Data Vendors Win Again

Ease of data querying is converging to "zero effort” and this is exactly what every data vendor would want.

LLM’s or large language models like ChatGPT became the topic of the day in the early months of 2023. They still are.

Broadly there are two camps commenting on LLMs:  1) some say LLMs are a profound new territory, and 2) some say “hey this is a joke, it lies, watch me trick it.”

Both views are correct in some regards, but the real winners in my view are the data vendors. I spend a lot of time with data ninja’s who can do magic with R or Python or VBA. They can cajole data to go where they need it, transform it and make it usable. It is truly a small miracle.

But in the past, if you were not a data ninja, and you’re just a normal person, you would be downloading CSVs via the classic data vendor interface and then the brute force work in Excel would begin.

LLMs change this paradigm, let’s dive in.

I’m pretty skilled at Bloomberg and FactSet, but using their interface, a normal person ends up with a spreadsheet.

With LLMs and similar technologies, you can use natural language to ask for what you want and in the form you want.

This is taking data vendor’s business model and turning it into a more nimble system, something more people can use, even non-coders.

For example, you can use ZoomInfo’s API to quickly get detailed information about a company using some simple code, written by AI.

So the winners here are of course those making the Natural Language processing tools, but also the data vendors as well. Why data vendors? Because the user is no longer constrained by the fat client, thin client, or their lack of coding skills.

FactSet recently announced a strategy called “Open FactSet” wherein FactSet would enable its data outside of its client UI. I don’t know if Open FactSet has taken off just yet, but I doubt it did because it required some coding chops. If you know differently, please comment., now that AI can write the script to query FactSet or Bloomberg or ZoomInfo (in my example) while skipping the part where the user interfaces with their local client and gets a CSV to work with, I think this is truly profound.

With AI and LLMs, practitioners will get to skip the intermediate step of downloading a CSV file and doing the brute force work of organizing the data in Excel.

I think this makes those who own/deliver the data even more powerful. Ease of data querying is converging to "zero effort” thanks to LLMs and I believe this is exactly what every data vendor would want.

Follow me on Twitter @DrewMMeister or on LinkedIn here:

I'm always open to connect and converse.