We're selling an AI SDR by using it on ourselves
Quotafuel sells one thing: AI agents that find, research and email your ideal buyers, then hand the interested ones to your sales team. So the fairest test of the product is to point it at our own market and see if it can book us a meeting.
That's what we're doing, in public. This series logs every week: what the agents did, what a human still had to do, and the real numbers. No cherry-picking. If week 4 is a zero, you'll read about the zero.
The finish line
One booked meeting with a qualified buyer. Not opens, not "interesting, send more info." A sales leader who agrees to a call. When that happens, we'll publish the whole thread (with their permission) as a case study.
How the agents work
1. Pick targets by trigger, not by industry list. The agents look for B2B companies showing a sign they need pipeline right now: hiring SDRs, posting more sales roles than usual, hiring a new CRO or Head of Sales, or listing "cold calling" and "lead generation" in job posts. Our first list is 50 accounts, each with the trigger and a link to the evidence (the job post, or the data source).
2. Find the right person. Usually the VP Sales or CRO, or the founder at smaller companies. Contact data comes from a B2B data provider, and every address is checked before sending.
3. Research each one. Before writing, the agent reads the job posting, recent news, and what the company sells, then picks one specific angle. Example from this week, anonymized: a company opening two new market segments at once got an email about how that usually means months of list-building before the first meetings.
4. Write like a person. Short emails (the best performers in Instantly's 2026 benchmark are under 80 words), plain text, no links in the first email, one clear ask, and a one-line way to opt out.
5. Follow up, then stop. Three touches over about two weeks. Hunter's 2026 data found three messages roughly double total replies versus one. After that we leave people alone.
What a human still does
Being honest about this matters, because "fully autonomous" is mostly a marketing phrase. So far, Aaron (our founder) has had to:
- buy domains and enter card details
- sign in to accounts and approve tool connections
- set the budget (a hard $50 a month for the whole launch stack)
- take the meeting, when one books
Everything else (the target list, research, writing, the website you're reading, this post) was done by AI agents.
Week 0 numbers
| Metric | So far |
|---|---|
| Target accounts researched | 50 |
| First emails sent | 11 (personal, one at a time) |
| Replies | 0 (first emails went out yesterday) |
| Meetings booked | 0 |
| Monthly tool spend | under $50 |
What we expect (and why it's modest)
Average cold email reply rates are about 3.4% to 4.5% (Instantly and Hunter, 2025 data), and many replies are "not now." For planning we assume 0.4% to 0.8% of new prospects turn into a meeting. So 11 emails will most likely produce zero meetings. That's normal, and it's why volume and deliverability matter.
The plan: warm up new sending inboxes through early December (sending cold email at volume from a brand-new domain is the fastest way to land in spam), then move to about 50 new prospects a day from December 8. At that pace our own planning math says roughly 1 to 2 meetings a week, if the targeting is good. We'll find out.
Next week
Second touches go out, the sending inboxes start warming up, and we'll report the first replies, good or bad.
Sources: Instantly Cold Email Benchmark Report 2026 · Hunter, The State of Email Outreach 2026
See what our agents would send in your market
We'll send you 3 real prospects and the emails we'd write them. Free, no call needed.