This lesson is the practical core of the course. Each workflow has three lines: how to do it, what it costs on the free app versus Google AI Pro, and what it would cost per run on the API if you automated it. API estimates use catalogue prices at the time of writing and a rule of thumb of about 1,300 tokens per 1,000 words.
1. Writing and editing
Paste a draft, state the audience and the change you want ("cut to 300 words for a LinkedIn post, keep the example"). Use a Gem for your house style. Free tier: fully sufficient for a few pieces a day. Pro: only needed if you want Gemini inside Docs and Gmail. API: a 1,000-word draft in and out on Gemini 2.5 Flash is roughly 2,600 tokens, which is well under a cent per run.
2. Research
Two modes. Quick research: ask a grounded question and request sources. Deep Research: give a brief and let Gemini run a multi-step web investigation and write a report. Free tier: quick research works; Deep Research runs are limited. Pro: meaningfully more Deep Research. API: a research pipeline that reads 50 web pages of about 1,500 words each into Gemini 3.8 Flash is roughly 100,000 input tokens, or around $0.08 in input at the time of writing, plus a few cents of output.
3. Coding
Attach the relevant files (or the whole small repo — long context handles it), describe the change, and ask for a diff or a full file. Use Canvas for iterative edits. Free tier: fine for scripts and small fixes; the stronger model is limited. Pro: better for multi-file refactors that need the top model. API: a 20,000-token codebase prompt with a 2,000-token answer on Gemini 3.1 Pro Preview is about $0.04 in and $0.024 out per call at the time of writing; the same on Gemini 3.8 Flash is about $0.015 in and $0.0075 out.
4. Data and spreadsheets
Upload a CSV or a screenshot of a table and ask for the analysis you want: totals, anomalies, a chart description, a formula. On paid plans Gemini can also work inside Sheets directly. Free tier: uploads work well; you copy results back manually. Pro: Gemini in Sheets removes the copy step. API: for a recurring report, send the data as text to Flash Lite — a 50,000-token monthly export on Gemini 2.5 Flash Lite is about $0.005 in at the time of writing.
5. Planning
Turn messy notes into a plan: paste everything, then ask for a structured output — milestones, owners, risks, a week-by-week table. With Workspace connected, Gemini can pull in calendar and email context. Free tier: everything except the Workspace pull. Pro: worth it only if the calendar and Gmail integration saves you real time each week. API: rarely worth automating; this is an interactive task.
| Workflow | Stays free? | Pro adds | API makes sense when |
|---|---|---|---|
| Writing | Yes | Docs/Gmail integration | You generate hundreds of pieces |
| Research | Mostly | More Deep Research runs | You read many sources on a schedule |
| Coding | Small tasks | Top model, higher limits | You build it into tooling or CI |
| Data | Yes | Gemini in Sheets | A recurring report or extraction |
| Planning | Yes | Calendar/Gmail context | Rarely |
Knowledge check
You want to extract three fields from 5,000 invoices a month. Which is the cost-first choice?