About the company
As AI reshapes how software is built, a top-10 US bank is making a defining bet: technology excellence is its primary growth engine. The company is transforming from a bank that builds software into a technology company that delivers financial services, with engineering velocity at the center of that strategy.
To get there, leadership needed hard data on where capacity was being consumed, which tools were accelerating delivery, and where bottlenecks were limiting the organization's ability to move at the speed AI now makes possible. The company launched a developer analytics initiative to build the measurement foundation required to answer those questions at scale. "We wanted to empower engineering teams to become more efficient. That meant giving them metrics and insights into performance, bottlenecks, and their ability to serve customers," says the CTO of a major revenue division.
"We wanted to empower engineering teams to become more efficient. That meant giving them metrics and insights into performance, bottlenecks, and their ability to serve customers." — CTO
Challenges
With the arrival of a new CTO, along with several leaders from tech-forward companies, the company set out to unlock a 20% velocity increase within one year and accelerate further with AI across a global engineering organization of 9,000 developers. This raised big questions: Do we have the right talent? Are we building things the right way? What are our actual bottlenecks? What is the true developer experience? Before these questions could be answered, leadership had to confront several critical gaps:
| Challenge |
Business Impact |
| No visibility, no leverage |
Leadership lacked reliable, real-time data on basic engineering performance. Pulling DORA and PR-based metrics by hand was the only option, unsustainable at scale and too slow to act on. Without a clear picture of where capacity was going and what was driving inefficiency, there was no way to make confident decisions about where to invest, where to cut, or how to improve. |
| Teams unable to drive their own improvement |
There was no self-service access to performance data. Teams couldn't see their own bottlenecks, which meant problems went unresolved and improvement stalled. The existing enterprise metrics initiative was neither extensible nor scalable, leaving individual teams without the tools to understand or change how they worked.
|
| AI adoption outpacing the ability to measure it |
As the organization began adopting AI coding tools, leadership had no infrastructure to measure whether those investments were paying off. Without the ability to correlate AI tool usage with delivery outcomes, there was no basis for making smarter tooling decisions or demonstrating the business value of the AI transformation underway. |
Key challenges and their impact on scaling engineering and AI adoption
Why Faros
The company evaluated several tools on the market, including well-known qualitative and survey-based platforms, but none met their needs for quantitative rigor, GenAI impact measurement, AI insights, and self-service. When Faros was introduced, it quickly stood out as the clear choice due to several key advantages:
Quantitative, objective truth as a non-negotiable. While qualitative survey data is valuable, leadership felt it was simply not enough to paint the full picture of engineering operations. The company was looking for a platform with robust quantitative capabilities, and Faros’s quantitative-first approach was unmatched by competitors.
“We run on an Amazon operating philosophy: trust in God, but others bring data. Quantitative rigor was non-negotiable. Faros had that nailed.”
Extensible capabilities and connectors, built for enterprise realities. Because the organization uses heavily customized logic for CI/CD definitions and PR flows, they required a system that wouldn’t force them to change the way they work. Faros stood out as the only viable choice due to its flexibility, composability, and extensibility. In addition to SaaS tools and AI coding agents, it can ingest data from non-standard systems, support custom connectors, and enable querying in ways that align with how and where the company wants to access its data.
“The API capabilities, well-defined data model, ‘headless’ support, and library of connectors to developer systems were all incredibly strong. No other vendor had those things figured out together.”
Enterprise-grade security and scalability from day one. As a global financial services enterprise, the company needs a platform that can securely manage thousands of developer identities. Most tools cannot support this volume, which often forces companies to switch vendors as they scale. Faros is built to handle enterprise capacity, allowing the company to get started with an initial 1,000 engineer pilot, prove ROI, and then confidently scale to their full engineering base.
A team that executes like a partner. The CTO had been introduced to Faros from a trusted colleague who had first-hand experience with the platform. From the outset, Faros worked closely with stakeholders across the organization to provide a highly collaborative and responsive experience. The team moved quickly and followed through, turning a complex evaluation into a successful company-wide deployment.
