Bigeye, an enterprise AI trust platform, has launched Cost Anomaly Detection within its Agent Trust Hub. The feature automatically flags unusual increases or decreases in AI spending by agent, user, or workspace based on daily evaluations against historical patterns. The tool addresses a growing concern among enterprises deploying AI agents. According to an EY survey from July 2026, 82% of senior leaders investing in AI are worried about token usage and related costs. Cost Anomaly Detection runs automatically without configuration, catching two failure modes: agents entering costly retry loops and agents that stop working due to technical failures, appearing as false savings. The feature links directly to conversations driving cost changes. Building on Bigeye's data observability expertise, the system provides dataset-level cost visibility across cloud warehouses and legacy systems. It began rolling out to existing customers on 18 August 2026. Bigeye serves major enterprises including USAA, Zoom, Hertz, and Cisco, monitoring 3 million datasets daily.

SAN FRANCISCO, March 25, 2024 /PRNewswire/ -- Today Bigeye announces the availability of Bigeye Dependency Driven Monitoring. With this launch, Bigeye becomes the first data observability platform to allow enterprise data teams to connect their analytics dashboards, map every dependency across modern and legacy data sources, and deploy targeted data observability to ensure they stay reliable by default.According to Bigeye's 2023 State of Data Quality Report, 70% of business leaders don't trust analytics dashboards to make critical decisions due to regular data quality incidents."Bigeye's new capabilities will let us surface the health of critical Power BI dashboards in real-time." Post this Announcing Bigeye Dependency Driven Monitoring. Optimized data observability for every column powering critical analytics dashboards and data products.Regarding this challenge Kyle Kirwan, CEO and Co-founder of Bigeye, said, "We've spoken with hundreds of enterprise data leaders and, despite investing heavily in data quality tools and processes, they still struggle to deliver reliable data analytics to business users. Something the data observability industry hasn't yet solved is how to handle the complexity and size of large enterprise data pipelines. This is because enterprise dashboards have a long list of dependencies that span modern and legacy technologies and data observability platforms have yet to offer true support for the types of hybrid environments nearly all Fortune 500 companies have."Bigeye Dependency Driven Monitoring uniquely solves this challenge by combining enterprise-grade lineage technology with data observability to automatically trace the entire enterprise data pipeline at column-level precision through traditional and modern technologies, ETL stages, and even across the boundary from cloud to on-premises environments.Bigeye Dependency Driven Monitoring improves data observability ROIWithout a clear understanding of which columns are critical dependencies for business operations and which are not, data engineering teams are left with no choice but to deploy broad, blanket monitoring across a large number of tables and columns to reach adequate coverage. While this approach helps engineers identify anomalies anywhere they occur in the data environment, it also requires them to manage higher compute costs, unwanted alert noise, and additional overhead as a result of deploying unnecessary monitoring.Bigeye Dependency Driven Monitoring turns this paradigm on its head by allowing data analysts and business users to start from their critical dashboard and—in just a few clicks—enable data observability on each column that matters, and none that don't.When compared with other data observability approaches, Bigeye Dependency Driven Monitoring provides:Faster time to value and improved trust for data consumers.Clear visibility into the health of the entire analytics data pipeline for analysts.Reduced alert noise and faster issue resolution for data engineers.Lower total cost of ownership and less compute overhead for data leaders.When a data issue is detected, Bigeye will instantly notify each data source owner impacted through Slack or Microsoft Teams