days
GitHub Copilot Agentic Software Engineering Intermediate

A four-day workshop that takes you from GitHub Copilot fundamentals to full agentic software engineering. You will master agent sessions and the Agents Window, implement agentic coding with multi-agent orchestration, use the Copilot CLI and SDK, explore the Copilot desktop app, apply governance and observability patterns, and deliver features through spec-driven workflows. Exercises span both greenfield and brownfield environments so the patterns apply directly to the work you already do.

Explore Agentic Software Engineering using GitHub Copilot

Designing & Implementing Cloud Native Applications using Microsoft Azure

Jun, 2023 - Last Updated: Jun, 2023 4 Days
Azure Cloud Native .NET Aspire Dapr Intermediate

Design and implement cloud-native applications on Azure using Container Apps, .NET Aspire, Dapr, event-driven architectures, microservices, and API management. Learn to evolve from monolithic to scalable distributed systems with hands-on labs across the full Azure cloud-native stack.

Explore Designing & Implementing Cloud Native Applications using Microsoft Azure

Three shapes, depending on whether the goal is a named title, a set of solved tasks, or a team that runs on its own.

Customized Delivery

Delivery starts with the business cases, not the syllabus. A one-week survey collects the tasks that recur every week, a scoring pass ranks them for AI suitability, and a 90-minute workshop selects five. The guides then teach the tool by finishing those five real tasks.

Specialist Classes

Ten titles, each with its own detail page covering modules, audience, prerequisites and duration. The range runs from agentic software engineering and business-user automation to Microsoft Foundry, Angular and cloud-native Azure. Every title runs privately in-house on the organisation’s own scenarios.

Team Enablement

The class runs first. The harness then goes into the organisation’s own repositories: conventions, expert agents, skills with reference leaves, guard hooks and a router constitution file. From there the team’s real work is the exercise, and the learning loop keeps conventions from drifting.

Frequently Asked Questions

Which class fits, and what if none of them do?
The catalogue includes [advanced Angular](/classes/angular-agentic-swe/) and [cloud-native applications on Azure](/classes/cloud-native-apps/) alongside the agentic AI titles. Where no title fits, delivery is built from collected business cases instead: recurring tasks are surveyed, scored for AI suitability, and the five that score highest become the exercises.
How large is a class, and what has to be installed?
A maximum of 12 participants per class, so every exercise gets attention. Labs run on GitHub Codespaces or a DevContainer, which means nothing has to be installed on corporate machines and a locked-down laptop is not a blocker. Some titles need a participant licence for the tool being taught, for example an AI coder subscription.
Is delivery remote or on-site?
Both work. Remote delivery runs over Microsoft Teams with the lab environments live in the browser. On-site suits a private in-house group, particularly where the exercises touch systems the team wants to work on in the room. Every title in the catalogue can be booked privately either way, and where a room is the wrong format the same material is produced self-paced, up to a fully agentic class with generated video and avatar presenters.
Can a class run on our own systems?
Yes, and private delivery is built for it. Exercises run against the organisation's own repositories, tenant and deployment targets, so the harness configuration, skills, agents and pipelines written during the class stay in place afterwards. Nothing has to be translated from a sample project into the real one before it is useful.
What is a team left holding afterwards?
In the blended team enablement shape, the class is the first block. The harness then lands in the organisation's own repositories: conventions, expert agents, skills with reference leaves, guard hooks and a router constitution file. Accompanied days after the class use the team's real work as the exercise, and the learning loop turns what each session reveals into a skill or an agent, so conventions improve instead of drifting.