Amazon Web Services Inc. today introduced an artificial intelligence tool designed to help organizations lower their cloud bills.

AWS FinOps Agent is initially available in public preview. It’s the latest in a series of AI tools that the Amazon.com Inc. unit has released over the past year to ease cloud administrators’ work. 

AWS provides several services that can help customers track their cloud spending and find optimization opportunities. FinOps Agent makes the data collected by those services available through a ChatGPT-like interface. Developers can describe the information they wish to access and have the AI display it in a tabular format.

One of the tasks that FinOps Agent eases is detecting anomalous expenses. For example, an engineer could ask it to check if there are any applications that used an unusually large amount of infrastructure in the past month. Users can optionally have the tool filter small cost fluctuations that likely don’t indicate the presence of a major issue.

When FinOps Agent finds an unusual expense, it displays the cloud service behind the cost spike and the user activity that caused it. For example, the tool could trace an increase in Amazon Relational Database Service spending to a workload that ran more queries than usual. FinOps Agent can also display the business unit responsible for the affected workload to help developers coordinate troubleshooting efforts.

The tool generates spending reduction suggestions based on data from AWS Cost Optimization Hub. The latter service uses AI models to identify more than a dozen types of cost cutting opportunities. For example, it can spot workloads that underutilize the instance in which they run and should be moved to a smaller machine.

Companies often attach pieces of data called tags to the resources that make up an AWS environment. For example, instances that power an AI training workflow might include a “TensorFlow” tag. Developers can increase the quality of FinOps Agent’s output by providing it with explanations of their company’s AWS tags and cloud management practices.

AWS says the tool can not only answer ad hoc questions but also generate periodic spending reports. Furthermore, FinOps Agent is capable of proactively notifying administrators when it finds an anomalous expense. It can display its findings in a standalone chatbot interface, Slack and the popular Jira bug tracking tool. 

FinOps Agent’s debut comes about three months after AWS launched two other AI-powered administrator tools into general availability. The first, AWS Security Agent, can detect vulnerabilities in workloads deployed on the Amazon unit’s cloud and external infrastructure. The tool rolled out alongside the AWS DevOps Agent, which helps administrators fix technical issues such as application outages.

Photo: AWS