Shah: Early multi-agent AI systems got stuck endlessly thanking each other
Amit Shah · He left a $2B ARR company to build AI agents—then hit $1M ARR in ＜ 6 months | Amit Shah, Founder ... · PMF Show · Nov 20, 2025 · at 37:54
Amit Shah is the founder and CEO of enterprise AI agent company InstaLily, discussing early technical hurdles when building multi-agent systems.
0:00 / 0:21exact quote · 21.8s
▶ Watch the full episode on YouTube → 720p mp4 · rendered on demand · StarZero watermark
“When we first started deploying this multi-agent, multi-step architecture, we would get caught up in something called a gratitude loop, and the gratitude loop was that the agents would just pat each other on the back and say, great job, but they would not go to the next step because the reward function was that they would get plus one if they just complemented each other, right?”
quote is from the automated transcript, cleaned for reading: filler sounds and stutters are removed, nothing is rephrased. names can be misheard (the analysis reads context, assessments check outside sources). how →
More from Amit Shah
Instalily's AI agents amplify commercial service technicians by 20x to 50x
“Now that what we have done is that this Insta workers surround each one of these service technicians, and they are able to amplify them, 20 x to 50 x. So suddenly the business can now go out and bid for more contracts, because they have not just escalated and …”
Shah: InstaLily will deliver over $150 million in growth for one client
“So just for one customer, we are on track to deliver more than one hundred fifty million in annualized growth.”
Shah: Enterprise software deployment actually degrades the quality of human work
“Because more and more software is getting written, more and more software is getting deployed, but the quality and quantum of human work was getting degraded.”
Shah: Enterprise AI memory layers must be editable like human memory
“The key thing that we discovered early on, which I think is a differentiated IP that we have, is that very much like the human brain, and actually one of our founding engineers is a neuroscientist and a computer scientist, and really helped us think through th…”