Epsilon layers natural-language UI checks over its Playwright regression suite as a managed platform service.

Epsilon built intent-driven UI testing into its shared agent platform using Amazon Nova Act and Amazon Bedrock AgentCore Browser. The company’s 50-engineer Agentic Platform team now offers the capability as a managed service. Product teams can add natural-language UI checks without building browser infrastructure or identity integration themselves.

Epsilon is a data, technology, and marketing-services company within Publicis Groupe. Its existing Playwright regression suite remains the backbone for deterministic checks. The platform team treated intent-driven testing as an additive layer, because selector-based tests can break on cosmetic UI changes even when functionality works.

For builders, the service turns UI testing into another capability on a multi-tenant agent platform. Teams consume it through the same API surface they use for other agent tools, while the platform handles federated identity, pre-flight authorization, MCP integration, and cross-account provisioning. Operators avoid duplicating browser infrastructure, dependency management, and identity work for each application.

Epsilon’s published architecture includes the core Nova Act and AgentCore Browser integration code plus an 18-step test workflow. The next signal is measured coverage and maintenance burden across product teams, along with adoption of the shared testing service. Operators evaluating similar systems should compare intent-driven checks against selector-based suites, not replace what already works.

What matters

• Epsilon’s Agentic Platform team delivered UI testing as a managed capability through Amazon Nova Act.
• Teams gain shared identity, security, and browser infrastructure without rebuilding each capability.
• Watch how Epsilon measures coverage gains and whether other product teams adopt the managed API.

Why it matters

Watch how Epsilon measures coverage gains and whether other product teams adopt the managed API.

This GenAI News article was prepared in original wording using reporting and materials published by AWS Machine Learning Blog. Source reference: https://aws.amazon.com/blogs/machine-learning/building-intent-driven-ui-testing-with-amazon-nova-act/.

Drafted by the GenAI News review pipeline.