OpenRouter's Rasp sales agent returns about 600 hours a month to a five-person team
OpenRouter built Rasp, an in-house agent that qualifies inbound leads, drafts first emails, writes call briefs and fills the CRM. The company reports about 600 hours a month saved at about $30 a day.
- Organization
- OpenRouter
- Industry
- AI infrastructure
- When
- Oct 2026
- Evidence
- Company reported
Reported outcome
About 600 hours a month returned to the sales team and a 34% faster deal cycle, at about $30 a day, per OpenRouter. The post says not all of the gain is Rasp.
- ~600 (company)Hours returned per month
- ~$30 (company)Cost per day
- 34% faster (company)Deal cycle
- 93% (company)First-touch emails sent without help
Caveats
Stated by the company that uses the agent (blog, earnings call, interview). Not independently audited.
- OpenRouter says close-rate gains happened at the same time as pricing and market changes, so 'not all of this is Rasp'.
- The post is the company's own marketing, and the numbers come from its internal data.
- Follow-ups, re-engaging quiet deals and dropped-deal detection are still done by hand.
- The post has small inconsistencies, for example the per-call time savings add to about 40 minutes while the text says about an hour.
What the agent does
Rasp researches and qualifies inbound leads, sends first-touch emails, writes a pre-call brief, drafts notes from call transcripts and fills most CRM fields. It flags edge cases and sensitive items for a person to approve. Slack is the place where briefs, drafts and flags show up.
Why this works
The work it takes over is prep and paperwork around sales calls, not the calls. People keep calls, negotiation and final approval. OpenRouter also built it as one pipeline with a feature flag on each feature and metrics logged from day one.
What to copy
- Automate the prep and notes, and keep the human on the call.
- Put each new feature behind a flag that starts off.
- Log cost and time saved from the start, and say when other changes may explain a result.
Starter prompt
Written by us, not by OpenRouter.
Lead form: {lead_form}
Company website text: {site_text}
1. Is this a sales lead, a support question, or spam? Say why in one line.
2. If sales: write a 5-line pre-call brief (who they are, what they asked, likely need, 2 questions to ask).
3. Draft a short first email. Mark it NEEDS_REVIEW if the lead is large, unusual, or sensitive.
Sources
- How an AI Sales Agent Saved Our Sales Team 600 Hours a Month — OpenRouter, Oct 7, 2026
Added Oct 19, 2026. Spotted an error? Send a correction.
Get new use cases weekly
One short email. Sources included.
Related use cases
Monte Carlo had every team build an AI agent in one week and logged 150+
150+ agents, skills and projects catalogued in one week; one agent's accuracy on an internal eval rose from 78% to 97%; a 2-3 hour report now takes 15 minutes.
Vercel's lead agent took inbound qualification from a 10-person team to about one person
Inbound SDR work went from about 10 people to 1 plus a part-time person; the agent costs about $5,000 a year to run (as reported).
Uber's Genie answers on-call questions in Slack and saved ~13,000 engineering hours
70,000+ questions answered across 154 Slack channels, with an estimated 13,000 engineering hours saved.