The data observability specialist's new feature uses agentic AI to ensure that the data used to inform analytics and AI tools is high quality and won't lead to poor decisions.
Acceldata on Tuesday launched Adaptive AI Anomaly Detection, a new feature that automatically detects data anomalies before they reach business applications and potentially disrupt operations.
Analytics products, such as reports and dashboards, and AI tools, such as assistants and agents, are built on data. If the data used to inform them isn't high quality, the applications won't be high quality either, and they'll go unused or lead to poorly informed decisions.
Acceldata's Adaptive AI Anomaly Detection uses AI to detect the irregularities that lead to misinformed analytics and AI applications. However, unlike traditional data observability tools that detect one-dimensional errors such as a missing zero in a sales figure, Adaptive AI Anomaly Detection aims to discover hidden anomalies across multiple dimensions, reducing a process that previously took weeks for humans to perform to minutes, the vendor said.
Given its potential to detect complex anomalies and make data engineers and other experts more efficient, Adaptive AI Anomaly Detection is an important new feature for Acceldata users, according to Stephen Catanzano, an analyst at Enterprise Strategy Group, now part of Omdia.
"Adaptive AI Anomaly Detection is a significant new feature for Acceldata users because it moves beyond traditional one-dimensional error detection to uncover complex, multidimensional anomalies automatically," he said. "By drastically reducing manual effort and time needed for analysis, it enhances business-level clarity and enables organizations to catch problems before they impact operations."
Adaptive AI Anomaly Detection is a significant new feature for Acceldata users because it moves beyond traditional one-dimensional error detection to uncover complex, multidimensional anomalies automatically.
Stephen CatanzanoEnterprise Strategy Group
Based in Campbell, Calif., Acceldata is among a small group of data observability specialists whose platforms target data quality. Competitors include Monte Carlo, Metaplane -- recently acquired by Datadog -- and Soda Data.
Anomaly detection
While always important, given data's role as the foundation for analytics and AI applications, anomaly detection has taken on greater significance since OpenAI launched ChatGPT in November 2022.
ChatGPT marked a significant improvement in generative AI technology and led to surging investments in AI development, due to GenAI's potential to make business workers better informed and more efficient. However, because part of making business workers better informed is enabling nontechnical users to work with data, and part of making them more efficient is automating certain tasks and processes, GenAI demands greater emphasis on data quality.
Not only are some GenAI users less skilled than those who historically worked with data -- and were therefore better at recognizing incorrect and misleading GenAI outputs -- but by automating tasks and processes, some of the human oversight that caught mistakes has been removed.
Data observability platforms automate the process of checking data for anomalies, speeding a labor-intensive process that could slow the insight generation that leads to actions and growth. Now, like Monte Carlo, which unveiled agentic AI-powered anomaly detection capabilities earlier this month, Acceldata is helping to progress data observability by deploying agentic AI to detect more complex anomalies at greater speed.
Adaptive AI Anomaly Detection is part of Acceldata's xLake Reasoning Engine. Like the enterprises using agents to automate specific processes, the feature deploys agentic AI to monitor data for anomalies and then detect, prioritize and resolve them.
Key features include multidimensional detection, intelligent sampling that prioritizes high-risk anomalies and autonomous pattern recognition that detects unique patterns traditional rule-based anomaly detection tools cannot.
Because Adaptive AI Anomaly Detection targets data quality, it focuses on an important user need, according to Kevin Petrie, an analyst at BARC U.S.
BARC research shows that data quality is, to date, the most popular use case for agentic AI-powered data management, with about half of respondents to BARC surveys already using or considering using agents to monitor and improve data quality.
"Acceldata addresses this demand," Petrie said.
Other benefits include the following, according to Acceldata:
Cost savings related to detecting and resolving data quality issues.
Compliance breach alerts based on detecting unusual data access patterns.
Business impact forecasting to connect data problems with potential decision-making errors based on anomalous data.
"These agentic capabilities ... cut a broader swath than many other data quality tools," Petrie said. "What distinguishes this announcement is the breadth of use cases it addresses -- data quality, pipeline performance, FinOps and data governance. These are all high-priority items for analytics and AI adopters."
Perhaps the most significant potential benefits are business impact forecasting and compliance breach alerts, according to Catanzano.
"Business impact forecasting is critical because it links upstream data issues with downstream decision-making, preventing costly mistakes," he said. "Compliance breach alerts are equally important, helping organizations proactively catch and address security risks by correlating user behavior with data sensitivity."
Looking ahead
With Adaptive AI Anomaly Detection now available, Acceldata's next product launch will be the general availability of its Agentic Data Management platform in May, according to Rohit Choudhary, the vendor's co-founder and CEO.
Unveiled in private beta testing in February, the platform includes a series of targeted GenAI agents aimed at data quality, data governance and pipeline management.
"We will be building and releasing agents to cover these aspects and more," Choudhary said.
Using data collected from customer behavior across various deployment types, Acceldata has the context needed to develop agents and address users' needs, he continued.
Acceldata's focus on providing customers with more agentic AI capabilities is wise, according to Catanzano. Agents that enable greater autonomous decision-making, real-time optimization and predictive data governance are possibilities.
"Deepening integration with operational AI systems and refining personalization for specific industries could also help Acceldata continue leading in the shift toward fully autonomous data management," Catanzano said.
Eric Avidon is a senior news writer for Informa TechTarget and a journalist with more than 25 years of experience. He covers analytics and data management.
