25 real AI agent use cases in work & operations, each with its source and a prompt to try it yourself.
A home service business connected Grok Bot to its phone system after a few hours of setup. It now listens to every call and posts a summary to the team's Slack channel.
Result Also tags the right people and adds to the to-do list
Every Monday a 12-tab marketing tracker got a 30-minute scroll and a vague gut check. Claude Cowork read every tab and built a color-coded dashboard inside the live Google Sheet.
Small charities can't afford developers. This one runs its day-to-day operations on its own Manus-built website, which saves a lot of time, for $10 a month plus volunteer hours.
Result Tens of thousands of dollars of software, avoided
A poorly designed Chinese API that had to be polled again and again. Codex built the integration (with human guidance where "AI was lost"), the load tests failed, and they advised backing out before a big deposit.
Result Advised backing out before a big deposit
Digging through SaaS APIs one record at a time is slow for an LLM. So a small company dropped several SaaS tools for self-hosted ones and plain files, and cut licensing fees on the way.
Result Lower licensing fees, faster data access
Excel and Office Script workflows, SharePoint JSON formatting, and now statistics: with Copilot chat's deep thinking, the regression model lives in the workbook. Editing it still takes some know-how.
Result Workflows, SharePoint JSON and statistics
Every day the Hermes agent hands out tasks and remembers everything about the business. It caught things they'd have missed, from fundraising and incorporation to compliance.
Result Caught things they'd have completely missed
A small company's ops director, not a developer, also had Claude migrate a Laravel 5 app to Laravel 11 in about 2-3 days, something they'd never have dared touch alone.
Result Laravel 5 to 11 in about 2-3 days
Pulling and cleaning public data into a dashboard, turning call recordings into action points, skimming podcasts by transcript: Claude did the grind around their meetings all day.
Result Podcasts searched by transcript, not listened to
Their company wants a weekly update in a strict template. A custom GPT turns the week's Obsidian notes into it, and 10 minutes of cleanup saves them a couple of hours, even as an LLM-writing skeptic.
Consistent marking and actionable feedback, every time. A UK teacher's custom GPT marks essays for their subject, and the students now use it to revise.
Result Repeatable marking, actionable feedback
Running a SaaS and an AV rental business, they use OpenClaw as an autonomous employee. It traced the dead SMS bot to a legacy app version and rewrote its prompt from real chats. It also once invented revenue figures.
Meetings all day, ideas at 1am. An OpenClaw 'tough coach' skill tears each one apart, why it's weak and what could make it different, so those late hours turn into real work instead of 'tomorrow'.
Result The 2-3 hours before sleep, now productive
An agency's day is morning briefs, meeting prep and client action items. They built their own OpenClaw-style assistant with Claude Code, plugged it into WhatsApp, and now run it all from phone or laptop.
Result Next up: family travel and childcare
No OpenClaw needed: Codex on a ChatGPT Plus plan, kept alive by systemd, with a UNIX account, a mailbox and a diary for memory. It does all kinds of company tasks, and a second bot may follow.
Result No memory problems so far
A small publisher couldn't afford in-house developers for its tedious ops and finance processes. A former engineer in a business role automated them with Claude Code; what's left is mostly QA.
Result Real shipped code, mostly QA left
Invoices, contractor payments in cash or Bitcoin, bookkeeping, 1099s. Sheila drafts and prepares it all; the founder watches and approves before anything goes out. Hundreds of iterations made it reliable.
Result Reliable after hundreds of iterations
A scanner flags CVEs; the real question is which ones can actually be reached. For every one Wiz flags, Claude Code runs a reachability analysis, for about $200 a month in usage.
Result Real risk checked for about $200 a month
When the time-tracking and invoicing tool they relied on hiked its price, they had Claude Opus recreate it and load in 13 years of their own data. Read-only works; editing is still in progress.
Result Read-only version working; editing in progress
A full .Net website and a React Native app, built entirely with Claude Code. It works, but the code is "definitely overengineered", the bugs keep coming, and the lesson was to commit early and often.
Result Lesson learned: commit early and often
Vermont town meeting reports locked the budget away in PDFs. A volunteer had Claude Code and Gemini pull the numbers out and build a comparison site, deployed in under an hour. Rough, but it works.
Result Deployed in under an hour
Research and hand-written proposals ate more than a week per client. Pragmatic's Lindy agents now research and draft, the founders review and send, and they save 20 to 30 hours a week.
Security coverage had to grow without matching headcount. Vodafone's teams built 33 n8n workflows for monitoring, checks and threat response, avoiding £2.2M in costs, according to n8n.
IT handled around 800 lockout requests a month, 35 minutes each. An n8n workflow now sends each one to the employee's manager to approve, then restores access automatically.
Repetitive IT requests piled up as HR platform Remote scaled. An AI help desk built on Zapier now resolves common tickets automatically, saving 616 hours a month and $500K in hiring, per Zapier.