An Operator's Blog is hosted by Soumitra Sharma, General Partner at Operators Studio, a micro-VC firm focused on the US-India corridor. This podcast takes you into the world of startups and venture capital, featuring candid conversations focusing on sharing real operating insights, tips, and playbooks that you can immediately start using to solve your biggest challenges as a founder and operator.
Join Soumitra as he cuts through the BS & interviews top founders, operators, and investors who have built businesses in the trenches for decades.
This episode delves in-depth into building a Prosumer AI startup.
We cover everything from overlooked spaces in AI, how to choose problem statements as a founder, learnings on upfront market research & customer discovery, creating defensibility in Prosumer AI, managing the foundation model layer, including closed source vs open source, Prosumer GTM, particularly influencers and affiliate marketing, new usage-based pricing models, fundraising learnings and finally, specific learnings for US-India founders.
Joining me for this discussion is Pratyush Rai, Co-founder of Merlin, an AI productivity assistant that you can use for everything from getting one-click answers to summarizing & repurposing content, and much more. Pratyush is an IIT grad, worked with BCG in the past, before starting Merlin with two of his batchmates, and has now scaled it to 2M+ users. Merlin was recently covered in A16Z's top 50 AI apps where startups spend money.
TIME STAMPS:
(01:27) What spaces in AI are being overlooked?
(05:26) Building a Prosumer AI startup: the Merlin backstory
(09:41) Choosing AI problem statements to work on
(20:21) Doing effective customer discovery
(21:12) Competition & differentiation
(24:15) Creating defensibility as a Prosumer AI product
(30:33) AI front-end experiences: Browsers and Chrome Extensions
(34:03) Managing the foundation model layer: closed vs open source
(42:46) GTM Approaches for Prosumer AI: Influencers & Affiliate Marketing
(57:08) Pricing learnings for Prosumer AI startups: usage-based
(1:05:55) Fundraising learnings for Prosumer AI founders
(1:12:54) One hard-earned lesson for US-India founders
Connect with Pratyush on LinkedIn - https://www.linkedin.com/in/pratyush-rai/ and follow him on X - @pratyush_r8
*****
Check out "An Operator's Blog" (https://operator.blog/) for weekly posts on Startups, Investing, and Life.
I am a venture investor focused on the US-India corridor. Feel free to reach out on LinkedIn (https://www.linkedin.com/in/soumitrasharma5/) and follow me on X (@soumitra_sharma).
23 Oct 2025TRANSCRIPT
Agentic AI Meets Compliance: Lessons From Sprinto
This episode delves in-depth into building an AI Applications startup in the area of Compliance.
We cover everything from how to choose problem statements as a founder including a 2x2 framework for idea evaluation, approaching building in “boring” verticals, executing an effective customer discovery process, thinking through product differentiation in the Applications layer, evolution of GTM in the age of AI, new-age pricing models, present state of AI fundraising in the Bay Area, and finally, specific learnings for US-India founders.
Joining me for this discussion is Girish Redekar, Co-founder of Sprinto, a unified, AI-powered compliance platform that enterprises can use to manage compliance standards such as SOC 2, ISO 27001, HIPAA, and GDPR. Before Sprinto, Girish founded RecruiterBox in the HRTech space, which was acquired by TurnRiver Capital.
TIME STAMPS:
(01:17) What parts of the AI ecosystem are under-hyped?
(05:31) Choosing unsexy verticals: the Sprinto backstory
(11:48) Choosing the right problem statements in AI
(17:21) Product Risk vs Market Risk: a 2x2 framework for evaluating startup ideas
(21:12) Competition & differentiation
(25:41) Staying on top of rapid changes in the foundational model layer
(31:37) AI’s impact on GTM
(37:41) AI adoption apprehensions of enterprise customers
(42:27) Ground reality of AI-first pricing models
(47:03) Bay Area fundraising advice for AI Applications startups
(51:18) A hard-earned lesson for US-India founders
Connect with Girish on LinkedIn - https://www.linkedin.com/in/girishredekar/ and follow him on X - @grease_
*****
Check out "An Operator's Blog" (https://operator.blog/) for weekly posts on Startups, Investing, and Life.
I am a venture investor focused on the US-India corridor. Feel free to reach out on LinkedIn (https://www.linkedin.com/in/soumitrasharma5/) and follow me on X (@soumitra_sharma).
25 Sept 2025TRANSCRIPT
Building An AI-Infra Startup In 2025: Seed-Stage Lessons From Pipeshift
This episode delves in-depth into the present market landscape, challenges & opportunities being navigated by AI infra startups in the Bay Area.
We cover everything from parts of the AI ecosystem that are overhyped and underhyped, real AI infra pain points & apprehensions of enterprise customers, nuances of leveraging closed source vs open source models in enterprise infra, discovering gaps & wedges as a startup, creating differentiated products against competition, AI-first GTM, new-age pricing models and finally, lesser-known aspects of fundraising in the Bay Area.
Joining me for this discussion is Arko, Co-founder and CEO of Pipeshift, a YC-backed startup that's building reliable and SLA-compliant inference-as-a-service for real-time AI applications in production. He and his co-founders started building products together as undergraduates, participating in hackathons over six years ago, and today they are building a best-in-class AI infrastructure company right here in the Bay Area.
TIME STAMPS:
(01:43) What parts of the AI ecosystem are under-hyped?
(07:48) The Pipeshift backstory
(11:23) The unique value of building in the Bay Area
(16:31) Choosing the right problem statements
(21:34) Competition & differentiation
(25:22) AI infra pain points of enterprise customers today
(32:24) Staying on top of rapid changes in the foundational model layer
(42:48) Customer approach on closed vs open source models
(48:02) Ground reality of AI-first pricing models
(51:44) Early stage GTM learnings in AI infra
(59:55) Bay Area fundraising advice
Connect with Arko - https://www.linkedin.com/in/arkoc/
*****
Check out "An Operator's Blog" (https://operator.blog/) for weekly posts on Startups, Investing, and Life.
I am a venture investor focused on the US-India corridor. Feel free to reach out on LinkedIn (https://www.linkedin.com/in/soumitrasharma5/) and follow me on X (@soumitra_sharma).
Inside a recent episode
Building A Prosumer AI Startup: The Merlin Story
Published 11 Dec 2025 · Transcript excerpt
[…] And you can like use it to access ChatGPT on any website and rewrite content and so on. But because of some random coincidence, even marketers and other people started using it. And then we realized that the application is actually very horizontal. So we did not turn out to be a horizontal company. We launched it as something designed for developers and engineering managers. But somehow, everyone started using us. And when we put in a paywall, everyone started paying us. So it ended up becoming a horizontal product. Then when we really tried hard to try to narrow it off and only focus on developers or marketers and so on, we were struggling because we realized that when we are trying to cater to one persona, actually a lot of use cases which that persona also ends up using Merlin for. they get compromised. […]
Pod Engine is an independent podcast discovery and analytics service and is not affiliated with or endorsed by this podcast. Artwork and show content belong to their owners. Full legal notice.
Explore this show Podcast research with Pod Engine
