For early-stage startups, the mandate is clear: grow fast or face obsolescence. In the current economic climate, the fastest route to scale is often through the "big fish"—the Fortune 500 enterprises that command massive budgets and deep IT infrastructure. However, the graveyard of startups is littered with those who possessed revolutionary technology but failed to cross the "enterprise chasm."
Arsalan Tavakoli-Shiraji, co-founder and senior vice president of field engineering at Databricks—now a $190 billion data and AI titan—argues that the biggest mistake founders make is the belief that superior technology is a golden ticket. Speaking at TechCrunch Disrupt, Tavakoli-Shiraji offered a sobering reality check: while product quality is a prerequisite, enterprise sales are won through people, processes, and strategic positioning, not just code.
The Illusion of the "Better Mousetrap"
The narrative that "the best technology always wins" is a dangerous myth in the enterprise software space. When Databricks was in its infancy, Tavakoli-Shiraji assumed that if their data processing speeds outperformed competitors, the contracts would follow. He quickly learned that enterprise procurement is rarely a meritocracy of engineering; it is a complex web of risk mitigation, existing relationships, and internal politics.
The Personal Stakes of Enterprise Procurement
Tavakoli-Shiraji recounts a formative moment early in his career: securing a meeting with a high-level executive at a target firm, only to realize the individual had recently staked their reputation—and their career—on a major competitor’s solution.
"He had no interest in listening to me because he bet his career on that competitor," Tavakoli-Shiraji noted. This underscores a fundamental truth: enterprise buyers are not just buying software; they are buying security, predictability, and a buffer against professional failure. If a startup doesn’t account for the buyer’s internal incentives, the best technical pitch in the world will fall on deaf ears.
The Operational Blind Spot
Beyond the psychological hurdles, many startups collapse under the weight of administrative inertia. Deals that should have been "home runs" often fizzled because the startup was unprepared for the grueling gauntlet of enterprise legal, security, and contractual requirements. To an enterprise, a startup is a risk vector. Failing to provide a robust answer to a security questionnaire is often interpreted as a lack of maturity, signaling to the client that the startup might not be around in five years to support the product.
Chronology: From Scrappy Startup to Enterprise Titan
The transformation of Databricks from a niche data-processing play into a platform serving over 70% of the Fortune 500 provides a masterclass in enterprise entry.
- The Early Struggles: Initially, Databricks operated as a standalone entity trying to break into the banking sector. The cold-calling approach to institutions like UBS proved largely unsuccessful. The startup lacked the "trusted vendor" pedigree required to pass through the iron-clad procurement gates of global financial institutions.
- The Pivot to Partnership: Recognizing that they were hitting a wall, the leadership team shifted strategy. Instead of selling directly, they became an Azure cloud partner. By nesting their product under Microsoft’s umbrella, Databricks gained immediate legitimacy.
- The Distribution Catalyst: The partnership didn’t just provide "brand cover"; it provided a sales force. Microsoft’s sales team, already deeply embedded in Fortune 500 accounts, began advocating for Databricks. As Tavakoli-Shiraji recalls, "All of a sudden, that meeting wasn’t hard to get. It was a huge boon to us that we were able to partner with somebody who had that access and distribution."
- Scale and Ecosystem: Having matured, Databricks has now reversed the roles. They now act as the enterprise anchor for smaller startups like Hightouch and Addepar, providing the security and contractual framework that allows these companies to bypass the initial "risky vendor" assessment phase.
The Strategic Framework: Winning the Enterprise
For founders looking to replicate this success, Tavakoli-Shiraji outlines three specific pillars for engagement.
1. Leverage the "Umbrella" Strategy
Startups should stop viewing themselves as standalone entities and start viewing themselves as part of a larger ecosystem. Aligning with an established tech giant—whether through cloud marketplaces or official partner programs—effectively "outsources" the trust building. When a customer buys a startup’s software through a Microsoft, AWS, or Databricks agreement, they are relying on the established vendor’s legal and security vetting, which is a massive friction-reducer.
2. The Four Cs of AI Adoption
As AI becomes the primary focus for enterprise budgets, the complexity of the sale has shifted. Tavakoli-Shiraji advocates for addressing the "Four Cs" to convert prospects:
- Context: The AI must be able to absorb the company’s unique terminology and internal processes. It needs to act like an employee who understands the corporate culture, not a generic chatbot.
- Choice: Enterprises are terrified of vendor lock-in. A successful solution allows the enterprise to switch between different large language models (LLMs) rather than being trapped in a proprietary loop.
- Cost: Transparency is paramount. Enterprises need a predictable way to forecast what an AI deployment will cost as it scales.
- Control: This is the most critical factor. Enterprises require robust, immutable safeguards that prevent the AI from leaking sensitive data or hallucinating facts that could lead to regulatory or reputational damage.
3. Targeting the "Legacy Debt"
While many startups are chasing the "shiny object" of new generative AI features, the real gold mine lies in the modernization of legacy systems. Every bank, insurance company, and logistics firm is currently sitting on "zombie" infrastructure—mainframe code written decades ago that nobody dares touch for fear of breaking the entire business.
"If you go to any bank today, they still have [an IBM] AS/400 and mainframes," Tavakoli-Shiraji notes. "Nobody wanted to touch it because nobody knew what that code did." By using modern AI agents to "decompose" and document this legacy code, startups can offer a bridge to the future, effectively creating a path for enterprises to migrate away from expensive, fragile legacy environments.
Implications: The State of Enterprise AI
Tavakoli-Shiraji’s assessment of the current AI landscape is intentionally provocative. While the tech industry is buzzing with the potential of AI agents, he estimates that the vast majority of enterprises remain in the early stages of adoption.
"Everybody loves picking, ‘I’m three leaning towards four’ [on a scale of 1 to 5]," he says, referring to their AI maturity. "But 99% of the ones I talk to are in bucket one and two."
The implication for founders is clear: the market is not yet saturated. Most enterprises have barely moved beyond using AI for simple code-writing tasks. There is a massive, multi-year window of opportunity for startups that can solve the "meaningful deployment" problem. The firms that win won’t necessarily be the ones with the most advanced LLMs; they will be the ones that can navigate the bureaucratic, legal, and operational realities of the enterprise.
Conclusion: A Shift in Perspective
The journey from a "tiny startup" to an industry giant is rarely a linear progression of technical breakthroughs. It is a series of strategic compromises, alliances, and a deep understanding of the customer’s internal pain points. As Tavakoli-Shiraji’s journey proves, if you want to land the enterprise, you must stop selling to the IT department and start selling to the business—by de-risking the future and cleaning up the past.
For the next generation of founders, the lesson is simple: build for the tech, but sell for the human, the legal department, and the bottom line.
