@Kelly Lynch yes sure

@Malik Ahmed Thanks, I won't disappoint you.

@Okasha Khan thanks

@Brendan Jowett Good timing on this one — I'm actually mid-build on a client project that's exactly this use case: WhatsApp outreach agent, text + voice, targeting a few thousand contacts. My original stack was Twilio as the BSP + Retell for the voice layer, bridging the two for WhatsApp Calling. Solid on paper, but I've hit real friction getting the Twilio account through compliance review this past week — nothing technical, just their KYB process dragging way past their own stated turnaround. Watching this, the appeal of ElevenLabs' native WhatsApp integration is obvious: no separate BSP account, no bridge to build/prove between two platforms — chat and voice through the same agent, same channel connection. Trying to figure out if it's actually the better call for a business messaging use case (not just 1:1 assistant chat) or if there's a catch I'm not seeing yet — anyone here run ElevenLabs WhatsApp at real volume, or hit the same 2000-conversation Meta threshold for voice either way? Comparing both seriously before I commit either direction.

@Yamada Ryo Really solid breakdown, thanks. A few thoughts on your specific points: Template approval — that one's actually a wash either way, since Meta owns that process regardless of which BSP or platform sits in front of it. Twilio, ElevenLabs, doesn't matter, same review queue. The control-over-backend-integration point is the one that actually matters for my use case. My current setup keeps all the business logic (scoring, qualification state, consent tracking, budget gating) in my own orchestration layer, with Twilio/Retell just handling the channel. From what I saw in the demo, ElevenLabs' agents do support tool-calling out to external systems + a knowledge base, so my read is it's not necessarily all-or-nothing — the interesting path might be using ElevenLabs purely as the channel/voice layer while keeping my own backend as the source of truth, rather than migrating the actual business logic into their platform. Latency at volume and template throughput under real load, I genuinely don't have data on yet — that's exactly what I want to pressure-test before committing either direction. I'll report back once I've actually pushed real volume through it instead of just going off the demo.

Love this thank @Brendan Jowett

Gonna try Claude Code + Upwork since I’ve never used it. Might be worth exploring!