Spanish Grammar Recognition

[luckyllama]

Hello,

We have some Spanish call flows within our application. Our callers are experiencing an almost 100% no match rate against our Spanish grammar list and we're troubleshooting to find any potential configuration changes that may help with the matching.

One thing we're looking at currently is the language setting within our grammars. We are currently specifying the language to be 'es'. However, when I reviewed the documentation, it looks like we may need to be using es-US or es-MX. Plum's grammar documentation page says to use 'es-US', but also have noticed 'es-MX' referenced frequently by Plum employees on the dev forums. We've tried both and haven't seen any noticeable improvement with either. Please advise on the correct settings for Spanish.

Here's an example of our current Spanish grammar for reference:

Code: Select all

<grammar type="application/srgs+xml" xml:lang="es" root="KEYWORD_ES" mode="voice">
<rule id="KEYWORD_ES" scope="public">
	<tag>lang='es';</tag>
	<one-of>
		<item>
			<one-of>
				<item>c p l</item>
				<item>ce pe ele</item>
			</one-of>
		</item>
	</one-of>
</rule>
</grammar>

[support]

Thanks for your question. We'll be happy to help as best we can.

What is the actual Spanish phrase that should match the prompt?

[luckyllama]

Our prompt asks the caller to say a "keyword" phrase they heard on the radio (as part of an advertising campaign). The match phrases could be literally anything the advertiser wants to use as the match phrase. In the example I provided, the match phrase was an acronym with the letters C P L, but we were also listening for a few other specific words like 'uno' and 'chipotle.'

[support]

We were able to get a match using the following:

<item>C.P.L.</item>

Please give this a try and let us know if you still have any issues.