Thoughtly Office Hours - March 17, 2025

Join us weekly for Office Hours to discuss all things Thoughtly. RSVP here to be notified about upcoming sessions.

​

Voice Cloning Feature Update [00:06:00]

• Introduction of the new voice cloning feature within Thoughtly which allows users to clone voices directly from the dashboard. Users can find this feature in the voice library section where they can clone their voice to use as a voice agent. This service now supports V1.8 agents, and will soon support V2 agents.
• The voice cloning currently supports English with plans to include more languages. This can be particularly useful for local personalities, such as real estate agents with a recognizable voice.
• Discussion about the voice cloning vendor, Cartisia, and their claim that their voice clones are preferred by users over the leading competitor, 11 Labs. Testing and feedback on this feature were encouraged.

​

Status Page Introduction & SIP Integration [00:15:00]

• Announcement of the new Thoughtly status page where users can check real-time metrics like latency and system performance across different server regions (Virginia, Oregon, Belgium, Taiwan). This will help users identify any performance issues in their respective areas.
• Update on SIP integration stating it is close to completion. SIP URI will be available for each agent on the dashboard, which can be utilized by users with multiple locations without needing to assign separate phone numbers.

​

Transition to V2 [00:20:00]

• Upcoming rollout of V2 announced for April with a focus on new user interface which will be more intuitive and easier to navigate. V2 promises improved backend functionalities for smoother mid-call actions and enhanced dialogue management.
• The conversation addressed improvements in managing interruptions during calls. The transition to V2 aims to resolve issues with unintentional agent interruptions.
• Clarification about beta testing of V1.8 which is expected to resolve some previous common issues and to garner feedback for further development.

​

Voice Agent Troubleshooting and Scheduling [00:25:00]

• Discussion starting with member Tony Q’s question about integrating vehicle inventory into the Thoughtly system for car dealerships. Recommendations included utilizing a dealer’s API for dynamic inventory management or using smaller datasets directly.
• Explanation given on using Thoughtly’s features effectively by focusing on using APIs for retrieving relevant data in large inventories, instead of attempting to utilize Excel sheets directly within the agents.
• Tips were shared to improve interaction consistency when managing a large inventory through APIs as a best practice instead of relying on bulk upload methods.

​

SIP and Infrastructure Questions [00:30:00]

• Community concern about latency particularly in international calls and the steps Thoughtly is taking to address connectivity like ensuring efficient data center links particularly in Europe.
• Clarification on prioritizing infrastructure by setting local numbers and better understanding how international call routing works to naturally improve latency experiences.
• Acknowledgment of suggestions for infrastructure enhancements such as adding more data nodes.

​

Community Engagement and Collaboration Tools Feedback [00:45:00]

• Discussion on transitioning from Slack due to user dissatisfaction with Slack’s limitations including message retention and organization issues. Suggestions included exploring alternatives like Discord, Reddit for forums, or proprietary community platforms like ‘School’.
• Reflection on the merit of having split tiers or groups for high-engagement vs. general user communities.
• Emphasis on community feedback being crucial in deciding which communication tools to effectively integrate for better collaboration.

​

Final Updates and Meeting Wrap-up [02:00:00]

• Suggestions on improving mid-call action reliability, and specifically discussed was adding features that improve the integration of payment processing during calls.
• Proposal of Objective Nodes prompted as an idea to hold conversational context which would potentially improve complex interaction handling in future updates.
• Closing the meeting participants were encouraged to continue the dialogue in community forums and prepare questions or feedback for the next session.