ChatGPT With Selenium: Automate Tests Using AI in 2026

Creating Test Scenarios

Manually deriving test scenarios from requirements takes time and is prone to gaps. ChatGPT does this in seconds once you describe the feature.

Sample Prompt:

Sample ChatGPT Response:

Generating Test Cases

Converting scenarios into structured test cases with steps, data, and expected results is the most time-consuming part of the testing cycle. ChatGPT handles this on demand, in any format you specify.

Sample Prompt:

Sample ChatGPT Response:

Below are detailed positive test cases for login functionality using Email and Password presented in a structured tabular format. These focus only on valid scenarios where login should succeed.

Generating Test Scripts

ChatGPT can scaffold a full Selenium framework from a single prompt, including project structure, base classes, Page Object Model classes, and test classes. This is where it saves the most time, especially for manual testers stepping into automation.

Sample Prompt:

Sample ChatGPT Response:

ChatGPT first outputs the project structure, then generates each class in order:

project
 └── src/test/java
       ├── pages
       │     └── LoginPage.java
       ├── tests
       │     └── LoginTest.java
       └── base
             └── BaseTest.java

BaseTest.java

package base;
import org.openqa.selenium.WebDriver;
import org.openqa.selenium.chrome.ChromeDriver;
import org.testng.annotations.*;
public class BaseTest {
    protected WebDriver driver;
    @BeforeMethod
    public void setUp() {
        driver = new ChromeDriver();
        driver.manage().window().maximize();
        driver.get("https://example.com/login");
    }
    @AfterMethod
    public void tearDown() {
        if (driver != null) {
            driver.quit();
        }
    }
}

LoginPage.java

package pages;
import org.openqa.selenium.*;
public class LoginPage {
    WebDriver driver;
    // Locators
    private By emailField = By.id("email");
    private By passwordField = By.id("password");
    private By loginButton = By.id("loginBtn");
    // Constructor
    public LoginPage(WebDriver driver) {
        this.driver = driver;
    }
    public void login(String email, String password) {
        driver.findElement(emailField).sendKeys(email);
        driver.findElement(passwordField).sendKeys(password);
        driver.findElement(loginButton).click();
    }
}

LoginTest.java

package tests;
import base.BaseTest;
import pages.LoginPage;
import org.testng.Assert;
import org.testng.annotations.Test;
public class LoginTest extends BaseTest {
    @Test
    public void verifySuccessfulLoginWithValidCredentials() {
        LoginPage loginPage = new LoginPage(driver);
        loginPage.login("[email protected]", "Test@123");
        Assert.assertTrue(loginPage.isDashboardDisplayed(),
                "Login failed: Dashboard not displayed.");
    }
}

ChatGPT writes the code, but you still copy, paste, compile, and run it manually. TestMu's KaneAI closes that gap with a purpose-built AI testing agent for Selenium.

• Natural language test authoring: Describe the test in plain English and KaneAI generates Selenium code automatically.
• Native test execution: Run tests directly from the platform without copying code into an IDE or build tool.
• AI self-healing: Broken locators are fixed automatically when the UI changes, cutting maintenance effort significantly.
• Selenium export: Export generated tests as Selenium Java, Python, or JavaScript for any CI/CD pipeline.
• Jira integration: Convert Jira ticket descriptions into executable Selenium test cases using AI, without writing a prompt.

See the KaneAI getting started docs to connect it with your existing Selenium workflow.

Generating Test Data

The same flow behaves differently with different inputs, valid, invalid, boundary, and empty. Building that data manually is slow and often incomplete. ChatGPT generates it in one prompt, including a ready-to-use DataProvider in TestNG.

Sample Prompt:

Sample ChatGPT Response:

Below is sample test data that can be used to test a login functionality with email and password. The data covers valid credentials, invalid credentials, boundary values, and format validations to ensure comprehensive testing.

Follow-up Prompt:

Follow-up Prompt Code Response:

public class LoginDataProviderTest {
    @DataProvider(name = "loginTestData")
    public static Object[][] loginData() {
        return new Object[][] {
                // Valid credentials
                {"[email protected]", "Password@123", "success"},
                // Invalid email format
                {"usertest.com", "Password@123", "error"},
        };
    }
    @Test(dataProvider = "loginTestData", dataProviderClass = LoginDataProviderTest.class)
    public void verifyLogin(String email, String password, String expectedResult) {
        driver.findElement(By.id("email")).sendKeys(email);
        driver.findElement(By.id("password")).sendKeys(password);
        driver.findElement(By.id("loginBtn")).click();
        AssertHelper.assertResult(expectedResult);
    }
}

Handling Dynamic Web Elements

Dynamic web apps generate element IDs at runtime and change the UI frequently, making locators the top cause of flaky test failures. Paste the HTML into ChatGPT with a prompt asking for stable locators and it avoids auto-generated classes automatically.

Sample Prompt:

Sample ChatGPT Response:

To create stable Selenium locators and avoid flaky tests, we should avoid dynamic CSS classes like sc-kAyceB, cgUUbz, etc. (these are usually generated by CSS-in-JS frameworks and change frequently). Below are recommended reliable locators.

// 1. Email Address Field
// Best Locator (ID)
By emailField = By.id("email");
// Alternative (Name)
By emailField = By.name("email");
// XPath Using Label (Very Stable)
By emailField = By.xpath("//label[text()='Email Address']/following-sibling::input");

// 2. Password Field
// Best Locator (ID)
By passwordField = By.id("password");
// Alternative (Name)
By passwordField = By.name("password");
// XPath Using Label
By passwordField = By.xpath("//label[text()='Password']/following::input[@type='password']");

Best practices to prevent flaky tests:

• Prefer id or name
• Avoid auto-generated classes
• Use label-based XPath

ChatGPT gives multiple locator options per element ranked by stability, so you can pick the best fit for your app.

Debugging and Fixing Selenium Test Scripts

Paste any Selenium error or stack trace into ChatGPT and it returns the probable cause along with a ranked fix list. NoSuchElementException in Selenium is the most common failure, so that is the example used here.

Sample Prompt:

Sample ChatGPT Response:

org.openqa.selenium.NoSuchElementException means Selenium could not find the element in the DOM when the command executed. This is one of the most common issues in Selenium automation and usually happens due to timing, incorrect locators, or frame/context problems. Below are the most common causes and fixes.

1. Element not loaded yet (most common cause)

If the page loads slowly or elements appear after AJAX calls, Selenium tries to locate the element before it exists.

// Problematic code
driver.findElement(By.id("loginBtn")).click();

// Fix using Explicit Wait
WebDriverWait wait = new WebDriverWait(driver, Duration.ofSeconds(10));
WebElement loginButton = wait.until(
        ExpectedConditions.visibilityOfElementLocated(By.id("loginBtn"))
);
loginButton.click();
// Explicit waits are the best way to avoid flaky tests.

2. Incorrect or unstable locator

The locator may be wrong or dynamically changing.

// Bad locator
By loginBtn = By.className("btn-primary");

// Better locator - use stable attributes like id, name, data-*, aria-*
By loginBtn = By.id("loginBtn");

Quick checklist when you see NoSuchElementException:

• Check locator in browser DevTools
• Verify element exists in DOM
• Add explicit wait

The checklist alone cuts investigation time significantly. Most NoSuchElementException failures are solved by one of the three fixes above.