Question · The class
What do you build in a two-day AI class?
You practice email triage, document summaries, spreadsheet analysis, research checking, document review and one build lab, all on provided case files.
Short answer
In Learn AI Fast's two-day class, attendees practice AI workflow automation on provided case files for a fictional company: email triage and drafting, document summaries, spreadsheet cleanup and analysis, research with source checking, first-pass document review, and one build lab where everyone builds the same report generator with an AI coding assistant. The class runs two days.
“Hands-on” can mean many things, so here is the concrete answer for Learn AI Fast, taken from its published agenda. Days run 9am to 4:30pm, with two days in all.
Day one: foundations and everyday automation
- Hour 1, how AI assistants and agents work: how assistants and agents work, their limits, the tool landscape, permissions and safe settings. Claude and Claude Code are used hands-on; others are demonstrated.
- Hour 2, the CRAFT framework (Context, Request, Alternatives, Feedback, Transfer): the whole room takes the same board update memo from a naive prompt to a strong one.
- Hour 3, agents and automation concepts: a live demo of an AI agent carrying out a multi-step task; see what Claude Code is and what AI agents are.
- Hour 4, email triage and drafting (lab): triage a provided inbox, draft replies and check them. Our Gmail guide previews the idea.
- Hour 5, documents and files (lab): summarize a provided transcript and project folder, extract action items and write a status update.
- Hour 6, spreadsheets (lab): clean a provided workbook, generate formulas from plain English and verify the numbers. See the Excel guide.
Day two: workflow design, review and responsible use
- Hour 7, AI workflow design: trigger, AI step, human review, output, with a standard worksheet for estimating time saved.
- Hour 8, case study: map and critique three recurring workflows for the same fictional company.
- Hour 9, research with AI, and checking it: build a brief on a provided scenario and verify every claim and citation.
- Hour 10, build lab: use a provided starter project and a written specification with an AI coding assistant to generate, test and correct a multi-step weekly report from provided data.
- Hour 11, first-pass document review (lab): flag clauses in provided sample agreements against a checklist; see whether AI can review contracts. It is not legal advice.
- Hour 12, responsible AI and policy: what never goes into an AI tool, catching hallucinated data, the core elements of a workplace AI policy, and a recap.
There is a lunch break each day. Optional take-home worksheets help you put the skills to work at your own company afterward.
What the tools in the agenda are, per their makers
- Claude Code The Anthropic Claude Code docs say Claude Code can read your design docs in Google Drive, update tickets in Jira, and pull data from Slack through MCP.
- Codex OpenAI describes the Codex CLI as a way to inspect code, make changes, run commands, and automate repeatable work from your terminal.
- Copilot in Excel Microsoft Learn lists these Excel features in Microsoft Copilot: analyze data, generate insights, create formulas and visuals.
- OpenClaw The OpenClaw README says OpenClaw is an open-source AI assistant that runs on your own computer and meets you in the channels you already use.
- Perplexity The Perplexity Agent API docs describe answers grounded in real-time web search, with inline citations in a single call.
What you take home
- A weekly report you generated, tested and corrected with an AI coding assistant, plus practice on email, documents and spreadsheets with provided files.
- A reusable method for designing and checking AI workflows.
- Optional take-home worksheets for applying the skills at your own company on your own time.
Who it fits
Non-developers such as managers, administrators and senior people in business, finance and operations. No coding or prior AI experience is needed. It is not designed for engineers wanting to build machine-learning models.
Practicalities
$1,299 per person, 20 seats per class. The next class is planned for early 2027 (dates to be announced). For a self-paced preview, read the Claude Code guide for non-programmers.
Sources
- Anthropic, Claude Code overview: “Claude Code can read your design docs in Google Drive, update tickets in Jira, pull data from Slack”
- OpenAI, Codex CLI: “Inspect code, make changes, run commands, and automate repeatable work”
- Microsoft Learn, What is Microsoft Copilot?: “Analyze data, generate insights, create formulas and visuals”
- GitHub, openclaw/openclaw README: “runs on your own computer and meets you in the channels you already use”
- Perplexity, Agent API quickstart: “grounded in real-time web search, with inline citations in a single call”
Frequently asked
Do I need to bring anything?
Your laptop and a power adapter. The class provides the exercise files and templates. Bring your own AI assistant account; we send setup instructions.
Will I really finish something?
Everyone builds and tests the same report generator from a written specification in the day-two build lab. The class does not use your own company data.
Applied AI Workflow Automation · Boston
Build this with us, in person
Two days, hands-on, 20 seats. You practice AI workflow automation on provided case files for email, documents and spreadsheets. Next class: early 2027, dates to be announced · $1,299 tuition.