lablab.ai ran a roundup titled "Seven Builders Who Won Hackathons by Making AI Agents Safer," and Deals Machine is one of the seven. The through-line of the piece is that the problem with autonomous AI agents isn't that they can't act, it's that nothing stops them from acting badly. Our answer to that is about as blunt as it gets.
A human on every call
Deals Machine handles the parts of selling that scale: prospecting, research, and outreach prep. But the conversation itself always routes through a live human rep, with Speechmatics surfacing context and suggested responses in real time as the call unfolds. The AI does the preparation; the person does the talking.
A human on every call — by design, not as a disclaimer.
That constraint is also the compliance story. Fully autonomous agents placing live sales calls run into regulation in most markets. Keeping a person on the line is what lets the pipeline stay fast and consistent while landing on the right side of those rules. The safety layer isn't bolted on afterward; it is the architecture.
Recognized at Milan AI Week
The write-up covers the win at Milan AI Week 2026, where Deals Machine took first place in both the Vultr and Speechmatics tracks, built solo. Good company, too: the other six projects in the piece each made a governance or trust layer the headline feature rather than an afterthought.
Read the full story on lablab.aiopen_in_new