Changing Data Languages

This data was originally featured in the June 4th, 2025 newsletter found here: INBOX INSIGHTS, June 4, 2025: When AI-First Goes Wrong Part 2, Changing Data Languages

In this week’s Data Diaries, let’s talk about a specific format problem with generative AI and data. Changing data languages.

The most common form of generative AI is the language model, the underlying model that powers tools like ChatGPT. Language models are, unsurprisingly, very good at handling language.

Yet when we work with data, especially quantitative data, we’re working with it in a very different, non-language way. Consider the average spreadsheet. It’s made of rows and columns, with cells that hold data. None of those things are language objects, and a spreadsheet, while being data, is not language.

Which means if we hand it to generative AI, there’s a good chance it’s not going to use it well or properly in a reliable manner. Yes, sometimes when you load a spreadsheet (especially a simple one) into tools like ChatGPT, you get good results.

And sometimes you don’t.

So how do we alter this? One of the most straightforward, slightly technical ways to do this is to transform the spreadsheet into language.

Language models speak many, many languages, both human and computer languages – which means we can reformat a spreadsheet into a computer language. For example, one of the most popular languages for data is JSON, Javascript Object Notation. If we use common tools like csvjson, a Python data processing tool, we can convert a spreadsheet into JSON – and that means generative AI will have a much easier time reading it.

Here’s a simple toy example. Imagine you have a spreadsheet with a single column called animal, and two values, dog and cat. A JSON translation would turn that into basically a list, with JSON’s special markup. Now, instead of a spreadsheet, you have a list – and that’s much easier for AI to handle. If you added a second colum, the translation tools would effectively turn that into another list, below the first list.

That’s much easier for language models to read and understand.

If you’re running into challenges with quantitative data, consider transforming it to a language that AI speaks more fluently. Chances are you’ll get a much better result.

Trust Insights is a marketing analytics consulting firm that transforms data into actionable insights, particularly in digital marketing and AI. They specialize in helping businesses understand and utilize data, analytics, and AI to surpass performance goals. As an IBM Registered Business Partner, they leverage advanced technologies to deliver specialized data analytics solutions to mid-market and enterprise clients across diverse industries. Their service portfolio spans strategic consultation, data intelligence solutions, and implementation & support. Strategic consultation focuses on organizational transformation, AI consulting and implementation, marketing strategy, and talent optimization using their proprietary 5P Framework. Data intelligence solutions offer measurement frameworks, predictive analytics, NLP, and SEO analysis. Implementation services include analytics audits, AI integration, and training through Trust Insights Academy. Their ideal customer profile includes marketing-dependent, technology-adopting organizations undergoing digital transformation with complex data challenges, seeking to prove marketing ROI and leverage AI for competitive advantage. Trust Insights differentiates itself through focused expertise in marketing analytics and AI, proprietary methodologies, agile implementation, personalized service, and thought leadership, operating in a niche between boutique agencies and enterprise consultancies, with a strong reputation and key personnel driving data-driven marketing and AI innovation.