Podcast: A conversation on AI Anomaly Detection for Cloud Cost Optimization with Frank and Stephen "the FinOps guys".

I have the privilege of witnessing firsthand how automated machine learning is transforming industries. One of the most impactful applications is time series anomaly detection, particularly in cloud cost optimization.

In a recent conversation with The FinOps Guys Stephen Old and Frank Contrepois for their Podcast series 'What's new in Cloud FinOps?', I had the opportunity to discuss this topic in depth—how automation in feature engineering is revolutionizing anomaly detection, why real-time insights matter, and how businesses can dramatically reduce the time between event occurrence and resolution.

Key takeaways from our conversation include:

• The importance of automated feature engineering in anomaly detection,
• How businesses can flag anomalies in cloud costs before they become costly problems,
• The role of root cause analytics at scale in making anomalies actionable, and
• The need for speed in resolution to maximize cost efficiency and system performance.

‍

The Cost of Unidentified Anomalies in Cloud Spending

Cloud environments are inherently dynamic, and cost structures are complex. Spikes in spending can be triggered by misconfigurations, unexpected workload surges, inefficient scaling policies, or even billing errors. Without a robust anomaly detection system, these issues can go unnoticed until they have already significantly impacted the bottom line.

Traditional monitoring tools often rely on basic thresholds or predefined rules, which can lead to missed anomalies or false alerts. The real challenge is detecting irregularities dynamically, identifying their root cause, and responding in real time.

‍
That’s where Tangent Works’ technology comes in.
‍

How Tangent Works Automates Anomaly Detection

‍
At Tangent Works, we specialize in automated feature engineering and predictive analytics for time series data. Our technology enables businesses to rapidly build, deploy, and scale machine learning models without requiring extensive manual intervention.

Why Automated Feature Engineering is a Game Changer

‍
Feature engineering—the process of selecting and transforming the right data inputs—is traditionally one of the biggest bottlenecks in machine learning. Without automation, businesses spend weeks or even months manually fine-tuning models to detect anomalies accurately.

With Tangent Works’ technology, feature engineering happens automatically, allowing organizations to:

• ‍Detect cost anomalies in real-time,
• ‍Improve accuracy by eliminating noise,
• ‍Scale anomaly detection across multiple cost centers and services.

By reducing the complexity of setting up machine learning models, Tangent Works democratizes advanced analytics, enabling both data scientists and business users to extract value from time series data instantly.

From Anomaly Detection to Root Cause Analytics‍

‍
Detecting anomalies is just the first step. The real business value comes from understanding why they happened and how to prevent them from recurring.

A major challenge in cloud cost management is the sheer volume of events. If an organization receives hundreds of anomaly alerts per day but lacks context or prioritization, actionable insights get lost in the noise.

‍
With Tangent Works, businesses can go beyond simple anomaly flagging by enabling root cause analytics at scale:

• ‍Pinpoint which cloud resources are causing cost spikes,
• ‍Correlate anomalies with system logs, workload trends, and user behaviors,
• ‍Automatically rank anomalies based on severity and impact.

This makes it easier for cloud operations teams to focus on what matters most—reducing waste and ensuring cost efficiency.

‍
Speed is Key: Reducing Time-to-Resolution

‍
The faster an anomaly is detected and resolved, the less financial and operational damage it can cause. The key to effective cloud cost optimization is reducing the time between anomaly occurrence and resolution.

With Tangent Works’ real-time anomaly detection and automated insights, businesses can:

• ‍Detect and flag anomalies as they happen,
• ‍Trigger automated workflows to address issues before costs escalate,
• Continuously refine models based on real-world feedback.

By integrating automated machine learning into cloud cost monitoring, companies can transition from a reactive approach (where anomalies are investigated after the fact) to a proactive strategy that prevents unnecessary expenses in the first place.
‍

Final Thoughts

‍
Anomaly detection is not just about identifying irregularities—it’s about making those anomalies actionable. With Tangent Works’ technology, businesses can automate the detection, diagnosis, and resolution of cloud cost anomalies, ensuring that resources are used efficiently and cost-effectively.

‍
If you’re interested in learning more about how automated machine learning can optimize cloud cost management as well as other interesting innovations in FinOps and GreenOps, be sure to check out the Podcast 'What's new in Cloud FinOps?' by The FinOps Guys Stephen Old and Frank Contrepois who are doing a great job in strengthening and educating the growing FinOps community!

‍