HighLevel Voice AI now uses OpenAI speech-to-speech models

TL;DR

HighLevel Voice AI now runs on OpenAI's new speech-to-speech models. HighLevel has not confirmed if you need to act.

Who this affects

HighLevel has not confirmed who is affected by this change. It is unknown if this update impacts all users of Voice AI or specific plans. Readers can safely ignore this until more details are released.

What changed

HighLevel Voice AI now runs on OpenAI's new speech-to-speech models. These models process audio input directly to audio output, eliminating the need for speech-to-text and text-to-speech conversion steps. The new models are designed to preserve tone and urgency during calls. They also allow for handling interruptions without conversation breakdown. HighLevel has not confirmed if this replaces or breaks existing functionality. Limits or plan gating for this feature are unknown. The rollout status is also unknown.

Where to find it

HighLevel has not published the location of this updated functionality yet. Since this update pertains to the underlying technology of Voice AI, it will likely be integrated into existing Voice AI configurations within workflows or specific settings related to AI phone agents. Users should monitor HighLevel's official announcements for specific navigation paths.

How agencies will use this

Agencies utilizing HighLevel's Voice AI for client lead qualification or customer service automation could see improved call quality. For example, in a workflow designed to qualify inbound leads, a Voice AI agent previously experienced delays due to speech-to-text and text-to-speech conversions. With the new models, the AI agent could now engage in more natural, fluid conversations. If a lead interrupts the AI, the new model should handle it without the conversation breaking down, leading to a better user experience and potentially higher qualification rates. The preserved tone and urgency could also make the AI interaction feel more human, increasing lead engagement in a multi-step workflow before passing the lead to a human agent via a calendar booking.

Our take

This is a major underlying technical update to HighLevel's Voice AI. While the specifics of its impact on existing functionality, user access, and rollout status are currently unknown, the move to OpenAI's direct speech-to-speech models suggests a significant improvement in AI-driven voice interactions. The ability to preserve tone, urgency, and handle interruptions more effectively addresses key friction points in current AI call experiences. Agencies invested in Voice AI should pay attention, but HighLevel has not provided enough information yet to act this week.