Acceldata supercharges data observability with new AI copilot

As the demand for generative AI assistants continues to grow, vendors across the data ecosystem are moving to implement them front and center into their products. Big names like Snowflake and Informatica were the first to make a move, and now, Acceldata, the known player in the data observability space, has jumped on the bandwagon with its own AI copilot.

Available to use right away, the copilot has been designed to sit across the Acceldata platform and to help with different tasks associated with their data observability workflows, right from monitoring the data pipelines for anomalies to defining policy rules.

“Our innovative approach empowers enterprises to tailor AI-assisted data observability to adapt and conform with their specific operational and business needs, setting us apart in the industry. Built on this AI technology,… we are delivering an AI copilot that eliminates manual configuration hassles, reduces setup time, enables automatic monitoring of data anomalies, and fosters collaboration and contributions from non-technical users,” Rohit Choudhary, CEO and co-founder of Acceldata, said in a statement.

How does AI copilot help with data observability?

According to the demos seen by VentureBeat, the new AI copilot sits across the Acceldata platform as a dedicated button and allows users to automate and accelerate previously manual tasks by entering simple natural language inputs. 

This touches on multiple aspects of data observability but most notably helps with anomaly detection and cost control. Essentially, users can tap the AI to easily study and fix what’s wrong with data freshness, profiling and quality as well as learn consumption patterns to make changes to prevent runaway consumption. 

“We have fine-tuned models to understand enterprise data taxonomy, data anomalies, policy recommendations, predictions and forecasting of costs. In addition, our co-pilot uses RAG with industry-standard LLMs (like GPT) to answer user’s questions on a variety of aspects related to the observability of their data,” Choudhary told VentureBeat. This covers all data used across applications, including gen AI models.

In addition to core observability tasks, the copilot helps Acceldata users with error-prone workflow elements like the generation of data quality rules and policies in bulk. It also enables them to enter natural language commands to generate SQL rules as well as human-readable descriptions for data assets, policies and rules. The latter, Choudhary said, is particularly important as it can facilitate seamless communication between the technical and business stakeholders.

General availability in Q2

While the copilot has just been announced, Acceldata says that some of its largest customers are using it in preview. Choudhary did not share the names of these companies but confirmed that the offering will be generally available in the first few weeks of Q2. 

Acceldata has raised more than $100 million from multiple investors and counts leading banks, telcos, consumer product companies and data providers as its customers.

“The product is proven in the most complex and scalable environments,” Choudhary said while noting that they also plan to introduce capabilities to help enterprises monitor large language model performance. The timeline for this, however, remains unclear at this stage.

Notably, Acceldata’s AI ambitions got a push in September 2023 when it acquired NLP company Bewgle for an unclosed sum. However, it is not alone in the data observability space. Heavily funded players like Cribl, Monte Carlo and BigEye are also targeting the same problem with their respective solutions.

Monte Carlo has even started making its move with generative AI. Back in June 2023, the company debuted two AI features in partnership with OpenAI, one enabling users to create SQL code via natural language and the other suggesting code fixes. Both are now generally available.