About eXp Realty‍

 eXp Realty® is “the most agent-centric™ real estate brokerage on the planet.” As a publicly traded, cloud-first company (NASDAQ: EXPI), eXp runs a complex technology ecosystem. From agent dashboards and commission tracking to CRM integrations and AI-powered support tools, all of it needs to work flawlessly for the tens of thousands of agents who depend on them daily.

‍

The Challenge

eXp's Software Assurance team validates that every feature shipped by engineering works as specified. For each Jira ticket, QA engineers would manually walk through the acceptance criteria by logging in, navigating to the right page, clicking through flows, and verifying outcomes by hand.

‍

Each QA engineer spent hours per sprint clicking through checks that could be described in a few sentences of plain English. There was no easy way to turn those manual checks into reusable regression tests without investing in brittle automation that required constant maintenance.

‍

"We wanted a way to automate our acceptance criteria checks without having to allocate hours of QA bandwidth," said Sherri Delbridge, Sr Director of Software Assurance at eXp Realty.

‍

The Solution

eXp adopted Docket as a visual AI testing platform that lets QA engineers write tests in natural language by describing what to test the same way they'd describe it to a human. Docket navigates the actual UI using screen coordinates rather than code selectors, clicking through flows the way a person would.

‍

The workflow fits naturally into how eXp already works: a developer marks a Jira ticket ready for QA, and an engineer writes the acceptance criteria as a Docket test in plain English. Docket runs through the flow visually, validates each criterion, and reports the results. Over time, those one-off acceptance checks accumulate into a permanent regression suite with no extra effort required.

‍

"The ease of use and the self-healing are the big ones for us," said Delbridge. "People write a test once, and it keeps working even when the UI changes. If we did this with traditional automation, things would break the moment something moved."

‍

Because Docket uses vision-based coordinates with built-in self-healing, tests automatically adapt when UI elements shift. A banner appears at the top of the page and pushes a button down? Docket finds it and clicks it anyway. No one has to go back and fix a broken selector.

‍

eXp's teams also combine Docket's AI steps with recorded actions, using AI for dynamic flows where the content changes every time, and recorded steps for predictable sequences where speed and cost efficiency matter. This mix gives them flexibility without sacrificing coverage.

‍

The Results

Adoption spread organically. After an initial lunch-and-learn session, usage expanded from a handful of engineers experimenting to users across multiple teams, including eXp's Zoocasa team, which plans to integrate Docket into their CI/CD pipeline to run tests on every deploy. There are 100+ eXp realty team members with Docket accounts.

‍

What started as a tool for one-off acceptance checks is now embedded in how eXp's QA organization operates. Engineers across teams write tests in plain English, run them against live environments, and build up regression coverage with every sprint - without writing or maintaining automation scripts.

‍

"The people using it love it," said Delbridge. "And the more they use it, the more they find new things to test with it."

Key outcomes:

• Acceptance criteria validation that took hours per sprint is now automated in minutes.
• One-off QA checks naturally accumulate into a growing regression suite
• 100+ team members on Docket with adoption continuing to increase
• No additional QA hires required to keep pace with engineering velocity

‍

‍