DataGPT: An AI Analyst that Enables Any Company to Directly Engage with Data

One startup company, DataGPT, dedicated to simplifying how businesses gain insights from data, has officially launched its new AI analyst, a conversational chatbot that helps teams understand the content and reasons behind their datasets through natural language communication.

Starting today, this AI tool combines the creativity and understanding capabilities of self-deployed large language models with DataGPT's proprietary analysis engine's logic and reasoning abilities to perform millions of queries and calculations to determine the most relevant and impactful insights. This includes everything from how something affects business revenue to why something happened.

"We are committed to enabling anyone, any company to directly converse with their data," said Arina Curtis, CEO and co-founder of DataGPT, in a statement. "Our DataGPT software, based on conversational AI data analysis, not only provides instant, analyst-level results but also offers a seamless, user-friendly experience that bridges the gap between rigid reports and informed decision-making."

However, it is interesting to see how DataGPT stands out in the market. In the past year, many participants in the data ecosystem, including data platform providers and business intelligence (BI) companies, have made their moves in generative AI to make it easier for users to consume insights. Most data storage, connection, warehouse/lake, and processing/analysis companies are now turning to allow customers to have conversational AI interactions with their data.

DataGPT AI Analyst: How Does It Work?

DataGPT was founded over two years ago, targeting the static nature of traditional BI tools where one person had to manually dive into dashboards to get answers to ever-changing business questions.

"Our first customer, Mino Games, invested heavily in building an ETL process, creating many custom dashboards, and hiring an analytics team," Curtis told VentureBeat. "Despite exploring all available analytics solutions, they still struggled to get answers to critical business questions quickly and clearly. DataGPT enables them—and all customers—to efficiently and effectively gain deep data insights."

At its core, the solution only requires a company to set up a use case—a DataGPT page configured for a specific business domain or a predefined set of KPIs. Once the page is ready, end users get two elements: the AI analyst and the data navigator.

The former is a chatbot experience where they can input questions in natural language and get instant insights, while the latter is a more traditional version where they get visualizations displaying key metric performance and can manually dive into any combination of factors.

For the conversational experience, Curtis said there are three main layers on the backend: data storage, core analysis engine, and the analyst agent driven by a self-deployed large language model.

When a customer poses a business question to the chatbot, such as "Why did revenue increase in North America?", the embedded model in the core analysis engine finds the closest match in the data storage architecture ("Why did <monthly recurring revenue> increase in <country>['USA', 'Canada', 'Mexico']?"), and the self-deployed LLM accepts the question and creates a task plan.

Each task in the plan is executed by the data API algorithms of the analysis engine, performing comprehensive analysis on extensive datasets with capabilities beyond traditional SQL/Python functions. The results of the analysis are delivered in a conversational format to the user.

"The core analysis engine does all the analysis: calculating impacts, performing statistical tests, computing confidence intervals, etc. It runs thousands of queries in the blink of an eye in the lightning cache (data storage) and gets results. Meanwhile, the self-deployed LLM makes the response human-friendly and sends it back to the chatbot interface," explained Curtis.

"Our lightweight yet powerful LLM is cost-effective, meaning we don't need expensive GPU clusters to achieve fast response times. This flexibility gives us a competitive edge. We invested time and resources in creating an extensive internal training set tailored specifically to our models. This ensures unparalleled accuracy and robustness to any architectural changes," she added.

Benefits for Businesses

While Curtis did not disclose how many companies are working with DataGPT, the company's website indicates that multiple enterprises are leveraging this technology for benefits, including Mino, Plex, Product Hunt, Dimensionals, and Wombo.

These companies are able to accelerate their time to insights using the chatbot and ultimately make critical business decisions faster. It also saves analysts time for more urgent tasks.

The CEO pointed out that DataGPT's lightning cache database is 90 times faster than traditional databases. It can run queries up to 600 times faster than standard BI tools while reducing analysis costs by 15 times.

"These newly gained insights can unlock up to 15% revenue growth for businesses and save nearly 500 hours per quarter for busy data teams, allowing them to focus on higher-yielding projects. DataGPT plans to open-source its database in the near future," she added.

Future Plans

So far, DataGPT has raised $10 million in seed and pre-Series A funding and built a product that covers 80% of data-related questions, including key metric analysis, key driver factor analysis, segment impact analysis, and trend analysis. In the future, the company plans to expand its analytical capabilities further in these areas to cover as many domains as possible. This will include things like group analysis, forecasting, and predictive analysis.

However, the CEO did not disclose specific timelines for these capabilities. Nevertheless, the expansion of analytical capabilities could give DataGPT an advantage in a market where every data ecosystem provider is introducing or seeking to introduce generative AI.

In recent months, we have seen many companies, such as Databricks, Dremio, Kinetica, ThoughtSpot, Stardog, Snowflake, investing in LLM-based tools—either through internal models or integration—to improve data access. Almost every vendor conveys the same message of ensuring all enterprise users, regardless of technical expertise, can access and derive value from data.

DataGPT claims its strength lies in the power of its analysis engine.

"Popular solutions fall into two categories: LLMs with simple data interfaces (e.g., LLM+Databricks) or BI solutions with integrated generative AI. The first category deals with limited data volume and source integration. They also lack deep analysis of data and awareness of the data business context. Meanwhile, the second category leverages generative AI to moderately accelerate traditional BI workflows, creating similar types of narrow reports and dashboard outputs. DataGPT offers a new data experience... LLM is the right brain. It excels at contextual understanding. But you also need the left brain—the data API—our algorithms for logic and conclusions. Many platforms struggle when combining deep data analysis and interpretive, logical 'left brain' tasks with LLMs," said Curtis in a statement sent to VentureBeat.