Building an Infrastructure and Data Moat

In this episode, we sit down with Anish Agarwal, CEO and Co-founder of Traversal, for a deep dive into the substance behind the AI noise. Anish, a former Columbia professor and researcher in causal AI, shares his unique journey from academia to founding an organization disrupting the site reliability engineering (SRE) and observability space.

We explore the critical difference between "AI wrappers" and companies building genuine infrastructure, the emergence of the "Forward Deployed Engineer" in the sales pod, and how to identify technical moats in a world where models are rapidly evolving.

🙌 Thanks To Our Sponsor! Aurasell AI: https://www.aurasell.ai 🏹

Key Topics Covered

00:00 - Intro 01:53 – Traversal: the AI Site Reliability Engineer

03:14 – AI Wrappers vs. AI Native

04:58 – The Importance of a Technical Moat

07:27 – Human Error and Data Scarcity

08:56 – From Experimental Budgets to Production Reality

14:42 – The Rise of the "Forward Deployed Engineer"

22:50 – From PhD to Founder 34:50 – Rethinking the Observability Stack

💥 3 Biggest Lessons: A Technical Moat Lives in Infrastructure and Data: Anish argues that a sustainable AI company must be an infrastructure and data company at its core. If your value is purely in prompt engineering or simple workflows, model companies (like OpenAI) will eventually "eat" those features. A true moat is developed through context engineering—managing petabytes of data in a way that is consumable for LLMs—and functioning on systems elements haven't been trained on, such as private observability logs.

Target Workflows Where Humans Are "Bad": When assessing the longevity of an AI solution, look at the tasks it automates. If a human is already good at the task, data for that workflow is easily collected and fed into public models, making the solution easy to replicate. The most defensible AI companies solve superhuman problems—tasks humans are inherently bad at or data types (like time-series logs) that are not publicly available for training.

The Chasm Between Pilot and Production: We are entering a phase where the "experimental AI budget" is drying up. Winners in the next two years will be companies that can connect their year-long pilots to hard labor or software spend. This requires tight technical scoping and a "pre-sales mindset" that lasts long after the initial check is signed to ensure the product survives the first renewal cycle.

💬 Notable Quotes

"As much as we’ve been an AI company, we’ve been an infrastructure and data company."

"If you are creating prompts or simple workflows, the models will just keep eating those up."

"The un-sexy part of AI—how you actually deploy this and comply with regulation—that is what is now maturing."

"Sales is a first-class citizen in the company... a less good product can win if you have the right team to bring it to market."

"I’ve become a student of sales... it eventually becomes a systems engineering problem."

🙌 Thanks for listening!

This episode was hosted by Simon Kouttis and Oli Kuehne, founders of Hunters & Unicorns, and post-produced by the team at videoforce.pro.

If you enjoyed this episode, please drop a like/share and subscribe to our channel!

🦄 Connect with Hunters & Unicorns Website: http://huntersandunicorns.com Twitter: http://twitter.com/HuntersUn1corns Instagram: http://instagram.com/huntersandunicorns Blog: http://huntersandunicorns.com/blog #softwaresales #huntersandunicorns #playbookuniverse