Why Automation Testers Who Ask Questions Back Impress Me Most

Interviewer: Please automate the login page.

(Candidate opens IDE and writes a test that types
 admin / admin123 and checks the URL contains
 '/dashboard'.)

Interviewer: What about a locked account?

Candidate: Oh, I can add that.

Interviewer: And the CI server has no display.

Candidate: Then I need headless mode... let me change
the driver setup.

(Half the time is spent reworking the test.)

Interviewer: Please automate the login page.

Candidate: Before I start, three quick questions.

1. Which users should I cover - just a valid login, or
   also invalid password and locked accounts?
2. What proves login worked - the URL, or something on
   the dashboard like the user's name?
3. Where will this run - locally, or headless in CI?

Interviewer: Valid and locked users. Check the welcome
message. It must run headless.

Candidate: Great. I'll do those two cases with an
explicit wait for the welcome message, in headless
Chrome, and mention what I'd add with more time.

import java.time.Duration;
import org.openqa.selenium.*;
import org.openqa.selenium.chrome.*;
import org.openqa.selenium.support.ui.*;
import org.testng.Assert;
import org.testng.annotations.*;

public class LoginTest {
  private WebDriver driver;
  private WebDriverWait wait;

  @BeforeMethod
  public void setUp() {
    ChromeOptions options = new ChromeOptions();
    options.addArguments("--headless=new", "--window-size=1366,768");
    driver = new ChromeDriver(options);
    wait = new WebDriverWait(driver, Duration.ofSeconds(10));
    driver.get(System.getenv("APP_URL") + "/login");
  }

  @DataProvider(name = "users")
  public Object[][] users() {
    return new Object[][] {
      {"valid_user",  "Welcome, Valid User"},
      {"locked_user", "Your account is locked"}
    };
  }

  @Test(dataProvider = "users")
  public void loginShowsExpectedMessage(String user, String expected) {
    driver.findElement(By.id("username")).sendKeys(user);
    driver.findElement(By.id("password")).sendKeys(System.getenv("TEST_PASSWORD"));
    driver.findElement(By.cssSelector("button[type='submit']")).click();

    WebElement message = wait.until(
        ExpectedConditions.visibilityOfElementLocated(By.cssSelector("[data-test='login-message']")));
    Assert.assertEquals(message.getText(), expected);
  }

  @AfterMethod
  public void tearDown() {
    if (driver != null) driver.quit();
  }
}

"I'm reading the URL and password from environment
variables, so no secrets are in the code and CI can
set them.

The DataProvider has the two user types we agreed. If
you want expired passwords later, it's one more row.

I'm waiting for the message element to be visible,
rather than using a sleep, because login time varies.

I used a data-test attribute for the message, assuming
the app has one. If not, I'd ask the developers to add
it, because it's more stable than text-based XPath.

With more time, I'd move the locators into a LoginPage
class."

When you choose not to ask, state the assumption:

"I'll assume the test users already exist, so I won't
 automate user creation."

"I'll assume Chrome is the target browser; the driver
 setup is in one method if we need Firefox."

"I'll assume the error message text is fixed. If it is
 translated, I'd assert on an error code instead."

Assumptions give the interviewer a chance to correct
you without you spending time asking.

Pick two that you genuinely care about:

- How long does your main regression suite take, and
  when does it run?
- What happens when a test fails in CI - who looks at
  it first?
- Do developers write or review automated tests here?
- How do you handle test data: fixed users, generated
  data, or database resets?
- Which part of the product is hardest to automate,
  and why?
- What would make a new automation tester successful
  in the first few months?

Before any automation task, consider:

DATA
  Which users or inputs should I cover?
  Is test data available, or do I create it?

EXPECTED RESULT
  What exactly proves the feature worked?

ENVIRONMENT
  Local or CI? Headless? Which browser?

SCOPE AND TIME
  How many cases in the time we have?

Ask the 2 or 3 that change your code.
State assumptions for the rest.
Then start, and narrate.

With a friend acting as interviewer, practise these.
They should have answers ready but only give them if
asked.

1. Automate the search box on a shopping site.
2. Automate adding an item to the cart.
3. Write a test for a file upload form.
4. Automate a dropdown that loads options from a server.
5. Test that a user can log out.

For each: ask your questions (max 3), state
assumptions, then write or describe the test.

After each practice task, check:

[ ] I asked about expected result before coding
[ ] I asked no more than three questions
[ ] Each answer changed something in my code
[ ] I stated assumptions for what I did not ask
[ ] I narrated decisions while coding
[ ] I said what I would add with more time
[ ] I asked one good question at the end

Aim for all seven before your real interview.