Webinar 16: Back to Basics with Reflect AI A Step-by-Step Guide for All Users - Smart Bear
Objective of Webinar
The webinar aimed to introduce Smart Bear’s Reflect, a functional test platform for web applications, highlighting its
AI-driven test automation (e.g., generating test steps from prompts, self-healing selectors).
Record-and-playback functionality for streamlined test creation.
Integration capabilities with tools like Jira, Zephyr, and CI/CD pipelines.
Privacy and security under Smart Bear’s Halo AI platform.
Company Name and Presenter
Company: Smart Bear (host of the webinar).
Presenters:
Mel Gage, Solutions Engineer (5.5 years at Smart Bear).
Sarah Mahood, SE team member (supporting Reflect and Zephyr).
Brief Summary of the Webinar
Reflect is a SaaS-based test automation tool designed for:
AI-augmented testing: Generates test steps from natural language prompts and adapts to UI changes.
Visual validations: Validates text and images on web pages.
Self-healing: Uses AI to fix broken selectors during execution.
Integrations: Works with Jira, Zephyr, Slack, and CI/CD tools (e.g., Jenkins, Bamboo).
Security: Data sent to OpenAI is limited (test steps, screenshots); no training on customer data.
Key demo takeaways
Created tests via record-and-playback and AI prompts.
Showcased self-healing for broken selectors.
Demonstrated Zephyr integration for test case management.
Feature and Technical Aspects
AI Capabilities:
Generates test steps from text prompts (e.g., "Click the contact link").
Self-heals broken selectors using AI (optional feature).
Supports multilingual prompts (e.g., German).
Test Execution:
Runs in virtual environments (Chrome, Firefox, Edge).
Supports parallel execution and suites with decision trees.
Integrations:
Test Management: Zephyr, Xray, TestRail.
CI/CD: Jenkins, Bamboo, GitHub Actions.
Messaging: Slack, email notifications.
Security:
Data encrypted; no customer data used to train models.
Enterprise options for extended test-run storage.
How GoTestPro Can Compete
Niche Focus: Specialize in desktop/WPF testing (Reflect is web-only).
Cost Efficiency: Offer competitive pricing for mid-market teams.
Advanced AI: Provide deeper dynamic content handling or bias detection in AI models.
Open-Source Flexibility: Integrate with Playwright/Selenium for users wary of vendor lock-in.
Additional Observations
Audience Interest: Questions centered on AI reliability, CI/CD integration, and dynamic content handling.
Limitations: Reflect lacks desktop app testing and advanced conditional logic (e.g., branching).
Roadmap: Mobile testing (via BitBar/Sauce Labs) and nested test suites are upcoming.
Question & Answer:
How strict are AI prompts?
Be specific (e.g., "validate text X"), but Reflect adapts to varied phrasing (even multilingual).
Can Reflect handle dynamic UI elements?
Uses AI to self-heal selectors; documentation covers dynamic content strategies.
Does Reflect support API testing?
Limited to web UIs; API tests require Smart Bear’s ReadyAPI or JavaScript steps.
How is data privacy handled?
No customer data trains models; inputs are prompts/screenshots (deleted after 30 days by default).
Can tests run in private environments?
Yes, via tunneling or IP whitelisting for internal apps.
Will AI replace manual testers?
No; AI augments testers by automating repetitive tasks (e.g., generating steps).
Mobile testing support?
Coming soon (BitBar/BrowserStack integration).
Demo Screenshots:
