Investigate customer errors and draft tickets in support
Cursor's support team uses Cursor with MCP to pull Datadog logs, search past cases, and draft a Linear bug ticket and customer reply.
Done withCursor
- What they did
- Support engineers start in Ask Mode with the codebase in a multi-root workspace. MCP servers bring in Datadog logs and traces, Slack and Notion history, and Linear. Slash commands and skills cover log search, known-issue lookup, ticket drafting and customer replies. Parallel subagents (log investigator, known-issue miner, ticket writer, reply drafter) merge into one output that a person reviews and sends.
- How it went
- Over 75% of support interactions now run through Cursor, and the team estimates 5–10x throughput per support engineer. These figures are the team's own estimates, not measured results.
- Worth knowing
- Write guardrails into the reply prompts so drafts leave out internal service names, error codes and file paths. A human still reviews everything before it goes out.
Try it yourself with Cursor
A customer reported [error description] on [date/account]. Using my connected [Datadog / logging tool], [past case search], and [Linear / Jira], find the relevant logs and similar past cases, then draft a bug ticket and a customer reply. Show me both drafts with the evidence you found, and do not file the ticket or send the reply until I approve.
Source: cursor.com · Undated
Five of these in your inbox every morning
The best things people got an AI agent to do, each with the prompt to try it.
More like this
Keep coding agents running autonomously for long stretches
Cursor's write-up on scaling long-running autonomous coding.
Run a scheduled personal assistant that builds a task dashboard
A Rippling engineer's Cursor cron agent runs every two hours, reading Slack notes, GitHub PRs, Jira issues and mentions, deduplicating them and posting a dashboard.
Migrate 180 microservices to a new framework version
Brex built Cursor-based self-correcting agents that let one engineer migrate 180 microservices to a newer Micronaut version in about a month.
- Custom agent
Sell services through an agent that takes payment and delivers
An agent runs a transaction flow, creates Stripe checkout links, verifies payment before delivering a service, and asks the owner on Telegram to approve refunds or disputes.
Review Meta ad performance inside Ads Manager
An agency tested Manus on a real ecommerce client's Meta ads account; it reported a 14.25x ROAS against an actual 1.76x and corrected itself only when challenged.
- Claude
Run a Monday competitor-monitoring brief from job listings
Claude is set up as an agent that browses competitors and reads their job listings every Monday morning to infer what they may do next.