Building Call Scheduling Across Time Zones for AI Voice Agent using Temporal

Imagine scheduling thousands of calls daily, ensuring every time zone and scenario is handled with precision. I’ll take you through how we cracked this problem with Temporal for our AI Voice Agent, balancing complexity with simplicity in execution

Recently, we built an AI Voice Agent, which handles Incoming and Outgoing calls. We had a requirement to schedule mass outgoing calls. In the talk, we will be talking about how we built BULK outgoing call scheduler using Temporal and what challenges we faced.

1. Different users can be in different timezone. (You can’t call someone at night xD)
2. If call fails, rescheduling the call, based on legal compliance of different countries. (You dont want the number to be flagged as spam)
3. Managing outbound throughput, so it doesnt eat up resources for inbound calls.
4. Compliance of scheduling mass calls (Failing to filter out DNC numbers can result in fines)

We will discuss how we solved all these technical difficulties using Temporal.

I have given 3 DevFest talks, and few internal talks at workplaces, I host podcasts. I have built softwares used by 10s of millions of users each day. Also, I am a fun host.

I will be accompanied by ~2 other folks, Lead Engineer and AI Director at Observe AI. Who have previously spoken at FOSS United, SRE CON (APAC), EM NLP and a lot of other tech events and have also built the AI Voice Agent Platform.

Talking about technical requirements, we will just need a way to connect a macbook to projector.