WebOps Takes Control of its AWS Spend with Next-Generation Managed Services and Agentic FinOps

The Challenge

Understanding what was driving a growing cloud bill

WebOps is a software company. It carries no inventory and operates no warehouses, so its spending concentrates in two places: the people who build the product and the tools they build it with. AWS sits at the top of that second list. The difficulty was less the size of the bill than the visibility behind it. Consumption-based pricing meant costs could climb without a clear account of which teams, environments, or services were responsible.

Closing that gap fell to the engineering team, the same people responsible for building and supporting the product. Behind the day-to-day cost question sat a larger structural one. WebOps runs Microsoft SQL Server Enterprise Edition on Amazon RDS, and because that licensing is driven by CPU, the cost scales with the volume of data the platform manages. That made SQL Server licensing one of the largest single line items in the company’s AWS spend.

The Solution

Cost clarity first, then modernization

WebOps set out to find an AWS partner that would be responsive and proactive in equal measure. The engagement with EPI-USE Services for AWS began with AWS Managed Services and Cloud Cost Optimization: mapping where the money was actually going, then separating necessary spend from avoidable spend. Environments were right-sized and consolidated, subscribed services that were no longer needed were eliminated, and cost allocation tagging was introduced so spend could be attributed cleanly across development and production workloads.

What makes the monthly cadence different is the work that happens before the meeting. EPI-USE Services for AWS runs that analysis through its FinOps Hub, where AI agents built on Amazon Bedrock query the client’s own billing and usage data, establish what moved and why, and draft the findings in plain business language. Nothing reaches the client unreviewed. An EPI-USE FinOps specialist validates every finding and approves every recommendation before it is issued. The agents compress the analysis. The accountability stays with people.

Key Solution Elements

• Amazon Bedrock, EPI-USE FinOps agent AI agents analyze billing and usage data continuously and draft findings in business language, with specialist approval on every recommendation
• Cloud Cost Optimization Environment right-sizing and consolidation, elimination of unused services, and cost allocation tagging across development and production
• AWS Managed Services Day-to-day operations, monthly service reviews delivered as an actionable punch list, and early warning on AWS changes that could affect cost
• Modernization Viability Assessment Assessed whether Microsoft SQL Server Enterprise Edition on Amazon RDS could move to AWS native services, and recommended Babelfish for Aurora PostgreSQL
• Phase 1 proof of concept Verified the first WebOps database running on Amazon Aurora PostgreSQL, with AWS modernization program funding covering much of the work
• Estate under management Amazon EC2, Amazon ECS, Amazon VPC, Amazon ElastiCache, Elastic Load Balancing, Amazon S3, Amazon CloudWatch, and AWS Database Migration Service

Outcomes & Benefits

Confidence in the number, and a path to lower it

As a growing company running a high-volume, data-intensive platform, WebOps still invests significantly in AWS. What changed is the quality of the reasoning behind that investment. With the modernization path proven, WebOps took the production migration in-house and is building it with its own team, with EPI-USE Services for AWS continuing in an advisory role.

Cost Confidence

• CEO confidence that AWS spend is necessary spend, up from roughly 10 percent to around 80 percent

• Significant reduction in unnecessary AWS spend through right-sizing and environment consolidation
• Savings plans and available discounts applied with planning rather than discovered after the fact

• Validated path off Microsoft SQL Server licensing costs, with Phase 1 verified on Amazon Aurora PostgreSQL
• Engineering time returned from cloud cost analysis to product development

Engineering Productivity

• Monthly reviews that produce decisions rather than investigations