How Startups Can Win Big Enterprise Customers: Lessons from Databricks
Startups today face an unprecedented pressure to grow rapidly, often by landing contracts with large enterprises—the deep-pocketed giants of the IT world. However, winning these lucrative deals is far from straightforward. Arsalan Tavakoli-Shiraji, co-founder and senior vice president of field engineering at Databricks, shared valuable insights into the common pitfalls startups encounter and how to overcome them.
Many young startups make the critical mistake of assuming that their technology will sell itself. “Everybody when they’re selling to enterprises at startups… we think it’s all about the tech and having the best tech,” Tavakoli-Shiraji told attendees at TechCrunch Disrupt. Yet, as he emphasizes, “people and processes matter the most.”
Databricks, now an AI and big data storage powerhouse valued at around $190 billion with over 70% of the Fortune 500 as customers, began as a small startup facing these exact challenges. Tavakoli-Shiraji recalls early hurdles, including losing prospects who had already committed to competitors and failing to address enterprise concerns related to legal, contractual, and security issues.
Sell Under a Bigger Tech Company’s Umbrella
One of the most effective strategies for startups is aligning with larger, established tech companies. Early in Databricks’ journey, the company struggled to secure a meeting with UBS, a major Swiss bank. This changed after forming a partnership with Microsoft as an Azure cloud partner. This alliance allowed Databricks to leverage Microsoft’s sales agreements and credibility, making meetings with enterprise prospects significantly easier to obtain.
“All of a sudden, that meeting wasn’t hard to get,” Tavakoli-Shiraji recalled, highlighting the power of partnering with companies that have extensive access and distribution channels. This partnership model not only opened doors but also created new opportunities, even as it introduced challenges like meeting feature demands for international markets.
Today, Databricks itself has become a major platform for startups, offering them the ability to sell under Databricks’ own terms and cybersecurity framework. Notable examples include Hightouch, which offers AI-driven marketing and customer data tools; Addepar, providing data tools for investment professionals; and Tecton, a machine learning startup later acquired by Databricks.
When Selling AI, Address the “Four Cs”
When pitching AI solutions to enterprises, Tavakoli-Shiraji advises startups to understand and directly address what he calls the “four Cs”: context, choice, cost, and control.
“Context” refers to the AI’s ability to be trained similarly to a new employee, adapting to a company’s specific terminology, processes, and priorities. This personalization is critical for enterprise adoption.
“Choice” highlights the importance of providing customers with the freedom to switch AI models rather than locking them into a single provider, which can cause vendor lock-in concerns.
“Cost” pertains to transparency and control over how much the AI solution will cost to operate, an essential factor for enterprises managing budgets tightly.
Finally, “control” involves ensuring robust safeguards that govern the AI’s behavior and protect sensitive information, a paramount concern within highly regulated industries.
Attack the Expensive Legacy Systems Enterprises Think They’re Stuck With
Many enterprises are still in the early stages of AI adoption. Tavakoli-Shiraji humorously rates CIOs on a scale from one to five regarding their AI maturity, where one means employees only have access to tools like ChatGPT, and five represents full integration of AI capabilities that transform business processes. Most enterprises currently fall between one and two.
This slow adoption presents a significant opportunity for startups to modernize costly legacy systems. Tavakoli-Shiraji points out that banks and other institutions often rely on outdated mainframes like IBM’s AS/400, systems no one wants to touch due to the complexity and opacity of their code.
AI agents now have the capability to analyze and decompose these legacy codes, making it possible to create modern interfaces or even replace these systems entirely. This not only reduces costs but also paves the way for enterprises to innovate and become more agile.
Looking forward, Tavakoli-Shiraji predicts that meaningful enterprise AI deployment will take years, but startups that focus on these areas can position themselves as essential partners in this transformation.
For a deeper dive into Tavakoli-Shiraji’s experiences and his unconventional tactics—including a story about posing as a lawyer to close a deal—read the full interview Here.
When you purchase through links in our articles, we may earn a small commission. This doesn’t affect our editorial independence.
