Automation with AI
AI inside an automation does things that used to need a person: reading and sorting an email, pulling data out of an invoice or summarizing a long form. You learn to add it with judgment in Zapier or Make, with human review where it matters and an eye on costs and limits.
Automation with AI
8 h · 4 sessions · Intermediate
- Level
- Intermediate
- Duration
- 8 h · 4 sessions
- Format
- In person or live online
- Language
- Spanish
Who's it for?
What's included
- Live classes with guided practice
- Downloadable guides and materials
- Reusable templates and prompts
- A project applied to your work
- Zevra Learn completion certificate
- Support for questions after the course
Who it's for
- People already automating with Zapier or Make who want the next step.
- Teams that receive lots of emails, forms or documents to process.
What you need
- Having built simple automations in Zapier or Make.
- A Zapier or Make account and access to ChatGPT; we show you the options to connect them.
- No coding needed.
What you'll learn
Choose which steps of a flow benefit from AI and which don't.
Build steps that classify, extract and summarize information with consistent results.
Add human review, control costs and plan for AI mistakes.
The inbox that sorts itself
You automate incoming inquiries or documents: AI classifies each message, extracts the key data (for example, an invoice's tax ID and amount) and logs it to a spreadsheet, with human review before anything important.
Syllabus, session by session.
- 1
Where AI adds value in a flow
Session 1 · 2 h- Classifying, extracting, summarizing and drafting
- When not to use AI
- Designing the flow with human review
- Choosing your case
You leave withThe design of your AI flow.
- 2
Classifying and extracting
Session 2 · 2 h- Prompts for structured answers
- Extracting data from emails and documents
- Validating formats before saving
- Example: the data on a SIAT invoice
You leave withAn AI step that returns structured data.
- 3
Summarizing and drafting inside the flow
Session 3 · 2 h- Summaries for the team
- Reply drafts to review
- Approvals before sending
- Tone and limits
You leave withA flow that prepares drafts for approval.
- 4
Costs, errors and maintenance
Session 4 · 2 h- Cost per run and how to control it
- What to do when AI gets it wrong
- Logs and monitoring
- A plan to scale it
You leave withThe full project, tested and with controls.
Who teaches this course
Every Zevra Learn instructor uses these tools on real projects every day.
Iván Sangueza
Zevra co-founder · Product & technology · La Paz, Bolivia
Engineer and Zevra co-founder. He builds with AI and automation every day for real companies, and that's what he teaches: how to use these tools at work, not the theory. He trained about a hundred CAINCO staff in an in-person program.
Teaches
Keep learning
Make: automations with logic
Scenarios with routes, iterations and error handling for processes with more than two steps.
ChatGPT for everyday work
Write, analyze files and build your own assistants in ChatGPT for the tasks you repeat every week.
What do you want to learn?
We'll build it with you.
Book a short call: we'll recommend the course or path that makes sense for you or your team.