Databricks Co-Founder’s Playbook for Winning Enterprise Customers

Summary

In remarks delivered the week before October 26, 2026, Databricks co-founder and field engineering chief Arsalan Tavakoli-Shiraji said enterprise sales depend more on people and processes than superior technology. Databricks, now valued at $190 billion and used by more than 70% of the Fortune 500, initially lost deals because buyers had backed competitors or raised legal, contractual and security questions the startup could not answer. After Databricks failed to secure a UBS meeting, an Azure partnership placed it under Microsoft’s commercial terms and gave Microsoft salespeople incentives to arrange meetings and advocate for it, generating demand even in countries Databricks had not entered.

Databricks now provides a similar sales and cybersecurity umbrella for Hightouch, an AI marketing and customer data provider in which it invested, Addepar, which serves investment professionals, and Tecton, a machine learning startup it later acquired. Tavakoli-Shiraji said AI vendors must address four Cs: company context, model choice, operating cost, and control of behavior and sensitive data. Although executives often rate their AI maturity as three leaning toward four on a one to five scale, he estimated 99% remain at levels one or two, often merely providing ChatGPT, with valuable agents largely limited to coding. He expects meaningful enterprise deployment to take years, but sees an opening for agents that decipher IBM AS/400 and mainframe code, enabling modern interfaces or complete replacement.