From UI to AI: How Effectively Can Joule Replicate System Actions?Can Joule accomplish tasks faster?

[SwethaMariyappa]

Initial Assumption

At first, I was skeptical. My early interactions with Joule suggested that its behavior didn’t align precisely with traditional UI-based transactions. However, after experimenting with a few scenarios, I realized that Joule not only mirrored the UI flow but also simplified it—achieving the same outcomes with fewer steps and in less time.

Test Scenario: Changing an Employee's Location with an Effective Date

Traditional UI-Based Flow

To change an employee’s location through the standard UI, a manager would typically follow these steps:

1. Log in as a Manager.
2. Locate the direct report whose location needs updating.
3. Navigate to Job Information → Take Action → Change Job Information.
4. Select an Effective Date and choose the New Location from the dropdown.
5. Click Save.

Joule-Based Flow

Now let’s see how the same task plays out through Joule:

1. Open Joule.
2. Ask, “Show my direct reports.”
3. Select the desired employee.
4. Instruct Joule, “Update location.”
5. Joule prompts for the new location.
6. Provide a value, e.g., “New York.”
7. Joule identifies multiple valid locations and prompts for clarification.
8. Select the correct location.
9. Submit.

Result: Location is updated in the employee’s profile.

The Catch: What About the Effective Date?

On comparing both methods, I noticed a subtle but important difference.

In the chatbot approach, the system defaulted the effective date to the current date—without prompting me. This behavior is problematic in real-world scenarios, where managers often schedule changes for future dates.

So, what happens if I try to update the effective date afterward?

Unfortunately, Joule cannot update the effective date of a previously submitted record. This is due to system restrictions—chatbots typically lack permission to modify historical records.

Can we simply re-submit the same data with a future date?
Not quite. If the location remains unchanged between the original and new record, the system ignores the update due to redundancy.

Here’s the error message that illustrates this behavior (see screenshot):

Solution: Crafting an Effective Natural Language Query

To overcome this limitation, I tried submitting a more complete command:

"Update the location for Faith Marshall to New York with an effective date of July 2025."

What happened next surprised me:

• Joule asked me to confirm the location (since "New York" maps to multiple valid options).
• Once confirmed, Joule returned a summary for my review.
• The summary included all the correct details, including the effective date.
• I clicked Submit.

Result:
✔️ Record successfully updated in the system.
✔️ Effective date was correctly set.
✔️ Verified via the UI.

Final Thoughts

This test clearly demonstrates the potential of AI-powered chatbots like Joule to handle complex HCM transactions with ease—when prompted correctly. With the right natural language input, Joule not only mimics the UI behavior but streamlines it, improving both efficiency and user experience.

For organizations adopting digital assistants, the key takeaway is this:

“Train your users not just to use the chatbot—but to converse with it intelligently.”

This shift in approach can unlock significant productivity gains, reduce manual steps, and support self-service operations at scale.

This is just an example- the options to perform smarter tasks in shorter period of time can be accomplished by Joule for Success Factors HCM pages. To give more insight-:

UI-Based Flow: Manual Steps and Clicks

To perform this action via the UI, a manager typically goes through the following steps:

1. Login as Manager (1 click: Login)
2. Search for the employee (1–2 clicks depending on how search is configured)
3. Go to Job Information (1 click)
4. Click Take Action → Change Job Information (2 clicks)
5. Select Effective Date (1 click)
6. Open Location dropdown and choose new location (2 clicks)
7. Save the changes (1 click)

Estimated total: 8–10 clicks
Time: ~1–2 minutes (depending on network speed and UI responsiveness)

Chatbot-Based Flow: Conversational Efficiency

1. Open Joule (1 click)
2. Ask to show direct reports (typed or selected)
3. Select employee from list (1 click)
4. Ask to update location (typed)
5. Respond to prompt with new location (typed)
6. Choose the right match from suggested locations (1 click)
7. Review and confirm (1 click)

 Estimated total: 3–4 clicks, 3 messages typed

 Time: ~30–45 seconds

Summary Comparison Table

Action	UI-Based Flow	Chatbot (Joule)
Clicks	8–10	3–4
Typing	Minimal	Moderate (3 messages)
Time Taken	~1–2 minutes	~30–45 seconds
Effective Date Flexibility	Manual Input	Requires explicit mention
Error Risk	Medium (missed fields)	Low (guided prompts)