It works the same as the existing `wait_until_text_in` method,
except that it checks the inner HTML instead of text content.