AI for your work · agents and processes
In 8 lessons, you’ll take a real process from your company and leave with a redesigned version: what stays with people, what goes to AI agents, and a one-page specification with quality standards and permissions spelled out. No coding, no video, at your own pace.
get started now →~20 to 25 min per lesson · nothing to download · works on mobile and computer
The thesis
The serious focus for companies isn’t making pretty pages with AI. It’s redesign the work so humans can direct agents as they carry out increasingly large parts of the process. Anyone who knows how to do this with control, quality, and responsibility is one of the most valuable professionals from 2026 to 2028.
74%
of leaders expect that, in four years, almost half of the processes has been redesigned around AI agents (Deloitte).
5%
consider their own processes highly prepared today (Deloitte). The opportunity lies between the two numbers.
21%
of companies say they have mature governance for AI agents (Deloitte). Lesson 7 is about joining this group.
What you’ll know how to do
Map a real process in six links (input, decision, action, verification, exception, result) and mark what can be delegated to agents.
Delegate goals, not tasks: write a long-horizon request with an approval checkpoint and test it in the AI chat you already use.
Build the orchestration chain: human, coordinator, specialists, tools, systems, verification, human.
Write the quality sheet and permissions matrix: what counts as the right answer, when to stop, when to call you, and what the agent cannot do.
Who it’s for
this is for you if
isn't for you if
Learning path
Leave knowing the difference, in one of your own processes, between “AI on top of the old process” and a true redesign.
Leave with a real process mapped and its delegable links marked.
Leave with a long-horizon request tested in an AI chat, with an approval checkpoint.
Leave with the roles assigned across the human, coordinator, specialists, and verification.
Leave with the context inventory: what the agent needs, where it is, and whether it’s accessible.
Leave with the quality checklist: correct answer, tolerance, stop, call, test.
Leave with the permissions matrix decided and sensitive data out of reach.
Leave with a one-page specification and a simulated run completed in chat.
After the course, the Agent Architect Kit includes the seven fillable templates that generate your agent's specification on one page.
Workbooks by profession
Agents in the Clinic — healthcare, 3 lessons. Administrative work (scheduling, confirmations, denials) goes to the agent; medical records and clinical decisions always stay out of scope.
Agents in Accounting and Finance — 3 lessons. The person who enters the information doesn’t approve it: the agent goes as far as the signature step, and certificates and passwords stay locked away.
Agents in the Law Office — 3 lessons. The agent reads, counts, and prepares; the lawyer decides, drafts the legal argument, and signs. Deadlines always require people.
Sources
Microsoft, Work Trend Index 2026 · Deloitte, State of AI in the Enterprise 2026, Agentic AI strategy, Agentic AI is scaling faster than guardrails e AI readiness research · OpenAI, How agents are transforming work · World Economic Forum, Future of Jobs Report 2025 · McKinsey, Rewiring talent strategy in the age of AI.