Nov 20, 2025 · 40m · product-market-fit
He left a $2B ARR company to build AI agents—then hit $1M ARR in ＜ 6 months | Amit Shah, Founder ... · PMF Show
0:00 / 0:00
▶ Watch on YouTube →
gold bands on the timeline = statements, start to end. Hover to read, click to jump. CC turns on captions
In this episode of The Product Market Fit Show, host Pablo interviews Amit Shah, founder of InstaLily, about building AI agents that amplify human enterprise operators, scaling rapidly to $1M ARR, and executing an effective enterprise go-to-market strategy.
How this conversation actually went
Every chapter scored 0–10 on four independent dynamics. Hover any point for the reasoning behind the score. Pablo holds 13.2% of the talking time here. How this is scored →
Pablo as informed peer 3.2 Guest teaching 3.6 Guest disagreement 0.1 Pablo pushing back 0.7
speaking balance: gold is Pablo, purple is the guest (3 minute bins)
Sharpest disagreement ▶ 7:15 Reframing enterprise drudgery beyond management oversight
Amit respectfully dismisses Pablo's hypothesis that enterprise software bloat is purely a management visibility problem, reframing it as a fundamental structural breakdown in how software mediates human cognition.
Hardest push from Pablo ▶ 23:17 Challenging startup access to enterprise dataPablo presses Amit on the severe credibility gap faced by a tiny early-stage startup asking multi-billion-dollar corporations for direct access to highly sensitive, proprietary internal data.
Biggest teaching moment ▶ 17:40 Neuroscience-inspired dynamic context pruningAmit educates the audience and host on the necessity of editable, pruneable memory in multi-step AI agents to prevent stale quarterly business context from degrading current execution.
Pablo holds their own ▶ 12:53 Synthesizing the three paradigms of AI value creationPablo articulates a comprehensive framework categorizing AI applications into replacement, workflow automation, and creating entirely new tiers of previously unfeasible analytical work.
the scores for every segment, with the reasoning behind each
| Chapter | Topic | Pablo as informed peer | Guest teaching | Guest disagreement | Pablo pushing back | Why |
|---|---|---|---|---|---|---|
| Amit Shah's Background and Scaling 1-800-Flowers | 2 | 3 | 0 | 0 | Pablo welcomes Amit and asks for his professional background. Amit outlines his two decades in operationally complex supply chain and logistics businesses like 1-800-Flowers and Blue Apron. | |
| The Philosophy of AI-Powered Human Amplification | 4 | 4 | 1 | 2 | Amit explains how systems of record trap human operators in low-value data entry. Pablo offers hypotheses on management disconnects, prompting Amit to reframe the problem around a three-tier operational stack. | |
| Enterprise AI Use Cases and Market Unlocking | 5 | 3 | 0 | 1 | Pablo synthesizes the AI paradigm into three distinct categories: replacement, automation, and unlocking previously impossible work. Amit validates Pablo's framing with real-world distribution examples. | |
| InstaLily Architecture: InstaBrain and InstaWorkers | 2 | 5 | 0 | 0 | Amit details InstaLily's core architecture, explaining how InstaBrain acts as an editable, pruneable contextual layer inspired by human neuroscience to prevent context pollution in multi-step AI agents. | |
| Initial Enterprise Customers and Governance Strategy | 3 | 3 | 0 | 2 | Pablo presses Amit on how a tiny early-stage startup convinced multi-billion-dollar enterprises to share proprietary systems. Amit explains their deliberate focus on unglamorous enterprise infosec and SOC compliance. | |
| Measuring Enterprise ROI and Value Creation | 4 | 3 | 0 | 0 | Amit outlines the three pillars of enterprise value measurement: rapid time to value, expansion velocity, and engagement depth. Pablo jumps in to emphasize why rapid time to value beats delayed massive value. | |
| Go-To-Market Execution and Industry Trade Shows | 4 | 3 | 0 | 0 | Pablo and Amit discuss enterprise go-to-market mechanics, highlighting the resurgence of in-person trade shows and learning from frontline workers rather than just pitching executives. | |
| Recruiting AI Talent and Rapid Business Scale | 3 | 4 | 0 | 1 | Amit explains his unconventional hiring philosophy of selecting AI-native new college graduates over seasoned enterprise veterans, enabling rapid ARR milestones and early profitability. | |
| Product-Market Fit, Technical Overcoming, and Founder Advice | 2 | 4 | 0 | 0 | Amit shares early technical edge cases, including agent 'gratitude loops' where multi-step models stalled by continuously praising each other, before concluding with advice on picking a focused lane. |
Statements from this episode (15)
Shah: Almost half of global GDP is still 90% manual work
“Almost half of the global GDP
Is 90% still manual?”
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: AI's true value is unlocking work organizations previously couldn't afford
“The work that we think can be mediated by AI and AI agents is not just the work that we see exists right now, but it's actually work that the organization was never able to get to because either the scale required or the cost required to get that done was extr…”
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: Enterprise AI demands vertical domain specificity because one size fits none
“Our belief is one size fits none. So you have to really understand the nuances because ultimately a lot of impactful work requires domain specificity, right?”
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…”
Shah: Dominating AI in industrial verticals means billions in ARR
“Whether it's construction vertical, industrial goods, We think those are trillion dollar verticals. They have a lot of space to grow. And if you could dominate any one of those verticals, you know, that's like literally billions in ARR.”
Srugo: Fast, modest value beats delayed massive enterprise value in SaaS
“That's a huge one by the time to value, I think is an underrated, underappreciated concept just in general, like less value, but in less time, you know, a lot of times beats massive value in a very long time because you don't get there.”
InstaLily's GTM strategy involves attending over 100 trade shows this year
“So this year we are on pace to have attended more than a hundred sort of trade shows.”
Shah: Enterprise sellers must interview frontline workers before pitching to VPs
“Because a lot of people will just jump to, hey, can I get in front of that VP? Well, the VP's pain points are actually reflective of the line management's pain points. And you're better off going and talking to that frontline worker and really asking me, hey, …”
Shah: InstaLily reached $1 million in ARR within a matter of months
“We were able to hit that first million ARR within a matter of months.”
InstaLily surpassed standard triple-triple SaaS growth benchmarks in its second year
“And then we are in the second year of our sales cycle and confidently passed that triple triple and actually accelerating our growth.”
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: InstaLily reached profitability before raising its Series A round
“By the way, we were profitable before raising series A”
Shah: Early multi-agent AI systems got stuck endlessly thanking each other
“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 t…”