Sonny Valindra
The relentless pursuit of rapid growth has placed unprecedented pressure on startups, compelling many to target large enterprises – the financial powerhouses of the IT landscape – as key customers. However, securing these substantial contracts is a notoriously arduous undertaking, a sentiment echoed by seasoned professionals in the enterprise sales domain. Arsalan Tavakoli-Shiraji, Co-founder and Senior Vice President of Field Engineering at Databricks, a global leader in AI and big data analytics valued at $190 billion and boasting over 70% of the Fortune 500 as clients, shared critical insights at TechCrunch Disrupt last week, highlighting a common misstep among emerging companies: the overreliance on technology to sell itself.
Tavakoli-Shiraji, who played a pivotal role in securing Databricks’ initial major deals during its nascent stages, emphasized that while robust technology is fundamental, it is by no means sufficient. "Everybody when they’re selling to enterprises at startups… we think it’s all about the tech and having the best tech," he stated during his address at the prominent technology conference. "But," he cautioned, "people and processes matter the most." His journey, he recounted, was paved with hard-won lessons. He recalled instances where crucial meetings were secured only to find the prospect had already committed to a competitor, rendering his pitch moot due to pre-existing allegiances and career investments. Early deals also faltered due to a lack of preparedness in addressing critical legal, contractual, and security vetting processes inherent to enterprise procurement cycles.
The insights shared by Tavakoli-Shiraji at TechCrunch Disrupt, a premier event for startups, investors, and industry leaders that showcases groundbreaking innovation and fosters critical discussions on the future of technology, come at a time when enterprise adoption of AI and advanced data solutions is accelerating. The event, held annually, typically attracts thousands of attendees and features keynotes, panels, and startup competitions, providing a vital platform for such strategic dialogues.
Leveraging Partnerships for Market Access
One of Tavakoli-Shiraji’s foremost recommendations for startups aiming to penetrate the enterprise market is to strategically align themselves under the umbrella of established technology giants. He cited Databricks’ early struggles to secure a meeting with the prominent Swiss bank, UBS. The breakthrough came after forging a partnership with Microsoft, becoming an Azure cloud partner. This collaboration fundamentally altered their sales approach. Under this new arrangement, sales agreements were integrated within Microsoft’s existing framework, simplifying the procurement process for enterprise clients.
Crucially, the partnership empowered Microsoft’s sales force to actively advocate for Databricks, thereby facilitating introductions and securing vital meetings that had previously been elusive. "All of a sudden, that meeting wasn’t hard to get," Tavakoli-Shiraji recalled. "It was a huge boon to us that we were able to partner with somebody who had that access and distribution." This symbiotic relationship proved transformative, propelling Databricks into broader market conversations and leading to an influx of opportunities. While this rapid expansion presented new challenges, such as meeting demand for features in regions where Databricks had yet to establish a presence, these were characterized as "good problems to have," indicative of significant market traction.
This strategy of co-selling or reselling through larger partners is a well-documented avenue for startups seeking to bypass the significant hurdles of direct enterprise sales. Companies like Salesforce, through its AppExchange, and Amazon Web Services (AWS) Marketplace, provide similar ecosystems where emerging vendors can gain visibility and access to a vast customer base. The success of this model hinges on the value proposition offered to both the larger partner and the end customer, typically involving a synergistic technology offering that enhances the partner’s own portfolio.
Today, Databricks itself embodies this model, now acting as the established entity that supports emerging companies. Startups like Hightouch, a provider of AI-powered marketing and customer data tools (in which Databricks has also invested), Addepar, offering data solutions for investment professionals, and Tecton, a machine learning startup later acquired by Databricks, have all benefited from being integrated and sold under Databricks’ established terms and robust cybersecurity framework. This evolution underscores the cyclical nature of market dynamics, where successful startups eventually become platforms for the next generation of innovators.
Addressing the "Four Cs" in AI Sales
When venturing into the complex domain of AI sales to enterprises, Tavakoli-Shiraji advocates for a structured approach, focusing on addressing what he terms the "four Cs" – the primary concerns and objections typically raised by potential enterprise clients. These are: Context, Choice, Cost, and Control.
Context refers to the AI’s ability to adapt and learn the specific nuances of an organization, akin to onboarding a new employee. This involves understanding the company’s unique terminology, operational processes, strategic priorities, and data landscape. An AI solution that can seamlessly integrate and learn within this specific environment is far more likely to be adopted than a generic, one-size-fits-all offering. The ability of an AI to ingest and interpret domain-specific language and workflows is paramount for its effective deployment.
Choice addresses the enterprise’s desire for flexibility and avoidance of vendor lock-in. Customers want assurance that they are not irrevocably tied to a single AI model or provider. This necessitates offering architectures that allow for the integration of various AI models, or the ability to switch between them as needs evolve or superior alternatives emerge. This principle aligns with the broader trend towards open standards and interoperability in enterprise software.
Cost encompasses the critical factor of financial predictability and management. Enterprises require a clear understanding of the total cost of ownership, not just the initial purchase price, but also the ongoing operational expenses associated with running the AI product. This includes compute, storage, maintenance, and potential scaling costs. Transparency and robust cost management tools are therefore essential selling points.
Control pertains to the essential safeguards and governance mechanisms surrounding AI behavior and the protection of sensitive data. Enterprises are acutely aware of the risks associated with AI, including potential biases, unintended consequences, and data breaches. Demonstrating a strong commitment to security, ethical AI development, and compliance with relevant regulations (such as GDPR or CCPA) is non-negotiable. This includes providing clear audit trails and mechanisms for oversight of AI decision-making processes.
Disrupting Legacy Systems with Modern AI
Tavakoli-Shiraji also offered a pragmatic perspective on the current state of enterprise AI adoption, humorously probing Chief Information Officers (CIOs) about their AI maturity on a scale of one to five. While many aspire to be at a "three leaning towards four," indicating a strong adoption of transformative AI capabilities, Tavakoli-Shiraji asserted that the vast majority of organizations he encounters are still at a "one or two." This suggests that many enterprises are still in the early stages of exploring AI, primarily limited to basic applications like empowering engineering teams to write code more efficiently using tools like ChatGPT, rather than fully integrating AI into core business processes to drive fundamental change. He projected that widespread, meaningful enterprise-wide AI deployment will likely take "many, many years."
This assessment highlights a significant opportunity for startups that can bridge this gap. A key area for innovation lies in modernizing expensive and often unwieldy legacy systems that enterprises continue to rely upon. Tavakoli-Shiraji pointed to the persistence of outdated infrastructure, such as IBM AS/400 systems and mainframes, within major financial institutions. These systems, often decades old, are frequently poorly documented and difficult to manage, leading to a reluctance to modify or replace them.
However, the advent of advanced AI agents capable of understanding and decomposing complex legacy code presents a transformative solution. These AI agents can analyze and interpret the intricacies of these older systems, paving the way for new products that can either provide modern interfaces for these archaic platforms or facilitate their complete replacement. This capability addresses a significant pain point for enterprises, offering a path to unlock trapped value and improve operational efficiency without the prohibitive risks often associated with manual code analysis and system migration. The ability of AI to decipher and interact with COBOL or Fortran code, for instance, can unlock decades of business logic that would otherwise be inaccessible.
The implications of this trend are substantial. Startups that can develop AI-powered tools for legacy system modernization can tap into a massive market opportunity. This could range from AI-driven code translation and refactoring services to intelligent migration platforms that automate the process of moving data and functionality to modern cloud-native architectures. The demand for such solutions is likely to grow as enterprises face increasing pressure to digitize and innovate, while simultaneously managing their existing technological debt.
Tavakoli-Shiraji concluded his presentation by sharing further insights into his predictions for the future of enterprise AI adoption. He also recounted a memorable anecdote from Databricks’ early days, detailing a time when he humorously posed as a lawyer to secure a crucial enterprise contract, underscoring the often unconventional approaches required in the high-stakes world of startup sales. The full interview, offering a deeper dive into these strategies and experiences, was made available for attendees and the wider tech community.
