Is anyone tracking Meta AI (Llama 4 Scout) as a discovery surface?

Been thinking about this since Llama 4 Scout is quietly running Meta AI across WhatsApp, Instagram, and Messenger, which is a genuinely massive footprint, arguably bigger reach than ChatGPT in a lot of markets outside the US.

And yet almost nobody in the AI-visibility/GEO conversation talks about it. Every "which LLMs should I track" list defaults straight to ChatGPT, Gemini, Perplexity, Claude.

Makes sense in one way. Scout's a smaller, more efficient model (10M token context, MoE architecture) built for scale and cost, not for being anyone's primary "ask it everything" assistant the way ChatGPT is. But the distribution is the part that's hard to ignore: it's sitting inside apps billions of people already have open every day, recommending things in casual chat context rather than a dedicated search box.

Genuinely curious, for anyone building or using AI-visibility tools, does Meta AI even show up in what you're measuring?

Is it worth the effort to track, or is the actual query volume for anything commercial/brand-related just too low to matter yet compared to the "real" search-intent surfaces?