Anomalo Inc., an enterprise platform that uses artificial intelligence to monitor and ensure data quality, today unveiled AIDA — short for Anomalo’s Intelligent Data Analyst, a new generative AI system that lets users query and control their data in natural language.

The company uses machine learning to replace traditional rules-based systems for monitoring and correcting data quality issues in enterprise organizations. The platform can automatically detect anomalies and other problems across both structured and unstructured data.

Anomalo Chief Executive Elliot Shmukler told SiliconANGLE in an interview that the new offering extends the company’s monitoring capabilities by providing data scientists, engineers and business users a way to get answers from data. It combines generative AI capabilities with the company’s deep, contextual understanding of enterprise data to let users “talk to their data” in natural language.

“The issue that most folks have when they try to use generative AI for data analysis isn’t producing queries — the models are great at that,” Shmukler said. “It’s actually having the right business and data context to make sure those queries are correct.”

AIDA addresses this by grounding each generated query in Anomalo’s constantly updated knowledge of the organization’s database structure, high-quality data and established business metrics.

Shmukler’s firsthand experience with data errors when working at high-growth companies like LinkedIn Corp. and Instacart inspired him to found Anomalo. At Instacart, missing or inconsistent data feeds could lead to lost revenue — sometimes millions in sales.

According to Shmukler, some of the top analysts spent as much as 50% of their time chasing data quality issues and trying to resolve them.

“We had an issue where Costco, one of our partners, actually sent us a feed without anything from the meat department,” he said. “Turns out, if you put up a Costco storefront with no meats of any kind in it, you lose 50% of the orders.”

Since business data is constantly evolving and the company’s needs and policies are also subject to change, users can transparently view AIDA’s reasoning processes and intervene as needed. Once corrected, those insights into company policy and business context are stored. This improves the system’s understanding over time.

While Anomalo’s biggest footholds are in financial services, retail and e-commerce, Shmukler said data-driven decision-making is becoming universal. The company’s major customers include financial services, such as Discover Financial Services Inc. and Block Inc., formerly Square.

Bringing intelligence to data quality

AIDA was built in collaboration with leading data and AI developers, including Google Cloud, Databricks Inc. and Snowflake Inc. The system uses Anomalo’s existing connectors to more than 17 data sources, including Google BigQuery.

Under the hood, companies can enhance these connections by using AI capabilities provided by Google Cloud’s Vertex AI platform, Snowflake Cortex AI and Databricks Mosaic AI. Customers can also hook in their preferred AI provider, including Google LLC’s Gemini and Anthropic PBC’s Claude.

“It’s a traditional kind of chat interface,” Shmukler explained. “You ask your question, the agent will see if it can answer the question right away… or it might need to query the data, compose a SQL query, show you that query, execute it and then analyze those results.”

Users can also make statements such as, “Hey, this metric is really important. Let’s monitor it with Anomalo from now on.” From there, AIDA will set up a monitor and a trigger based on the specific business metric and then produce warnings when something goes awry.

AIDA is being rolled out as a preview feature to all customers, with feedback expected to guide future updates.

Images: SiliconANGLE/Microsoft Designer, Anomalo