When ChatGPT was launched, it revolutionized the way internet users interacted with AI. It offered an AI assistant that could handle a wide range of tasks, from generating natural language content to analyzing complex information. The underlying technology behind ChatGPT, the GPT series of large language models (LLMs), quickly gained attention and became a driving force in both individual and business operations.
Expanding Business Capabilities with LLMs
Enterprises are now using commercial model APIs and open-source offerings to automate repetitive tasks and improve efficiency across various functions. Tasks such as generating ad campaigns or accelerating customer support operations can now be handled with the help of AI. The impact of LLMs in these areas has been profound.
LLMs and the Modern Data Stack
One area where the role of LLMs is often overlooked is the modern data stack. Data plays a crucial role in training high-performance language models. When utilized correctly, LLMs can assist teams in working with their data, whether for experimentation or complex analytics.
Over the past year, as ChatGPT and other similar tools gained popularity, enterprises providing data tooling started incorporating generative AI into their workflows. The goal was to enhance the data-handling experience for customers, saving them time and resources. This integration of LLMs simplified tasks such as data experimentation and running complex analytics.
“Tap the power of language models so the end customers not only get a better experience while handling data but are also able to save time and resources – which would eventually help them focus on other, more pressing tasks.”
Conversational Querying Capabilities
One significant shift with LLMs was the introduction of conversational querying capabilities. This allows users to obtain insights from structured data using natural language prompts, eliminating the need for complex SQL queries.
“The LLM being used converted the text into SQL and then ran the query on the targeted dataset to generate answers.”
Notable vendors such as Databricks, Snowflake, Dremio, Kinetica, and ThoughtSpot have incorporated this capability into their offerings. For example, Snowflake provides two tools: a conversational assistant for querying data and a Document AI tool for extracting information from unstructured datasets.
Startups like DataGPT have also emerged in this domain, specializing in AI-based analytics. Their AI analyst runs thousands of queries to provide companies with conversational insights from their data.
LLMs in Data Management and AI Product Development
Besides generating insights from text inputs, LLMs are also being used in manual data management tasks and efforts to develop robust AI products. Informatica’s Claire GPT, for instance, is a multi-LLM-based conversational AI tool that allows users to interact with and manage their data assets using natural language inputs.
Refuel AI, on the other hand, provides a purpose-built large language model that assists with data labeling and enrichment tasks. Additionally, LLMs have shown promise in removing noise from datasets, a critical step in building reliable AI.
Data integration and orchestration can also benefit from LLMs. These models can generate the necessary code for tasks such as converting data formats, connecting to different data sources, or constructing Airflow DAGs.
Future Applications and Considerations
As LLMs continue to improve and teams innovate, their applications in the enterprise data stack will expand further. This includes areas like data observability, where companies like Monte Carlo and Acceldata are already leveraging LLMs to enhance their offerings.
However, as these language models become more integrated into various processes, it becomes crucial to ensure their performance is accurate and reliable. Any errors can have significant downstream effects, impacting the customer experience.
