CONTRIBUTOR
Chief Content Officer,
Techstrong Group

Most organizations have been investing in data analytics for decades, so with the rise of generative artificial intelligence (AI) platforms there’s a clear need to better delineate what type of platform to use for varying types of use cases. Most organizations are not going to abandon investments in data analytics that are the core of many digital business transformation initiatives simply because a platform such as ChatGPT now exists.

In fact, Teradata this week via an alliance with Google outlined how it envisions data analytics and AI will become much more tightly coupled. Teradata is making it possible by integrating Teradata VantageCloud and ClearScape Analytics with the Vertex AI platform that Google makes available as a cloud service.

The goal is to make the large language models (LLMs) that will be built and deployed on the Google Vertex cloud service accessible to the data analysts that are using SQL to surface insights into corporate data.

Of course, one of the primary benefits of generative AI is that it enables end users to make use of natural language to query data. However, there is a rich set of existing applications that makes use of SQL to query data, so the alliance with Google creates an opportunity to extend the reach of SQL queries in LLMs using an application programming interface (API) defined by Google, says Michael Riordan, senior director of product management for data science and analytics at Teradata.

It’s still early days as far as the adoption of LLMs to drive generative AI platforms. ChatGPT may be the best-known example of an LLM but it’s a brute force approach that relies on data scrapped from all over the Internet. As such, the results from a ChatGPT query need to be verified because not all the data being used to drive that response is valid. Longer term, there will soon be a plethora of generative AI platforms based on LLMs that were created using a much narrower range of validated data. Teradata via its alliance is looking to make it simpler for data analysts to invoke those LLMs alongside the data their organization already stores in a more structured format. “Organizations will be able to bring their own model,” says Riordan.

At the same time, organizations will be able to more easily lift and shift data from a Teradata database into an LLM running on the Google Vertex cloud service, he added.

Teradata is clearly making a case for preserving the value of investments made in its database, and Digital CxOs should expect to see other providers of databases and analytics applications moving down the same path.

It’s not likely generative AI platforms will replace legacy analytics platforms as much as it is the level of insights that can be surfaced are about to considerably increase. In the meantime, it’s incumbent on Digital CxOs today to start considering what might be possible tomorrow as the level of insights that come from data dramatically improves in the months and years ahead.