“Every engineer at our company uses this solution now. Multiple leaders wanted direct access during evaluation, and Faros delivered, keeping everyone informed and actually shipping something that works at scale. That execution under pressure is something I really respect.”
How this F100 financial services institution uses Faros to run its AI-forward engineering organization
Deploying a measurement infrastructure across an enterprise with heavily customized internal tooling requires a deliberate, phased approach. The organization executed a nine-month pilot and integration phase.
The first three months were dedicated to getting initial core metrics up and running. The subsequent six months expanded the pilot to include DORA metrics like deployment frequency and a custom definition of lead time (PR merge to deploy). Because every team utilized different deployment models (batch, daily, and CI/CD), rolling out these metrics required careful buy-in and organizational selling.
The limited pilot involved 1,000 developers. Once teams were given dashboard access, leadership established imperatives, such as an organizational goal to achieve 20% higher throughput, rather than top-down quotas. Leaders could review their standings, identify underperforming teams, and use the data for truth-seeking rather than punitive measures. Following the pilot's success, the platform is now being rolled out to the full 9,000 global engineering base.
Each VP oversees teams working on thousands of applications, each with its own pipeline and deployment stages. Faros provides a detailed view into the end-to-end delivery process for each one, so every team can see exactly which PRs are in a deployment, where work is waiting (whether in build, manual approvals, or specific deployment environments), and where test failure rates are elevated. Leaders have purpose-built views focused on the performance of the most critical, high-volume applications. This level of granularity and historical depth is simply not available in the individual tools.
“The pilot results were strong. With 1,000 developers, we demonstrated 20%+ increase in throughput and 15% reduction in cycle time. On pure developer cost alone, that ROI more than pays for Faros.”
With AI coding tools like GitHub Copilot and Windsurf now in use across the engineering base, Faros gives the organization the infrastructure to measure their actual impact, correlating AI tool adoption with delivery metrics to understand where AI is accelerating throughput and where new bottlenecks are emerging. Faros's AI insights provide natural language synthesis, explaining exactly why and where bottlenecks occur.
“I particularly like the AI insights, which can tell you in natural language where your bottlenecks are: 'Your merge-to-deploy time is elevated because your test failure rate is higher,' or 'You have a high PR churn rate.'
Benefits realized with Faros
| Capability |
Benefit |
| Decisive ROI on developer efficiency |
Tracking 1,000 developers in a pilot, the organization recorded a 20%+ increase in PR throughput and a 15% decrease in cycle time. This data provided the financial and operational confidence required to expand the Faros solution to the full 9,000 engineering base and gain efficiencies globally.
|
| Unified data for executive-to-pod visibility |
With Faros, the same underlying data powers team-level dashboards, director-level operational reviews, and CTO-level views across the full engineering organization. The organization built a unified measurement framework capable of generating derived metrics at every scale so every layer now operates from the same source of truth. |
| Accelerated AI transformation with diagnostic clarity |
With AI coding tools already in use across the full engineering base, Faros provides the infrastructure to track AI adoption, correlate AI usage with delivery metrics, and surface next-layer bottlenecks in natural language as AI changes where time is actually spent in the development cycle. |
| Immediate utility across fragmented systems |
Faros’s extensive connector library meant that out-of-the-box integrations handled the heavy lifting. This bypasses the massive switching costs and manual integration hurdles typical of evaluating developer analytics at enterprise scale. |
| A truth-seeking organizational culture
|
Rather than setting hard goals that lead to gamed metrics, the organization uses Faros to foster a truth-seeking environment. By giving teams their own performance dashboards, leaders replaced punitive top-down pressure with an empowering, self-driven operational cadence.
|
Benefits realized with the Faros partnership
“Visibility into how teams operate and our bottlenecks is an ongoing need. AI changes what you're measuring, not whether you need to measure.”
The system for running engineering with AI
Faros is the system for running engineering with AI. We give engineering leaders visibility into how work operates across code, people, and systems, and control over how that work progresses through enforceable workflows and policy. This enables organizations to deploy AI effectively and improve engineering throughput with stronger cost efficiency. Request a demo today.
.png)
.webp)
.webp)