For modern startups, the pressure to achieve rapid growth has never been more intense. To survive and scale in today’s fiercely competitive technology ecosystem, young companies are frequently forced to target enterprise clients—the deep-pocketed titans of the global IT world. However, as any veteran B2B founder will readily attest, securing lucrative contracts with massive multinational corporations is an uphill battle fraught with hidden pitfalls, bureaucratic hurdles, and unforgiving procurement cycles.
According to Arsalan Tavakoli-Shiraji, co-founder and senior vice president of field engineering at Databricks, one of the most critical and frequent mistakes made by early-stage entrepreneurs is the naive belief that superior technology will naturally sell itself. Today, Databricks stands as an undisputed AI and big data storage colossus, boasting a staggering valuation of $190 billion and counting more than 70% of the Fortune 500 among its active customer base. But the company’s current market dominance is a far cry from its humble origins as a tiny, resource-constrained startup where Tavakoli-Shiraji was personally tasked with hunting down and closing the enterprise deals that kept the enterprise afloat.
Speaking to founders and industry professionals at TechCrunch Disrupt, Tavakoli-Shiraji reflected on the hard lessons learned during those formative years. While acknowledging that foundational technology must function exceptionally well, he issued a stark warning that technical brilliance alone is never enough to win over risk-averse enterprise buyers.
"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 the audience. Yet, as he discovered through painful trial and error, people and processes ultimately matter the most in enterprise sales.
Recalling the early days of trial and error, Tavakoli-Shiraji shared memories of securing hard-fought meetings with prospective clients, only to realize the person across the table had already signed a massive, career-defining mega-deal with Databricks’ primary competitor. In those moments, the prospect had virtually no interest in listening to an alternative pitch because they had staked their professional reputation on a rival solution. Other early deals routinely fizzled out simply because the young startup was fundamentally unprepared to navigate complex legal, contractual, and cybersecurity inquiries thrown their way by corporate legal departments.
Drawing from these formative experiences, Tavakoli-Shiraji outlined core strategies and perspectives that startups must adopt if they hope to break through enterprise barriers and land high-value customers.
Sell under a bigger tech company’s umbrella
In the early days of Databricks, the fledgling company desperately pursued a meeting with UBS, the prominent Swiss banking institution, but repeatedly hit a brick wall. Traditional cold outreach yielded nothing. The breakthrough only arrived when Databricks forged a strategic partnership with Microsoft, integrating its platform to become an official Azure cloud partner.
That single partnership fundamentally transformed the startup’s market position. Alignment with Microsoft meant that Databricks’ sales agreements were suddenly covered seamlessly under established Microsoft enterprise terms, reassuring cautious corporate buyers. More importantly, it meant that Microsoft’s formidable enterprise sales force actively booked meetings and advocated for the startup’s capabilities to their existing clients.
"All of a sudden, that meeting wasn’t hard to get," the co-founder recalled, emphasizing the monumental shift in market access. "It was a huge boon to us that we were able to partner with somebody who had that access and distribution."
With Microsoft’s expansive sales apparatus helping drive momentum, Databricks found itself pulled into opportunities across industries at an unprecedented pace. While this sudden influx of enterprise demand created entirely new operational challenges—such as prospective clients demanding compliance features for countries where Databricks had not yet established a presence—they were, as Tavakoli-Shiraji noted, undeniably good problems to have.
The ecosystem has since come full circle. Today, Databricks has evolved into the established industry giant that actively works with and nurtures younger startups, offering them the ability to sell their specialized products under Databricks’ trusted corporate umbrella, complete with established commercial terms and robust cybersecurity frameworks. Notable examples of this modern ecosystem strategy include Hightouch, which delivers AI-powered marketing and customer data tools and counts Databricks among its investors; Addepar, a provider of specialized data infrastructure for investment professionals; and Tecton, a machine learning startup that Databricks ultimately acquired.
When selling AI, address the "four Cs"
As enterprise software conversations shift decisively toward artificial intelligence, startups attempting to sell AI solutions must navigate a distinct set of corporate anxieties. Tavakoli-Shiraji categorizes the primary concerns and objections raised by corporate buyers into four essential pillars known as the "four Cs": context, choice, cost, and control.
Context refers to the enterprise requirement for AI systems that can be trained effectively, functioning similarly to onboarding a new human employee. Corporate buyers want AI that absorbs company-specific terminology, internal workflows, operational priorities, and institutional knowledge rather than relying solely on generic pre-trained models.
Choice represents the enterprise’s strong desire for architectural flexibility. Large organizations want the freedom to swap out underlying AI models as technology rapidly evolves, rather than being locked into a proprietary single-provider ecosystem that could become obsolete or expensive.
Cost centers on predictability and financial governance, focusing on how companies can monitor, manage, and rein in the operational expenses associated with running heavy AI workloads at scale.
Finally, control encompasses the crucial guardrails, security protocols, and compliance mechanisms required to govern AI behavior reliably while safeguarding sensitive corporate data, intellectual property, and customer privacy.
Attack the expensive legacy systems enterprises think they’re stuck with
Assessing the current state of enterprise artificial intelligence adoption, Tavakoli-Shiraji frequently poses a lighthearted question to chief information officers, asking them to rate their organization’s AI maturity on a scale from one to five. On this scale, a rating of one signifies an organization where employees merely have informal access to tools like ChatGPT, while a rating of five represents a company that has successfully embedded advanced AI capabilities deeply into its core business processes, fundamentally altering how it operates.
"And everybody loves picking, ‘I’m three leaning towards four,’" Tavakoli-Shiraji observed. "But 99% of the ones I talk to are in bucket one and two."
According to his industry observations, the vast majority of enterprises have yet to figure out how to deploy autonomous AI agents in ways that generate genuine business value beyond allowing internal software engineering teams to write code slightly faster. He cautions that it will take many years before corporations can properly and securely deploy AI in a deeply meaningful, enterprise-wide capacity.
For nimble startups, this widespread hesitation opens up a massive strategic window: modernizing the expensive, entrenched legacy systems that legacy enterprises often feel permanently trapped using.
"If you go to any bank today, they still have [an IBM] AS/400 and mainframes," Tavakoli-Shiraji noted, highlighting a ubiquitous corporate pain point. "Nobody wanted to touch it because nobody knew what that code did."
However, modern advancements in generative AI and autonomous agents have changed the equation entirely. These systems can now inspect, analyze, and systematically decompose legacy codebases to reveal exactly how decades-old systems function. Tavakoli-Shiraji emphasizes that this capability unlocks unprecedented opportunities for inventive startups to build products capable of doing everything from layering modern, intuitive user interfaces over ancient banking mainframes to replacing outdated legacy infrastructure altogether.
