Daytective
BUILT FOR WORKDAY EXTEND

Your Extend has a workflow.
Test the whole story.

An open-source testing framework built for Workday Extend.

Turn application source into focused test cases with your AI agent. Run them across the roles your application uses, then see the results with supporting evidence.

Explore the frameworkDiscover the approach
Start with your sourcePMD · POD · AMD · SMD
Daytective showing the Work from anywhere application, its Country selection test, readiness, recent runs and connected source.
Your tests. In one place.Prepare. Run. Review.
Take a closer look

Designed for Workday Extend

Source contextAI-assisted designMulti-role testing

FIVE STEPS. ONE CONNECTED PROCESS.

From the code you write
to the results you can inspect.

Bring your application and your business context. Follow one test from its first suggestion to the evidence behind its result.

  1. 01 / CONNECT

    Start with your
    source code.

    Connect the PMD, POD, AMD and SMD files from your Extend application. Daytective builds context from its pages, components, routes and services.

    That gives test authoring a starting point in the application you actually built. You still confirm how its controls map to the running interface.

    You leave withA source model to build on.
    FROM FILES TO CONTEXTConcept diagram
    PMDPages & fields
    PODShared components
    AMDRoutes & structure
    SMDService definitions
    Your application, connected.

    Pages · Controls · Services

    Source describes the structure. Your knowledge adds the business meaning.

  2. 02 / DISCOVER

    Find what is
    worth testing.

    Review suggestions derived from the application model. Choose the fields, interactions and business rules that deserve a closer look.

    Turn a suggestion into a focused test. Give it a clear purpose and resolve anything still needed before execution.

    You leave withA deliberate starting point, not a blank page.
    A TEST TAKING SHAPEApplication UI
    Real local workspace with a Country selection test, its readiness, source suggestions and a Not run yet status.
    Explore the workspace

    A prepared test in the real interface, using sample data.

  3. 03 / AUTHOR

    Your business rules.
    Your AI agent.

    Work with your own AI agent to turn a testing idea into explicit roles, inputs, actions and expected outcomes. Use your corporate or personal subscription.

    You supply the business context and review the case. AI helps with authoring; the local execution engine does not call AI services to decide what to do.

    You leave withA readable scenario, ready for validation.
    MAKE THE EXPECTATION EXPLICITIllustrative case
    YOUR BUSINESS RULE

    “A request should keep the country the employee selected.”

    You + your AI agent
    Role
    Employee
    Action
    Select a country and submit a request.
    Expect
    The saved country matches the selection.
    Review
    Confirm the controls, test data and permitted write.

    An example of test intent, not an executable scenario or a test result.

  4. 04 / EXECUTE

    Follow the process.
    Across its roles.

    Run a validated scenario through Playwright using the configured browser sessions. An employee’s request and a manager’s decision belong to the same process.

    Identity checks and record correlation help keep each action tied to the intended role and request. If the outcome of a write is uncertain, stop for investigation rather than submit it again.

    You leave withA recorded sequence of actions and observations.
    ONE PROCESS, DIFFERENT PERSPECTIVESExample flow
    EmployeeCreate a request
    Own browser session
    Match the request, process and task
    ManagerReview and decide
    Own browser session

    Uncertain outcome? Pause and inspect before taking another action.

    Illustrative employee-to-manager flow. Available roles depend on your application and configuration.

  5. 05 / REVIEW

    See the result.
    Trace the evidence.

    Inspect what ran, which role acted and how the observations compare with your expectations. Local HTML and JSON reports keep those details together.

    A confirmed result, a failed expectation and an uncertain action need different responses. Read the supporting evidence before deciding what to change or investigate.

    You leave withA result you can explain, not just a status.
    WHAT SUPPORTS THE RESULT?Report anatomy
    Read beyond the status.HTML for review · JSON for analysis
    1. 01
      Expectation

      What should the application do?

    2. 02
      Observation

      What did the run actually observe?

    3. 03
      Supporting evidence

      Which recorded details support that conclusion?

    A guide to reading a report, not a screenshot or a claim of a successful run.

KEEP THE PROCESS IN YOUR HANDS

Bring your next workflow.

Local execution Your AI agent Inspectable reports

View on GitHub

A closer look at Daytective

Explore the interface with a sample project.

Daytective application workspace showing the Work from anywhere application, Country selection test, readiness, recent runs and connected source.

Scroll to explore the full-size screenshot.