Wen added that customer service agents need to sound human and give callers confidence they can actually solve problems. In his view, once the voice is good enough and a customer engages for the first two or three turns, trust builds over time, to the point where people may not feel they need to talk to a human.

Accuracy remains a sticking point. Wen argued that automatic speech recognition (ASR) models often miss important keywords, which breaks the capture of full context. Alex Gay, CMO of meeting notetaker Otter, agreed and said his company keeps working on transcription, because flawed transcripts make every downstream action wrong and erode trust in the platform. Gay also flagged language support as an area needing improvement.

Both executives also raised transparency. Otter wants people in meetings to know when they are being recorded, even notifying everyone in the chat when its bot is not present, while Wen said it is important to establish that callers are speaking to an AI in enterprise calls. Gay noted Otter is also exploring digital twins that could represent people in meetings, but argued that without the emotive expression of a real conversation, an avatar is just a question-and-answer chatbot.