Databricks Co-Founder Arsalan Tavakoli-Shiraji Shares Hard-Earned Lessons on How Startups Can Land Massive Enterprise Deals

Startups in the modern technology ecosystem have arguably never faced greater or more relentless pressure to achieve rapid growth. In order to scale effectively and secure their financial futures, many of these young companies find themselves forced to target large enterprises—the deep-pocketed titans of the global IT world—as their primary customers. However, as any seasoned veteran of enterprise sales will readily attest, securing multi-million-dollar contracts with legacy institutions is an arduous, high-stakes battle fraught with hidden pitfalls and complex institutional barriers.

According to Arsalan Tavakoli-Shiraji, co-founder and senior vice president of field engineering at Databricks, one of the most critical and pervasive mistakes younger startups make is operating under the comforting illusion that their technology will practically sell itself. Databricks has evolved into an artificial intelligence and big data storage giant, currently boasting a staggering valuation of $190 billion and counting more than 70% of the Fortune 500 among its robust customer roster. Yet, the company’s current market dominance obscures humble beginnings as a tiny, unproven startup, during which Tavakoli-Shiraji was directly responsible for helping secure the enterprise deals that built its foundation.

Speaking candidly to attendees at TechCrunch Disrupt, Tavakoli-Shiraji reflected on the hard-learned lessons of his early entrepreneurial career. He noted that founders universally enter the enterprise market believing that success is entirely a function of having the absolute best technology. Through painful experience, however, he discovered that people and institutional processes ultimately matter far more than raw code.

Recalling the early, often humbling days of outreach, Tavakoli-Shiraji described managing to secure a coveted meeting with a prospective enterprise client, only to quickly realize that the executive sitting across from him was the exact same decision-maker who had recently signed a massive, career-defining mega-deal with Databricks’ primary competitor. The prospect had virtually no interest in listening to a pitch because his professional reputation and career trajectory were already staked on the rival platform. On other occasions, promising early deals fizzled out entirely during the final stages simply because the inexperienced startup was unprepared to answer rigorous legal, contractual, or enterprise-grade security questions.

To help a new generation of founders navigate these treacherous waters, Tavakoli-Shiraji outlined several vital strategies and conceptual frameworks for capturing the attention of conservative corporate buyers.

Sell under a bigger tech company’s umbrella

When Databricks was in its infancy, the startup tried repeatedly and failed miserably to secure a meeting with UBS, the massive Swiss banking institution. The cold-outreach emails went unanswered, and the doors remained firmly shut. Frustrated by the lack of traction, the startup pivoted its go-to-market strategy by forming a strategic partnership with Microsoft, integrating deeply to become an official Azure cloud partner.

That single strategic alignment meant that Databricks’ sales agreements would suddenly be covered and sanctioned under Microsoft’s established, trusted corporate terms. More importantly, it meant that Microsoft’s formidable enterprise sales force was suddenly incentivized to actively book meetings and advocate for the startup’s solutions.

All of a sudden, doors that had previously stayed locked began to open with ease. Tavakoli-Shiraji recalled that getting meetings was no longer an uphill battle, noting that partnering with an established player possessing that level of market access and distribution infrastructure proved to be a massive boon for the young company. With Microsoft’s sales team helping grease the wheels, demand snowballed rapidly, creating entirely new operational challenges—such as prospective clients demanding localized features for countries Databricks hadn’t even expanded into yet—which he characterizes as the best kind of problems for a growing company to face.

Today, the dynamic has completely inverted, with Databricks now operating as the established tech giant that works hand-in-hand with innovative startups, selling their products under Databricks’ own established terms and rigorous cybersecurity frameworks. Notable examples of this modern ecosystem approach include Hightouch, which delivers AI-powered marketing and customer-data tools and in which Databricks has made strategic investments; Addepar, a provider of sophisticated data tools designed specifically for investment professionals; and Tecton, a machine learning startup that Databricks ultimately acquired.

When selling AI, address the “four Cs”

As enterprise interest shifts overwhelmingly toward artificial intelligence, Tavakoli-Shiraji has categorized the primary concerns, hesitations, and objections voiced by corporate buyers into four distinct pillars he refers to as the "four Cs": context, choice, cost, and control.

Context addresses the enterprise’s desperate need for artificial intelligence systems that can be trained effectively—akin to onboarding a brand-new corporate employee. This means the AI must thoroughly absorb and master company-specific terminology, internal workflows, historical data, and corporate priorities.

Choice represents the enterprise buyer’s insistence on maintaining the operational freedom to switch between different AI models and underlying providers rather than being locked into a single, proprietary ecosystem. Cost centers on predictability and financial management, requiring clear mechanisms for controlling and forecasting how much an AI product will actually cost to run at scale. Finally, control encompasses the robust corporate safeguards, security boundaries, and governance protocols required to strictly govern AI behavior and protect sensitive enterprise data from leakage or misuse.

Attack the expensive legacy systems enterprises think they’re stuck with

In his discussions with corporate leadership, Tavakoli-Shiraji frequently poses a lighthearted question to Chief Information Officers, asking them to evaluate their organization’s current enterprise AI adoption on a scale of one to five. A ranking of one indicates that employees merely have casual access to consumer tools like ChatGPT, whereas a five signifies that the company has successfully integrated deep AI capabilities that have fundamentally restructured its core business processes.

Although executives universally love to claim they are sitting comfortably at a three leaning toward a four, Tavakoli-Shiraji notes that roughly 99% of the CIOs he speaks with actually reside squarely in buckets one and two. Despite the explosive market hype, legacy enterprises still largely have not figured out how to deploy autonomous AI agents in ways that generate genuine business value beyond enabling internal software engineering teams to write code slightly faster. Consequently, he cautions that it will take many years before organizations can properly and maturely deploy AI in a deeply meaningful way across traditional enterprises.

For agile startups, however, this widespread hesitation represents a massive commercial opportunity to modernize notoriously expensive legacy systems that corporations believe they are permanently trapped with. Pointing to the banking sector as a prime example, Tavakoli-Shiraji observed that virtually every major bank still relies heavily on decades-old infrastructure, such as IBM AS/400 systems and aging mainframes, which corporate IT departments have actively avoided touching for years because nobody left on staff fully understands what the underlying legacy code actually does.

Today, advanced AI agents possess the capability to step directly into these archaic environments, analyze, and systematically decompose legacy code to reveal its precise operational logic. According to Tavakoli-Shiraji, this technological breakthrough finally cracks open the door for innovative startups to build products that achieve everything from slapping modern, user-friendly interfaces onto ancient systems to replacing them entirely.