Use agents to research leads and handle routine ops
A yacht-software CEO built a group of Lindy agents to qualify leads, answer support, and draft content without an engineering team.
Agent Uses · 1,359 real examples
Real examples, collected from Reddit, X, Hacker News, blogs and the press, of people handing work to AI agents: what they asked for, what happened, what went wrong, and a prompt you can copy to try it yourself.
The day's AI news plus 3 agent uses you can try, with ready-to-paste prompts. Free; unsubscribe anytime.
A yacht-software CEO built a group of Lindy agents to qualify leads, answer support, and draft content without an engineering team.
A tester used Operator to gather sales-prospect details from public sources, but it needed repeated guidance after a CAPTCHA and only handled one lead at a time.
Lindy reviewed past guides and a model in Slack, then prepared a Notion idea and waited for sign-off before making it.
A solo founder used Lindy to build support, calling, repurposing, and negotiation agents for small teams and niche businesses.
Cassie Kozyrkov tested Operator on several browser tasks, including paperwork, planning, and shopping, and found it was strongest when the job only needed a decent result.
Annika Helendi used Lindy to gather weekly AI marketing news and draft a newsletter plus promo content on a Monday schedule.
Debbie O'Brien had Grok Bot try to handle a flight reservation, but it stopped just before the end.
Have Lindy check trends each morning, write three LinkedIn draft posts in your voice, and send them to you for approval.
A user had ChatGPT Agent assemble benchmark screenshots in Canva, but it struggled with the interface and made a messy result after more than 75 minutes.
Interlaced used Lindy to summarize sales demos, draft follow-ups, book meetings, and suggest proposals, saving over 5 hours a week.
An enterprise team used Lindy to watch customer chatter, draft milestone posts, build calendars, and publish them after review.
An OpenClaw agent matched pet clinic charges to invoices, filled an insurer form, and left each claim as a draft for review.
Lindy can watch Slack or Google Sheets, then open Linear issues from matching items and save the issue link back to the sheet.
They used OpenClaw to organize video and paper summaries into two separate Markdown-based libraries that could be searched later.
Set up an agent to shop from a product link, complete checkout, and keep checking the order until it updates.
Jonathan Shriftman wired OpenClaw to Telegram so he could ask it to research companies and help draft internal tools.
A red-team benchmark let Claude search forked chains for contract weaknesses and draft exploit code in an hour.
They used Claude Code to clean up overdue Dutch filings by pulling records, matching proof, and preparing tax totals.
Have a computer-use agent play Doom on a remote machine and use its performance to spot workflow problems.
A user had Grok review posted card transactions to spot wrong-card purchases and benefits they had not used.
A developer had an agent make a clicker game from a chat, then push it to GitHub and deploy it to Vercel.
OpenClaw has been used to drive Forza Horizon and to create bots that could take part in a browser game.
A developer made an OpenClaw skill that turns chat requests into backend video edits and returns the rendered file.
A user had an agent compare shoe options and move one pair close to purchase, but bot checks and payment kept the handoff from finishing.