AIToday
Large Language ModelsAI Coding AssistantsAmazon AI BlogPublished: Oct 30, 2026, 04:00 JST

Epsilon cuts UI test authoring 4-6x with Amazon Nova Act

Send on LINE
Epsilon cuts UI test authoring 4-6x with Amazon Nova Act

3 Key Points

  1. What happened

    Epsilon's 50-engineer Agentic Platform team layered plain-English UI tests on top of its existing Playwright suite, exposed as a ui-test-agent that product teams consume through the same platform API.

  2. Why it matters

    Intent-based steps survive UI redesigns and can be read by non-engineers, broadening who can contribute tests while removing per-team browser infrastructure overhead.

WHO IT HITSProduct teams and QA leads at Epsilon can author and review UI tests without writing code, while the platform team avoids managing browser binaries or display servers per team.

Not sure about something? Ask the AI

Questions and answers are published on this page.

Summaries like this, every morning.

Context & Analysis

Epsilon's existing Playwright suite was not broken — it remains the regression backbone. The new capability is additive: scripted tests are tightly coupled to selectors and page structure, so cosmetic UI changes can break them even when functionality is sound. Intent-based steps, by contrast, are resilient to such changes and readable by non-engineers.

The platform team's goal was to avoid creating a custom CLI that every onboarded team had to operate. By delivering UI testing as a multi-tenant service through the same API surface used for other agent capabilities, Epsilon keeps onboarding straightforward: provision a cross-account IAM role, register target URLs, and grant access. Each team's tests then run in their own AWS account under a scoped role, maintaining isolation and auditability.

A key operational choice was isolating the Nova Act runtime in a dedicated runner service. The team learned that the signing endpoint for Nova Act targets us-east-1 regardless of the browser session's region, so a separate service with its own dependency lockfile prevents hard-to-debug conflicts with the main backend's Strands agent stack.

FAQ
How much faster is test authoring with Nova Act compared to Playwright?
Epsilon's internal benchmarks show a 4-6x reduction in authoring effort. An 18-step workflow took 5-10 minutes with Nova Act versus 30-90 minutes for the equivalent Playwright script.
What happens after a test workflow passes the initial explore phase?
Epsilon saves the successful trajectory and replays it deterministically. Saved trajectories achieve a 100 percent pass rate across repeated runs.
How do product teams access the UI testing capability?
They use the same platform API as other agent capabilities. The platform team provisions a cross-account IAM role and grants access to the ui-test-agent.
Amazon AI BlogRead Original Article
Send on LINE

AI news that matters for your work, delivered every morning.

Pick your industry and the AI tools you use, and get news related to your work every day.

Free · 30 seconds with Google · unsubscribe anytimeWhat is AIToday? →

Ask AI

Ask AI anything about this article. The AI reads this article, earlier AIToday articles, and Wikipedia, and cites its sources. Q&As are published on this page for other readers too.

Questions and answers are published on this page.

Related Articles

Next articleMeta partners with INCA to expand AI glasses access in Canada