Workflows, Automation and AI Agents for Professionals
Working professionals who already use AI confidently for daily tasks and want to stop doing it one chat at a time — analysts, project managers, comms and HR leads, consultants, team leads, solopreneurs, freelancers, and small-business owners. This is the Advanced half of the pair: it assumes the Foundations habits (PRISM prompting, TRUST evaluation, and the AI Safety Checklist) from AI+ At Work: Foundations (T1-02) are already familiar. If they are not yet second nature, start there first. No coding required.
Break the chat ceiling — stop using AI one task at a time and start building small systems that start and run inside your work, so the hours you save each week finally compound instead of plateauing.
Place yourself on the AI Capability Progression — Chat, Assist, Specialise, Chain, Delegate — and know exactly which stage you are ready to move to next.
Map your recurring work as workflows and use the AI Opportunity Matrix and the Workflow Worth-It Check to pick the two or three tasks actually worth automating — instead of automating the wrong thing.
Build and maintain a working Custom Assistant for a recurring professional task, designed with PRISM at system scope and reviewed with TRUST at elevated rigour.
Design and deploy a chained workflow that connects two or more of your tools with AI in at least one step and a human-approval gate that stays a real review, not a rubber stamp.
Judge honestly whether AI agents fit your work, govern anything that runs without you using the DELEGATE Protocol's eight components, and leave with a personal 90-day build plan that makes the systems stick.
Module by module, in the order they are assigned.
This module reframes capable AI use from individual chat gains to systems-level leverage because recurring work only compounds when AI sits inside workflows. You can locate your stage on the AI Capability Progression, map recurring work as workflows, and screen candidates with the AI Opportunity Matrix and Workflow Worth-It Check.
Custom Assistants move repeated setup into a persistent specialist tool, which matters when recurring work depends on the same reader, tone, rules, and review standards. By the end, you can choose the right platform category, write PRISM at system scope, build safely, apply elevated TRUST, and improve through use.
Connected workflows turn AI from a place you visit into one step inside work that starts from a real trigger, moves through tools, and pauses for accountable review. You can map triggers, actions, conditions, and approval gates, choose a no-code tool family, place AI with the AI Opportunity Matrix, and design one visible, stoppable chain.
Human-approval workflow design turns an approve button into an active review system for AI chains, where stakes, AI role, drift, and reviewer attention decide whether a step stays safe. You can choose the right approval pattern, apply Replace/Assist/Elevate/Create at step level, build active gates, and add observability and edges.
Delegating to agents and scaling AI across a team requires clear boundaries, because autonomy, shared access, and informal ownership can turn useful habits into risk. You can distinguish assistants, chains, and agents; run the Delegation Readiness Check; and design shared artefacts with ownership, sequencing, and review.
Delegated AI needs governance because systems that keep running face drift, new inputs, tool changes, and human absence after launch. Learners can apply DELEGATE across setup, runtime, and review, document each deployment on a canvas, and sequence Specialise, Chain, Delegate with a review habit.
The course closes with a timed final exam with a pass mark and a limited number of attempts. A pass issues a PDF certificate with its own serial, in the name of the Straits Institute for Applied AI. Cohort progress, at-risk learners and a CSV export are in the admin console throughout.
No. This course is written for working professionals — analysts, project managers, comms and HR leads, consultants, team leads, solopreneurs, freelancers, and small-business owners — not developers. You build everything with mainstream AI assistants and no-code automation tools; there is no coding.
This is the Advanced half of the pair, so it assumes you already use AI confidently for daily tasks — you prompt with PRISM, evaluate outputs with TRUST, and respect the AI Safety Checklist before sharing workplace data. Those come from AI+ At Work: Foundations (T1-02). If they are not yet second nature, start with T1-02 first; a fluent chat user can move through it quickly.
The course gives an honest 2026 picture of what agents can and cannot yet reliably do, and teaches the Delegation Readiness Check you apply before — not after — handing a task to an agent. That is how you avoid the time-sink that leaves a first automation attempt producing nothing useful.
The course is tool-agnostic. It works with the mainstream assistants you likely already use — ChatGPT, Claude, Gemini, or Copilot — and with no-code automation platforms such as Zapier, Make, n8n, or Copilot Studio. You apply everything in whichever tools your workplace already allows.
That question is the heart of the governance module. The DELEGATE Protocol extends the AI Safety Checklist for the case where you are not in the room — setting boundaries, logging, escalation paths, guardrails, an audit cadence, and a shutdown path — so you can delegate more of the work without ever abdicating what you are accountable for.
Six self-paced video modules of about 60 minutes each — roughly six hours in total — and every module ends in a usable artefact: a Custom Assistant, a chained workflow, a governance canvas, a 90-day plan. It closes with that 90-day build plan and a weekly review, so personal AI systems become a normal part of how you work rather than a one-off push.
Write, Speak and Present Better with AI
What AI Is and Why It Changes Everything
Use AI Tools Confidently in Your Daily Job
Study Smarter and Get Ahead with AI
Automate, Grow and Compete in the AI Era
Course libraries are assigned per user, team or department, and the reporting comes back to whoever has to answer for the programme. Tell us who needs it and we will show you the console.
