Use paired agents to plan and code a Godot game
One agent handles the game plan and task list while another writes GDScript, checks the build, and prepares commits for approval.
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.
One agent handles the game plan and task list while another writes GDScript, checks the build, and prepares commits for approval.
A developer set up Claude Code and Claude Cowork to handle research briefs, attendee lists, lead digests, and post-call CRM tasks.
An agent handled transcription, thumbnail prep, YouTube metadata, and draft promo posts before Sean approved the final uploads.
Debbie let Grok Bot learn from her past writing, convert a voice memo into a LinkedIn update, and post it after her review.
Use a weekly agent run to spot merged code changes that should have updated documentation.
A support admin used Claude Code to connect ticketing and knowledge tools, while another used a custom GPT to draft replies for manual paste-in.
They connected Cowork to Readwise with n8n to produce a morning briefing from newsletters and test an audio readout.
Tayla Burrell uses Claude Cowork to reshape a finished newsletter into four social assets and a Notion workspace.
Use a planner-worker loop with periodic review to keep coding agents moving on a large repository task.
Claude Cowork can turn existing content into platform-specific drafts, visuals, and scheduled social posts.
Tal Raviv linked Claude Desktop to a small robot and let it navigate, observe, and log a patrol with voice feedback.
One Cowork skill turns a customer interview into three draft assets, with approval checkpoints between stages.
A small-hotel owner used scheduled Claude tasks in Chrome to review Booking.com bookings and confirm which guests had checked in.
Jenny Ouyang used OpenClaw to monitor engagement for four weeks and let it steer a move from X to Bluesky.
They used Claude to inspect a minified desktop app and trace why it depended on macOS, then outlined three possible Linux paths.
A user had an agent connect social accounts, then review and schedule a large batch of videos for the following month.
Claude in Chrome can check Booking.com hotels for a business trip and copy the best matches into a spreadsheet.
They used Replit and ChatGPT to make a first iPhone app, then spent weeks getting it through App Store review.
Claude can enter flights and hotels in a company booking portal and then check restaurant reservation choices for a specific evening.
A user had Claude Code read Amazon invoice PDFs, catch duplicates, and help file FSA reimbursement claims in an Optum portal.
A team used parallel Claude agents to score 115 candidates against a rubric and produce a ranked shortlist.
A PM used Claude in Chrome to organize four Google Ads campaigns, write ads, and enter them in the UI.
A writer used Claude in Chrome to assemble complaint packets, draft forms, and guide disputes, recovering more than $12,000 across ten cases.
Have an agent compare many articles, pull out shared themes, and draft a source-cited summary document.