Why this is a common failure point
Speech recognition is trained mostly on a few accents, and voice agents inherit its blind spots. A caller with a strong regional accent gets misheard, the agent confidently acts on the wrong word, and the caller has to repeat themselves or gives up. It rarely shows in testing, because the people testing an agent usually share an accent with the people who built it.How Roark tests it
Accent handling runs the same flow with a caller from each region you pick. The caller says the same things in every call: only the accent changes, so a value that does worse points at the accent, not the conversation. It is a property sweep: every value runs the same flow with everything else held the same, and the report tells you which values did worse than the rest and which differences are just chance.The values
Every value is English spoken with that region’s accent, not another language. To test other languages, use Multilingual.
What it measures
The template seeds these, and your flow’s own metrics run alongside them. Only checks can mark a value Worse on; the other measures explain why.What to look for
- Comprehension first. An accent the agent struggles with shows up as comprehension failures and repeated questions long before the call outcome moves.
- Names, numbers and spellings. Account numbers, addresses and spelled names are where accents cause the most trouble. A flow that collects them tells you more than one that doesn’t.
- 23 values is a big run. At 6 calls per value that is 138 calls per agent. Start with the accents your callers actually have.
Setting one up
Pick the flow to test, then the values to run it across. Every value starts selected; deselect the ones you don’t need. Each value runs your flow again, so the values you keep are the cost of the run. Compared against sets the baseline: a reference value the report pins first, such as silence for background noise. The template suggests the natural one, and you can change it or choose none. It doesn’t change the verdict: whether a value did worse is always decided against every other value combined. Iterations live in the sweep panel rather than in Advanced. Each value runs every attached test case once per iteration:Calls per value = test cases x iterationsEach value needs at least 5 calls to be compared with the others, and the Iterations field won’t go below the count that reaches it. It starts at 6 calls per value, so one call excluded after the run doesn’t drop a value out of the comparison. More calls let the report see smaller differences. For a sweep of 10 values:
Below 5 calls a value can never be flagged, however badly it does, which is why the minimum is enforced rather than suggested. Over the API, a sweep template defaults
iterationCount to 6 and refuses a plan with fewer than 5 calls per value with a 400.How the results are decided
The sweep’s table leads the run report. Its Compared with the rest column gives each value one result:
How “too large to be chance” is decided:
- Each value is compared with all the other values combined, one check at a time, never with a single other value. Two values that happen to land at opposite ends are not a finding on their own.
- A sweep makes many comparisons at once (every value on every check), and some will look different by luck alone. The report accounts for that: of all the values it flags in a run, at most 1 in 10 is expected to be a fluke.
- A value is flagged only when it does worse. Doing better than the rest is not a problem to fix.