VoiceAI Update: Prompt Validation, Real-time Spam Detection

TL;DR

HighLevel's VoiceAI received updates for prompt validation and spam call detection. If you use VoiceAI for speech-to-speech models or handle inbound calls, review these changes.

Who this affects

This update affects agencies using VoiceAI, particularly those leveraging speech-to-speech models like GPT Live, GPT Realtime, and Gemini. Agencies receiving calls that may include spam will also be impacted. HighLevel has not confirmed who is specifically affected by these changes beyond these general use cases.

What changed

VoiceAI has two main areas of improvement:

• Prompt Validation and Optimization: Users can now validate and optimize existing prompts used with speech-to-speech models, including GPT Live, GPT Realtime, and Gemini. A before-and-after comparison is available, allowing users to preview prompt changes before applying them.
• Spam Call Detection: A new feature can identify spam calls in real time. This detection can automatically end unwanted calls and classify them afterward. The system can also automatically handle repeat offenders.

HighLevel has not confirmed any limits or plan gating associated with these features. HighLevel has also not confirmed the rollout status for these updates.

Where to find it

HighLevel has not published the exact navigation path for these new VoiceAI features. Based on the functionality, expect to find prompt validation tools within the VoiceAI configuration settings where speech-to-speech models are managed. The spam call detection settings will likely appear within call-related settings, possibly under the Phone System or specific sub-account communication configurations.

How agencies will use this

Agencies can now refine their AI communication. For example, an agency running an AI assistant for client intake using GPT Live can now test and optimize the prompts that guide the AI's conversation flow. Before pushing a new prompt live, they can use the before-and-after comparison to ensure the AI's responses are accurate and effective. This reduces the risk of deploying a prompt that misinterprets caller intent or provides incorrect information. For inbound calls, agencies can set up spam detection to automatically disconnect known spam numbers, saving valuable time for their clients' sales or support teams. The system's ability to classify calls afterward also provides data for analyzing call quality and further refining anti-spam measures, ensuring that legitimate calls are handled efficiently while noise is filtered out.

Our take

This is a major update for agencies actively using VoiceAI for advanced communication or those struggling with inbound spam calls. The ability to validate and optimize AI prompts directly within the platform is a significant improvement for managing AI assistant performance. The spam detection features address a common pain point for businesses, offering real-time protection and automation for repeat offenders. HighLevel has not confirmed the exact rollout status or specific availability details, so keep an eye out for further announcements.