Is Tabular Data Still Relevant?

Tabular data isn’t getting the attention it deserves, even though it often comprises the bulk of an organization’s data and contains its competitive edge and unique intellectual property – unlike generic AI tools such as ChatGPT. Furthermore, experience shows that tabular data is the most valuable data source for sales, marketing, churn management, operations, and risk management, among other business use cases.

So why the bias against tabular data? Text, voice, and images grab the lion’s share of media attention. Tabular data’s popularity challenges start with the lack of off-the-shelf foundational models and continue with the underperformance of neural networks (considered more cutting-edge and, therefore, attention-worthy than other algorithms) on tabular data.

Even the data science industry, which should know better than the public, tends to underestimate the relevance of tabular data for AI. There’s a natural tendency to conflate business intelligence, not AI, with tabular data, and snobbery about tabular data not being as sophisticated. It has been historically easy to build tabular data pipelines to a minimally viable standard, but extremely challenging to build them for optimal business value generation, which is why many tabular data use cases are not meeting expectations. 

Deep learning and unstructured data are not the solution to historically underperforming AI projects. Rather, we need to rethink how we execute AI projects using tabular data.

Don’t underestimate the value of tabular data. It’s time to unlock the untapped business value that lies within your database.