How to Apply a Quality Studio Recommendation

Created by Heather Perry, Modified on Thu, Jun 25 at 8:43 AM by Brandi Weyant-Ceron

When a check fails in Quality Studio, the system doesn't just flag it — it tells you exactly what to change and why. This article explains how to read a Quality Studio recommendation and apply it to your AI Receptionist.

What's covered in this article

  • What a recommendation includes

  • How to apply a recommendation

  • What happens after you apply

  • Frequently asked questions

What a recommendation includes

Every failed check that Quality Studio can act on generates a recommendation with three parts:

  • What should have happened: a description of the correct behavior for this scenario

  • The instruction change: the specific edit to make to your AI Receptionist instructions

  • Why: the reasoning behind the fix, connected back to the goal and scenario you defined

How to apply a recommendation

Applying a recommendation is a three-step process:

Step 1: Review the recommendation

Open the call details for any call with a failing check. If Quality Studio has generated a recommendation, it will appear below the failing check with the proposed instruction change. Read through it — if it looks right for your business, proceed to apply. If you want to adjust the wording, you can edit it before applying.

Step 2: Accept or modify, then apply

  • Click Apply to accept the recommendation as-is. The change is written directly to your AI Receptionist instructions, no need to navigate to a separate settings page.

  • Or, edit the proposed instruction change first, then click Apply to save your modified version.

Note: The edit is scoped to just the specific field the recommendation touches. Only that field is updated; the rest of your instructions are unchanged.


Step 3: Retest

Run the same scenario again, simulated or test call,  to confirm the fix worked. If the previously failing check now passes, you've closed the loop.

What happens after you apply

  • Your AI Receptionist instructions update immediately.

  • The change takes effect on the next call, simulated, test, or real.

  • Your AQI score updates in the next scoring window (default: last 7 days). After retesting, you should see the previously failing check move to pass and your AQI trend upward.

Frequently asked questions

What if I see an upgrade prompt instead of the Apply button?

Some recommendations require custom prompting, which is available on paid plans. If a recommended change involves custom prompting, you'll see an upgrade prompt instead of Apply. All other recommendations can be applied on any plan, including Free.

Can I undo a recommendation I've applied for?

You can edit your AI Receptionist instructions manually at any time through the dashboard. If you want to reverse a change, navigate to the instructions field that was updated and edit it back.

What if the recommendation doesn't look right?

Edit it before applying. If recommendations consistently seem off for a particular scenario, try adding more detail to your scenario description, the more specific your caller and receptionist descriptions, the more targeted the recommendations will be.

Do I have to apply every recommendation?

No. Recommendations are suggestions, not requirements. You can choose not to act on one. Your AI Receptionist continues handling calls as usual regardless.

What does Shared with Smith.ai to improve AIR mean on a failed check?

Some checks, especially Caller Experience checks, can fail due to factors outside your control (like a caller hanging up early). These are displayed for transparency with a note that the Smith.ai team is aware and working on it. No action is required from you on these checks.

Questions? Contact us!

  • Already a client? Email support@smith.ai or call (650) 727-6484.

  • Not yet a client? Schedule a free consultation at smith.ai/booking-va or email sales@smith.ai.

Was this article helpful?

That’s Great!

Thank you for your feedback

Sorry! We couldn't be helpful

Thank you for your feedback

Let us know how can we improve this article!

Select at least one of the reasons
CAPTCHA verification is required.

Feedback sent

We appreciate your effort and will try to fix the article