Caylent Services
Agentic Engineering and Enablement
Make AI the way your organization builds, ships, and operates software.
Smarsh Works with Caylent to Reimagine Compliance Software Delivery, Building an Agentic-SDLC framework on Amazon Bedrock
Generative AI
At a glance
Smarsh partnered with Caylent, AWS, and the AWS Generative AI Innovation Center to transform its software development lifecycle (SDLC) with generative AI. Together, we built an agentic SDLC framework on Amazon Bedrock, embedding Claude Code across the entire delivery lifecycle for code generation, automated testing, documentation synthesis, and security review. The result is faster software delivery, improved engineering efficiency, and stronger collaboration between product and engineering teams.
$5–10
Million saved in productivity
15–40%
Lower defect escape rates
15–30%
Improvement in test coverage
Projected 25%
Reduction in lead times for changes
Smarsh, a leader in financial communications compliance, wanted to transform its software development lifecycle (SDLC) with generative AI. Working closely with Amazon Web Services (AWS) experts Caylent and the AWS Generative AI Innovation Center, Smarsh built an agentic SDLC framework that increases delivery speed, improves test coverage and reliability, and deepens collaboration between engineering and product teams. And it has done so while maintaining the robust controls required by large financial institutions.
In year one, AI investment freed $5-10 million in engineering and product capacity, which was reinvested directly into innovation work. Defect escape rates were reduced by 15-40 percent; lead times for changes are now falling and projected to reduce by 25 percent; and test coverage has increased by 15-30 percent.
Business Challenge
Effectively Leveraging Generative AI to Ensure Compliance
Smarsh, operates in a tightly regulated environment. With nearly 500 engineers across multiple business units, the company wanted to use generative AI to transform its SDLC without compromising security, auditability, and governance. Smarsh needed a way to ensure rigor and explainability across its architecture, processes, and tooling.
"What excites me most about AWS and Amazon Bedrock is the speed of innovation. With AWS and Caylent as partners, it’s like walking into a room where all the doors are closed and having them open those doors for us. It’s a breath of fresh air.”
Blake Sherwood
Distinguished Technologist – Platform & AI
Solution
An Agentic SDLC Framework with Strong Governance on AWS
Smarsh chose to work with Caylent, a 2025 AWS partner of the year in three categories, as well as the AWS Generative AI Innovation Center, to accelerate and derisk their AI adoption. Together, they built the agentic SDLC framework on Amazon Bedrock and, using Claude Code, they have embedded code generation, automated testing, documentation synthesis, and security review directly across the entire delivery lifecycle. This enables them to enforce the auditability, data residency, and policy controls required by the world’s largest global financial institutions.
Adopting AI by Workflow, Not by Individual Seat
Smarsh adopted a “workflow, not seat” philosophy, so they could weave AI into end-to-end workflows and map real business outcomes. For example, product managers can now use AI to help generate and refine requirements, clarify problem statements, and align outcomes with outputs, while engineers use AI to support specification development, code changes, test generation, and long-running autonomous tasks.
Smarsh uses AI capabilities throughout the SDLC: for requirements, design, coding, testing, validation, and release.
“We’ve built telemetry, traceability, and capacity into the platform so we can explain every step of how workflows through the system. Combined with our control framework, it lets us bring the same level of trust and governance to AI that our customers expect from our products,” says Blake Sherwood, distinguished technologist at Smarsh.
This workflow-first approach sets the stage for a scalable, governed agentic SDLC framework that can be rolled out across teams and functions.
Outcome
Faster Delivery, Better Coverage, Stronger Collaboration
Although Smarsh sees AI adoption as a long-term journey, the organization is already recording significant benefits across the agentic SDLC. AI-enabled workflows have compressed timelines for major engineering tasks. Customers will start to see delivery times reducing, with projects estimated at six to nine months soon being delivered in just three, or even within a single sprint, and teams have increased test coverage from about 20 percent to almost 100 percent in a single day. Previously, this would have taken weeks.
Improvements are especially visible in test harness creation, which traditionally was an unpopular task that now generates genuine enthusiasm among engineers. As a result, Smarsh is seeing better alignment between business requirements and technical implementation, while vulnerability remediation and defect reduction are supported by AI-assisted workflows that address issues right at the developer’s keyboard.
Smarsh had recognized the potential for cultural challenge, but early skepticism has given way to growing enthusiasm as proof points accumulate and detractors become champions. The AI enablement team runs office hours, demos, and communities of practice to share success stories that drive organic, peer-driven adoption. Collaboration between product and engineering teams has increased, with both sides gaining a deeper understanding of each other’s work and technology choices, increasing the shared ownership of outcomes.
The pace of AWS innovation with Amazon Bedrock, and its partnerships with frontier model providers, enable Smarsh to continually explore new ways to enhance its development lifecycle.
“Having Caylent, AWS, and the Generative AI Innovation Center as resources has been incredibly important for Smarsh and for our engineers. They’re a sounding board with battle-tested experience, and that support is not just helpful, it’s essential,” says Sherwood. “Simply being able to sit down with their subject matter experts, talking through problem statements and what’s coming next, has really shaped our thinking. That engagement helped us validate our theses and gave us the confidence to take this step into AI.”
Over the next 12–18 months, Smarsh aims to integrate AI orchestration into every part of how it produces value for customers, turning generative AI from an experiment into the default way work gets done. With Caylent and AWS as strategic partners and governance embedded throughout its agentic SDLC, Smarsh is demonstrating how large enterprises can safely harness generative AI to enhance and enrich the software and services they provide to even the most highly regulated customers.
