Latest / Scrappy but Successful [Consultant's Personalities When No One's Watching] / 25. The Anti-Slop Builder: What 25,000 Hours of AI Looks Like with Piyanka Jain

Transcript

25. The Anti-Slop Builder: What 25,000 Hours of AI Looks Like with Piyanka Jain · Scrappy but Successful [Consultant's Personalities When No One's Watching]

• 0:03
  What's happening everybody? Welcome to Scrappy but Successful. It's a show where
• 0:18
  I sit down with your favorite consultants and we find out if they're successful today
• 0:33
  because they're sophisticated or because they're a little bit scrappy.
• 0:48
  I'm your host, Jack Tompkins, CEO of Pineapple Consulting Firm, always helping
• 1:03
  small businesses become more data-driven.
• 1:14
  Today I'm joined by my friend Piyanka Jain. She's the founder of two companies
• 1:29
  and spends most of her time as the CEO of AskEnola - an agentic AI platform
• 1:44
  that helps businesses move from data to decisions in real time.
• 1:59
  She's an international bestselling author. The book is called Behind Every Good Decision.
• 2:14
  She's been making AI models since 2002. She's a rock star in the industry.
• 2:29
  Piyanka, the question we have for you: do you fall more on the sophisticated side
• 2:44
  or on the scrappy side?
• 2:59
  Yeah, this is a juicy question. There are two parts to how my services are
• 3:14
  experienced by the world around me. The part where we are sophisticated
• 3:29
  is the fundamental technology behind Enola. Enola is serving mid-cap SMB CFOs
• 3:44
  who need to close books, run MRR schedules, P&L reporting - but they need high fidelity.
• 3:59
  No fractional CFO is going to sign their name on reports unless they know it's true.
• 4:14
  Where we are scrappy is our own internal go-to-market.
• 4:29
  If I want to meet somebody I will go through three people to get there. I'm a hustler.
• 4:44
  For the models we've built - the 7-Eleven ATMs - the cash loading is based on our
• 4:59
  algorithm. 75,000 ATMs, each with their own cash withdrawal pattern.
• 5:14
  That's an N number of machine learning AI models, all with very low latency.
• 5:29
  So that part has always been sophisticated. But internally? Sometimes I'm picking up
• 5:44
  the phone and calling people myself. That's scrappy. We don't have a call center.
• 5:59
  Yeah, I grew up in Delhi, India. I come from a community called Oswal
• 6:14
  in Rajasthan. About 500 or 600 years ago we were what was called Rajput - warrior class.
• 6:29
  A deity came and said you can't do violence anymore. So we threw our arms
• 6:44
  and we became business people. So I come from a community that is
• 6:59
  all business people. My dad was an engineer, more of an innovator - doing waste to energy.
• 7:14
  He was a pioneer in his field. I definitely have my innovation disruption
• 7:29
  from my dad, but the general culture I belong to is entrepreneurial.
• 7:44
  I was not born to be thinking about money. When I left PayPal and started Aryng,
• 7:59
  the core desire I have is transformation. I transform organizations from where they
• 8:14
  are to somewhere else where they have more skills, more ability to leverage data.
• 8:29
  If I can take hundreds of hours that CFOs spend manually pulling data from QuickBooks,
• 8:44
  HubSpot, NetSuite and reduce it to half an hour a month - to me, that's exciting.
• 8:59
  So Enola is a CFO command center. Anything a CFO needs to do monthly:
• 9:14
  close books, booking reports, P&L reporting, target vs forecast, MRR, churn, retention.
• 9:29
  If you're looking to raise funds, your investors want to see N months of your books -
• 9:44
  not just P&L, but all your booking reports. We pull data from QBO, HubSpot, ADP,
• 9:59
  Gusto, Chargify - wherever those source systems are. Typically five to six systems.
• 10:14
  And we have a sophisticated agent architecture where the core is deterministic
• 10:29
  and heuristics - which means it doesn't hallucinate. Your revenue will be true revenue.
• 10:44
  Trust is how we do business. I can't do business with a product I don't
• 10:59
  trust. And trust is how you build retention. The stickier they are, the more they trust.
• 11:14
  We started bootstrapped. I funded it initially. The build was never scrappy -
• 11:29
  because I started with high fidelity as the output. We took a lot of left and right turns.
• 11:44
  If you'd talked to me last summer, I would have said I was totally failing. I'd invested
• 11:59
