About this eBook

Most agent tutorials assume one request and one answer. Real work arrives in batches: four hundred invoices, a month of tickets, a spreadsheet someone exported at midnight. Make is built for exactly that shape of problem.

You will learn iterators and aggregators properly, because they are where most Make scenarios go wrong — data quietly collapsing to a single bundle, or fanning out into a thousand operations you did not budget for.

Then you layer agents on top: an AI Agent module or a routed LLM call that classifies, extracts and decides, with error handlers, break directives and incomplete-execution recovery so a failed run resumes rather than restarts.

What you will be able to do

• Read and design Make scenarios fluently — modules, bundles, routers, filters
• Use iterators and aggregators correctly, and see instantly when data collapsed
• Add LLM classification, extraction and decision steps inside a scenario
• Build a document-processing agent that reads PDFs and writes structured rows
• Handle errors with rollback, break, resume and incomplete-execution replays
• Budget operations so a scenario does not quietly burn a month of quota
• Use Make Grid and data stores for state across runs

Table of contents

1. The Make mental model: bundles, operations and the execution log
2. Scenario building blocks — modules, routers, filters and error routes
3. Iterators and aggregators, the two features everyone gets wrong
4. Data stores, data structures and typed mappings
5. Adding intelligence: AI modules, prompt design and structured JSON output
6. Agent one — an invoice and document extraction pipeline
7. Agent two — a ticket triage and routing agent with confidence thresholds
8. Agent three — a research agent that gathers, dedupes and summarises
9. Agent four — a content repurposing agent with a human approval gate
10. Error handling in depth: rollback, commit, break, ignore and resume
11. Scheduling, webhooks, instant triggers and queue behaviour
12. Operations budgeting, throttling and scenario cost forecasting
13. Versioning, cloning across teams, and handing a scenario to a client

Who it is for

• Visual thinkers who want branching logic they can see on one canvas
• Ops teams processing batches of documents, tickets or records
• Agencies delivering client automations that must be legible to hand over

Questions