Sigma’s AI Strategy Starts With a Simple Assumption That “AI Is Always Wrong”

“It doesn’t matter whether it works. It could be the best executed terrible idea ever.” Mike Palmer , CEO, Sigma Computing, offers one of the most direct critiques of AI-powered business intelligence to surface this year. He is talking specifically about the rising excitement around text-to-SQL bots, the AI tools that aim to convert natural language into perfect database queries. “This is not what AI was supposed to do for end users,” Palmer says. “We think these products are a joke.” That statement might seem surprising coming from the CEO of a company that recently crossed $100 million in annual recurring revenue and closed a $200 million Series D. But for Sigma, dismissing one of the most hyped use cases in AI is consistent with a broader strategy. Instead of chasing flashy features, the company is focused on rethinking how people actually work with data. Palmer’s critique is grounded in the belief that natural language queries often oversimplify how people interact with data. He argues that asking users to think in partial SQL logic wrapped in plain English does not make analytics more accessible. “Very few people think in terms of SQL,” he says. “No one says, ‘I SELECT the poppy seed bagel FROM the second shelf WHERE toasted equals true.’” Sigma's latest product, called Ask Sigma, reflects a different view of AI’s potential. Rather than trying to translate vague questions into perfect answers, it reveals exactly how each answer is formed. When a user inputs a question, the system breaks down the full path it took to respond. It shows the data sources used, the filters applied, and the logic it followed. Users can examine each step, revise it, or rephrase their question with full context. Palmer describes this level of visibility as essential to building trust with AI. “If I told you to go to your investment committee and said you should trust a number with 86 percent confidence, would you? No.” Ask Sigma also surfaces related data points to help refine thinking. A request for revenue might also include metrics like bookings, customer segments, average contract value, or historical trends. The product is designed to encourage exploration and iteration, helping users become more effective at analytical thinking over time. Much of Sigma’s strategy stems from its rejection of conventional business intelligence. Palmer is blunt in his assessment. “Forget the past 20 years of BI. It’s been boring.” Founded in 2014 by Rob Woollen and Jason Frantz , Sigma set out to do something different. With backing from Sutter Hill Ventures, the same incubator behind Snowflake, the team chose not to build another dashboarding tool. Early on, even JP Morgan told them the world didn’t need one. They agreed. Instead of chasing visual flair, Sigma focused on the most widely used business tool in the world: the spreadsheet. Palmer calls it the most common BI feature in existence. Rather than forcing users to adapt to a new interface, Sigma made its interface feel familiar while transforming what that spreadsheet could do. Sigma connects directly to cloud data warehouses like Snowflake, Databricks, and Redshift. Users can query data in real time and even write data back into the warehouse using features like Input Tables. That capability has made it especially valuable for teams updating forecasts, managing budgets, or adjusting operational metrics. “Legacy BI was built for passive consumption. Look, but don’t touch,” Palmer says. “That is not enough anymore.” The core of Sigma’s design philosophy is rooted in building for people who are not data scientists or SQL experts. These are finance managers, marketing leads, HR professionals, and other roles that depend on data but lack the technical tools to navigate it deeply. Palmer poses a simple question: “Do you have a working spreadsheet on your laptop today?” His point is that nearly everyone does. Yet for years, this group has been overlooked by enterprise software, which was often designed by and for engineers. By blending familiar interfaces with the power of cloud-native analytics, Sigma creates a space where both technical and non-technical users can work together. Teams can explore, analyze, and act on data without constant handoffs or dependency on analysts. One of the most interesting aspects of Ask Sigma is its ability to help users refine their thinking. By exposing the logic behind each result and suggesting adjacent metrics, it enables users to see how analysts frame questions and how answers evolve through exploration. The product is designed not only to deliver outcomes, but to build capability. Sigma believes that true value in analytics does not come from getting a quick answer but from developing the intuition to ask better, more relevant questions. Palmer explains, “AI is always wrong for one of two reasons. Either the model is wrong or the question is unclear. We address both.” Sigma’s bet on transparency, usability, and data literacy is paying off. The

Sigma’s AI Strategy Starts With a Simple Assumption That “AI Is Always Wrong”

Keep reading. It's free.

Related stories