How changes are tested
How every change to your agent is tested against simulated callers before it reaches your line, and how live calls are checked afterwards.
3 min read
Terms
- Scenario. A test situation, such as a new patient booking or a cancellation, with a caller, a goal and an expected outcome.
- Rubric. The list of things the agent must get right for your practice, from the critical to the minor.
- Regression. A case the live version handled well that a new version handles worse.
When testing happens
Every change to the agent's instructions is tested before it reaches a live line. That includes a new greeting, a new answer, or a change to how the agent books. The version your callers speak to stays live until the new one has passed.
How a test call works
- A simulated caller rings. Each scenario is played out by a simulated caller with its own goal, who speaks the way real callers do.
- The proposed version answers. The agent under test runs on the proposed instructions, not the live ones.
- Your schedule is never touched. The agent looks things up as it would on a live call, but nothing is booked, moved or cancelled in your practice system.
- The call is graded. What the agent said and did is graded against your rubric.
What gets tested
Scenarios are built around your practice's own calls, not a generic script. A typical set covers:
| Scenario | What the agent must get right |
|---|---|
| New patient booking | The right visit type, provider and slot, with the patient's details captured correctly |
| Reschedule or cancellation | Finding the existing appointment and following your cancellation rules |
| Questions about the practice | Hours, location, insurance and services, answered from your own information |
| Urgent symptoms | Recognising the call and passing it to a person straight away |
| Requests for medical advice | Declining to give it and routing the caller to your clinical team |
| Unclear or difficult callers | Asking again politely rather than guessing |
The pass mark
Every scenario must pass before a change goes live, and a change that does worse on anything the live version already handles well is blocked. The bar is set high on purpose: a change that is almost right does not reach your patients.
Going live
A person on your 2care team approves every change before it goes live. Each version is kept, so your team can see what changed and when.
After go-live
- Live calls are reviewed. Calls are checked against the same rubric your agent was tested on.
- Grades can be corrected. A person can override any grade, so the reviews stay accurate.
- New versions are watched. A new version's live results are compared with the version before it, and one that does worse is taken back to the previous version.
- Fixes go through the same tests. When the same problem repeats, a fix is prepared, tested like any other change and approved by a person.
Questions
Do test calls touch our schedule?
No. Nothing is booked, moved or cancelled in your practice system during testing, so your schedule stays as it is.
Who approves a change?
A person on your 2care team. A change goes live only after it has passed testing and been approved.
What happens if a change fails?
It does not go live. The current version keeps answering your calls while the change is reworked and tested again.
Can we add our own scenarios?
Yes. Send us a call that went wrong, or a situation your front desk worries about, and it becomes part of the tests every future change must pass.
What does testing focus on?
What the agent says and decides: understanding the caller, choosing the right action and following your rules.
Related
Was this page helpful?