  so much money and we weren't successful. But by end of summer we had cracked it.
• 12:14
  We raised a small angel round earlier this year. We launched in fall - okay, we have a product.
• 12:29
  We started targeting heads of data at Fortune 500. They all said if you build this we want it.
• 12:44
  But three months later I realized: selling to enterprise means POC takes six months.
• 12:59
  You're jumping through hoops, compliance, pen tests - and then leadership changes
• 13:14
  and you're out. So we pivoted to mid-cap companies, and then to the CFO specifically.
• 13:29
  We got so much traction from CFOs that we knew - okay, we've found it.
• 13:44
  I was at SaaStr recently and I was just amazed. Products going blank on the fourth
• 13:59
  click. Funded companies flying planes overhead with their ads - and it's slop.
• 14:14
  Jason Lemkin acknowledged it is slop. His point was slop is okay if it delivers value.
• 14:29
  I think differently. As soon as you have slop, you have loss of trust. Trust is how
• 14:44
  we do business. I can't give my data to a product where I don't trust the security.
• 14:59
  In the vibe coding world, you can build something in four days, but it still takes
• 15:14
  two months to test. Just because you built something over the weekend doesn't mean
• 15:29
  it's a product. It's an MVP. You really need to test and iterate before you call it done.
• 15:44
  Reason why Enola has 25,000 hours of build: wherever you include LLM, you have a
• 15:59
  probabilistic model. You need to test for all the cases. What if strings appear where
• 16:14
  integers are expected? All of that rigor - making it foolproof - is what took most hours.
• 16:29
  When do you know it's ready to ship? If customers can experience it for one full week
• 16:44
  and be delighted, you're good. They will have so much forgiveness after that first week.
• 16:59
  We also label things clearly. If a feature isn't ready, we mark it beta. The customer
• 17:14
  should see it the way we see it. Customers should be able to experience
• 17:29
  the product for one week without issues. After that, they join hands with you to fix it.
• 17:44
  My everything is automated - I can't manage costs manually. I found so many recurring
• 17:59
  subscriptions on old credit cards when we transitioned from Aryng to Enola.
• 18:14
  I grew up in Delhi, landing in Minnesota for my second master's in computer engineering.
• 18:29
  I was teaching calculus and statistics as an assistant lecturer. Once I had to
• 18:44
  ask a friend for a hundred bucks because I had no money left before my check came.
• 18:59
  Pitney Bowes - I got connected to a head of revenue through a senior finance person.
• 19:14
  He said: I'm taking this meeting because you harassed Todd. Todd told me you
• 19:29
  harassed him so I'd make the intro. I am a hustler. I'll text, email, call, WhatsApp.
• 19:44
  I do believe what Rumi said: what you're looking for is seeking you. Enola is right
• 19:59
  for the person who's also looking for Enola. I want the client who says: I can't believe
• 20:14
  I found you. And I want to feel the same. That's what I'm manifesting.
• 20:29
  We celebrate with team offsite trips. We've been to Kashmir, Kerala, Coorg in India.
• 20:44
  Once a teammate said: if we close this deal will you get me a Mac? Yeah, you'll get a Mac.
• 20:59
  I only have one monitor. I had a third big one but the cleaners broke it. That's scrappy.
• 21:14
  My inbox is thousands of emails unread. I miss deadlines. I'm paying fees and penalties.
• 21:29
  I need to build agents to clean this up. For all the agents I've built for others,
• 21:44
  I haven't gotten to building one for myself. It's totally out of hand.
• 21:59
  I've been chasing productivity too long. I want to feel worthy even if my desk is
• 22:14
  a disaster, my calendar is a mess, I miss meetings. Worthy without chasing efficiency.
• 22:29
  We've been sold this productivity thing too long - especially in the US. In Europe
• 22:44
  people are in the cafes, around the lake. They're not checking their phones.
• 22:59
  I want to enjoy this piece of chocolate. I want to enjoy this conversation. I'm here
• 23:14
  for that. The yummy and juicy part. The rest will happen.
• 23:29
  Come check us out at askenola.ai. If you're closing books, raising funds,
• 23:44
  or you're a fractional CFO, come find us. And connect with me on LinkedIn.
• 23:59
  Thank you so much for having me. Stay data driven, everybody.