Summary

• Inflection was replatforming to serverless and needed an operations model that minimized required engineering resources
• Inflection deployed Sedai’s Autonomous Optimization & Remediation capabilities to manage Lambda services
• Sedai helped Inflection deliver improved customer experience with a >95% FCI (Failed Customer Interaction) reduction and a 7% NPS gain
• Sedai also supported a 90% reduction in operational toil so that engineers could spend more time on core work that created competitive advantage

Problem

Inflection had decided to replatform from Kubernetes to serverless for both development and operational efficiencies.  This was a complex task given that Inflection served a large number of customers with APIs built to process massive transaction volumes and deliver actionable results in real-time. Inflection sourced and aggregated their data from more than 2,000 in-house sources, 3,100 county, and 94 federal courts and used advanced algorithms to classify and filter more than 230M aggregated criminal records.

Inflection wanted to free up engineer time to focus on core functionality and alongside the move to serverless was looking for a lower effort way to manage operations.  Inflection CTO Siddharth Ram defined operations as “the work done before and after code is deployed into production to maintain or improve operational quality.”

Solution: Autonomous Management

Inflection decided to implement Sedai´s autonomous cloud management platform in a way that would allow them to heavily prioritize latency improvements over cost reductions. Although the setup and integration of Sedai’s autonomous cloud management platform typically takes fifteen minutes for Lambda, the team from Sedai worked closely with Inflection´s SRE team during the implementation.  

Immediately after the setup was completed, Sedai’s autonomous system started accessing Inflection´s cloud accounts with the goal of understanding their infrastructure and detecting their production environment topology. Once the initial discovery process was completed, Inflection´s SRE team was able to see all identified resources including all serverless functions and applications.  Sedai then began collecting metrics data and developing optimization & remediation recommendations.

‍

Results

Improved Customer Experience with >95% FCI Reduction

Inflection saw major improvements in customer experience through the combination of replatforming and autonomous management.  In particular, Inflection reduced Failed Customer Interactions (FCIs) - Inflection was able to reduce FCIs from 3.2% to 0.02% (a greater than 95% reduction) over the space of 3 quarters, and FCIs approached near zero over the following year.

The FCI metric represents a subset of errors and slow services that impact customer experience; Amazon advises that companies should target an FCI rate of less than <0.025%, which Inflection achieved.  Furthermore, latency was reduced by 14%, improving service for Inflection customers, with individual services showing up to 81% latency reductions.

These improvements contributed to NPS (Net Promoter Score) gains.  NPS is calculated directly from customer ratings  of their likelihood to recommend Inflection to peers & colleagues. Inflection’s NPS improved from 63% to 68% in the first two quarters after the change, and then to 70% over the next year.

Improved Operational Productivity: ‘busywork’ reduced 90%

With Sedai´s autonomous cloud management platform up & running, Inflection´s SRE team was able to move away from human-led automated scripted management of cloud-native workloads towards a machine learning system for autonomous management. Being able to autonomously resolve issues significantly reduced the toil for their SRE team while optimizing for both cost and performance.

The operational productivity gains supported the team’s wider goal of spending time on ‘core’ work and minimizing time on ‘context’ activities.  Siddharth’s view is that core work is the secret sauce of the company - “anything that is a competitive advantage or the key value add in the company”.  In contrast ‘context’ work is necessary, but not necessarily done by your engineering team.

Release Quality Insights

Sedai also provided insights to the team on the underlying code quality as Inflection developers released new code.

Looking Ahead

Sedai helped Inflection deliver improved customer experience while also enabling Inflection’s engineering team spend more time on the “core” work that advanced the company’s competitive advantage.  Based on his experience, Siddharth now expects growing adoption of autonomous management for modern applications.

‍