Loom linked below.
When doing extractions, you have 2 options:
- automatic data extractions
- extraction tools
Why would you use one or the other? If you're looking for something easy and reliable, always use automatic data extractions. They are mapped on a per-run basis on chat and post-call on voice so they add no latency and are executed by a observation model (meaning not the AI doing the conversing). Extraction tools are done live during a call (and per-run basis for chat), this will add latency but will be done live if you process / client's business requires that.