AI Automation Without Coding

Learn to decide what is worth automating and to describe a process precisely enough to build it.

• Candidates for automation: frequency, rules, and low risk
• Process mapping with inputs, outputs, owners, and decisions
• Identifying exceptions that need a human
• Estimating time saved honestly using before and after measures
• Data sensitivity and what an automation may touch

Hands-on lab: Map one real workflow on a diagram and mark which steps are automatable, AI-assisted, or human-only.

Understand the building blocks shared by Zapier, Make, and similar tools, and build your first working flow.

• Triggers: new row, new form response, new email, schedule
• Actions and how data fields pass between steps
• Comparing Zapier and Make concepts: zaps, scenarios, modules
• Authentication, connected accounts, and permission scope
• Reading run history and task usage against free-tier limits

Hands-on lab: Build a form-to-sheet-to-email automation and trace one record through every step in the run history.

Control which records proceed, reshape data, and branch logic so automations behave predictably.

• Filters, conditions, and paths
• Formatting text, dates, and numbers
• Lookups and avoiding duplicate records
• Working with spreadsheets and Airtable as light databases
• Handling missing and malformed fields
• Naming conventions and organising multi-step flows

Hands-on lab: Extend your first flow with a filter, a lookup that prevents duplicates, and a branch for incomplete submissions.

Use AI inside a workflow for classification, extraction, summarising, and drafting with consistent, checkable output.

• Where AI helps in a workflow and where plain rules are better
• Prompting for structured output: fixed categories and field names
• Classification, extraction, and summarisation steps
• Draft-only AI replies placed in a review queue
• Cost, speed, and usage limits of AI steps
• Keeping personal data out of prompts when it is not needed

Hands-on lab: Add an AI step that classifies and summarises incoming records into set fields, then test it on twenty varied samples and record the accuracy.

Make automations dependable by testing edge cases and planning for failure.

• Test plans: normal, empty, duplicate, oversized, and unexpected input
• Error handlers, retries, and failure notifications
• Human approval and escalation steps
• Monitoring run history and spotting silent failures
• Rate limits and what happens when a service is unavailable

Hands-on lab: Run a written test plan against your automation, break it deliberately, and add error alerts and an approval step.

Document, secure, and hand over an automation so others can maintain it, then present your build.

• Documenting purpose, owner, data, and rollback steps
• Least-privilege access and revoking connections
• Privacy, consent, and retention for data moved between tools
• Bias and fairness checks when AI sorts people or requests
• Maintenance schedule and review of AI output quality

Hands-on lab: Complete your documentation sheet and demonstrate your automation live, including a failure case and its handling.