Mia Labs automates dealership voice agent testing
with Hamming
“It is the repetitive stuff that makes Hamming worth it for us. Being able to handle volume and testing at scale.”
Tosh Toida, QA Lead at Mia Labs
Meet Mia Labs
Mia Labs builds AI voice agents for car dealerships, handling appointment scheduling, store information, and customer inquiries. Mia answers calls, books service appointments, provides store hours and addresses, and manages the complete appointment lifecycle.
As one of Hamming's earliest customers, Mia has grown alongside the platform. Their QA team runs the appointment test series 100 times per week, testing different dealership configurations and after-hours behaviors. Appointment scheduling is the core revenue driver for dealerships, making reliable testing essential.
With nondeterministic AI responses, the team needs to run tests multiple times to ensure consistent behavior. Manual testing simply couldn't keep up with the volume and variety of scenarios they needed to cover.
With Hamming, Mia Labs was able to
The Challenge: Manual Testing Couldn't Scale with Dealership Growth
Testing Mia's appointment scheduling required making multiple phone calls per test: booking an appointment, calling back to verify, rescheduling, verifying again, and finally canceling. Each test cycle took 10+ minutes of focused effort.
With AI producing nondeterministic responses, the team needed to run each test multiple times to ensure consistency. And with new dealerships going live regularly, the testing workload grew faster than the team could handle manually.
Solution: Automated Testing That Scales with Growth
Automated Appointment Workflow Testing
Mia's core functionality centers on appointment scheduling for car dealerships. The QA team needs to test the complete appointment lifecycle: booking, rescheduling, and cancellation, then verifying each action was recorded correctly.
Before Hamming, this meant making multiple phone calls per test and manually checking the results. With Hamming, the team creates test cases that automatically:
- Call to schedule an appointment
- Call back to confirm the appointment was recorded
- Call to reschedule the appointment
- Verify the reschedule was processed
- Call to cancel the appointment
- Call back to confirm the appointment was cancelled
This entire workflow, which previously took 10+ minutes of manual effort per test, now runs automatically and can be executed hundreds of times per week.
Parallel Test Execution at Scale
One of the biggest limitations of manual testing is that a single person can only make one phone call at a time. With an AI voice agent that produces nondeterministic responses, the team needs to run the same test multiple times to ensure consistent behavior.
Hamming enables Mia's QA team to run four or five tests simultaneously instead of sequentially. This parallel execution is critical because:
- AI responses vary between calls, requiring multiple test runs
- Multiple dealership configurations need consistent testing
- Running tests in parallel saves significant time
The team can now execute an entire test suite at once, immediately see which scenarios pass or fail, and quickly identify any inconsistencies in Mia's responses.
Scheduled After-Hours Testing
Testing AI voice agents that can transfer calls to real humans presents a unique challenge. During business hours, test calls might accidentally connect to actual dealership staff, wasting their time.
With Hamming's scheduling features, Mia's team schedules tests to run after hours when no humans are available to receive transfers. This approach provides multiple benefits:
- No risk of test calls reaching real dealership employees
- Test Mia's after-hours behavior and configurations
- Verify voicemail handling and alternate workflows
- Review results in the morning without working late
Different dealerships have different after-hours configurations, and being able to test these automatically ensures consistent quality across all deployments.
“Not only can we have Hamming do it instead of us, we can have Hamming do it four or five times, all at once instead of having one person call and do it one time.”
Tosh Toida, QA Lead at Mia Labs
Why Mia Labs Chose Hamming
As one of Hamming's first customers, Mia Labs has seen the platform evolve and grow alongside their own needs. The relationship has been characterized by responsive support and features that often anticipate what the team needs.
The QA team particularly values the direct communication and engagement they receive. When they suggest features or report issues, they often find that the Hamming team is already working on solutions.
Mia Labs chose Hamming because:
The Impact
Hamming has transformed how Mia's QA team approaches testing. What once required dedicated hours of manual calling now runs automatically, giving the team more time to focus on analysis and edge cases rather than repetitive test execution.
As Mia continues to deploy to more dealerships, Hamming provides the scalable testing foundation they need. The same infrastructure that powers their current testing will enable production monitoring as they grow.
Significant Time Savings
Each appointment test cycle (book, reschedule, cancel, verify) takes approximately 10 minutes manually. With the team running these tests ~100 times per week, Hamming saves hours of tedious calling while providing better coverage through parallel execution.
Testing at Scale
The ability to run multiple tests in parallel means the team can achieve volume that would be impossible manually. When testing nondeterministic AI behavior, running the same test multiple times is essential for confidence, and Hamming makes this practical.
Foundation for Continuous Monitoring
The testing infrastructure built with Hamming positions Mia for the next phase: production monitoring. The team plans to expand into nightly monitoring runs as they scale.
“Hamming is, out of the box, pretty solid for our use case here at Mia Labs. I'm very happy with it.”
Daniel Kirwin, QA Engineer at Mia Labs
Featured customer stories
How Grove AI ensures reliable clinical trial recruitment with Hamming
How Hamming enables Podium to consistently deliver multi-language AI voice support at scale
How NextDimensionAI ships safer, faster healthcare voice agents with Hamming
How Grove AI ensures reliable clinical trial recruitment with Hamming