Autosana got 14× the replies in week one
Founder-led outbound, under 1% replies
Autosana builds AI agents for mobile QA. Like most early-stage founders, Yuvan Sundrani ran cold outbound himself. He found the leads, wrote the messages, and sent them from his own accounts.
It mostly did not work. His self-reported reply rate was under 1%. Prospects who fit Autosana perfectly never answered at all.
What driftwood did
driftwood is an AI sales agent for cold outbound. For each prospect on Yuvan's existing list, the agent researched the company, then built a working demo of Autosana for their business: Autosana's agent running against the prospect's own product.
Every message went through human review before it sent. Nothing left the queue without approval. And every message went out from Yuvan's own account, LinkedIn or email, so replies landed where he already works.
The review queue. A human approves every message before it sends.
The CTO who finally replied
One thread tells the story. A CTO at a company we will not name here had ignored Yuvan's outreach for 4.5 months. Yuvan's first message went out on February 26, 2026. It was never answered.
On July 9, the driftwood agent sent him one message. It opened with a real bug the agent had found on the prospect's own site, and attached a 19 second video of Autosana's agent catching it.
Feb 26 · Yuvan's first message, never answered
4.5 months of outreach, no reply
Jul 9
hey , found a bug on .
0:19
Jul 10 · about 12 hours later
send me a blurb + demos to and I'll forward to my team - best
Call booked · Jul 12
The thread, as it happened. Identities redacted on purpose.
The CTO replied about 12 hours later. The call was booked for July 12.
The result
In week one, the same lead list that had returned under 1% produced 14× the replies. Same leads, same founder's account. Different messages.
“amazing stuff, love the demo”
See a demo built for your business
driftwood researches your prospects, builds a working demo of your product for their business, and sends from your own account. A human reviews every message first.
Book a demo