How to control voice speed in Polly's neural voices?

Hi, as per documentation:

Prosody tag attributes are fully supported by the standard TTS voices. Neural voices support the volume and rate attributes, but don't support the pitch attribute.

(emphasis mine)

I just successfully synthesized below, observing slower speaking rate:

<speak>
    <prosody rate="slow">
        Ciao! Mi chiamo Adriano, e posso leggerti qualsiasi testo tu mi scriva.
    </prosody>
</speak>