Arsalan Tavakoli-Shiraji, co-founder and senior VP of field engineering at Databricks, recently shared his insights on how startups can successfully land large enterprise customers. Speaking at TechCrunch Disrupt, he emphasized that having superior technology alone is not enough to win big contracts with enterprises. Instead, startups must focus on building strong people and processes, navigating complex legal and security demands, and understanding the enterprise buying journey—lessons Databricks learned firsthand as it grew from a small startup to an AI and data giant valued at $190 billion with over 70% of the Fortune 500 as customers.

A key strategy Tavakoli-Shiraji highlighted was leveraging partnerships with established tech companies to gain access to enterprise clients. Early on, Databricks struggled to engage with major firms like UBS until it partnered with Microsoft as an Azure cloud partner. This relationship allowed Databricks to operate under Microsoft’s sales umbrella and contracts, quickly opening doors that were otherwise closed. The Microsoft sales team helped schedule meetings and advocate for Databricks, illustrating how aligning with a larger player’s distribution can accelerate enterprise customer acquisition.

When it comes to selling AI products to enterprises, Tavakoli-Shiraji identified four core concerns—context, choice, cost, and control—that startups must address. Enterprises want AI solutions that adapt to their specific methods and language (context), offer flexibility to switch models (choice), keep expenses manageable (cost), and ensure strict governance over AI’s actions and data (control). Addressing these "four Cs" directly can help startups overcome skepticism and demonstrate value to cautious enterprise buyers.

Lastly, Tavakoli-Shiraji pointed out the significant opportunity for startups to modernize legacy systems within enterprises. Many large companies still rely on outdated platforms like IBM AS/400 mainframes, whose operations are poorly understood by current staff. Advanced AI tools can analyze and deconstruct these legacy codes, enabling the creation of modern interfaces or complete replacements. This modernization is a critical avenue for startups to prove their worth and drive meaningful AI integration in enterprises that remain early in their digital transformations.