Hey everyone! I am looking to speak to folks who are building customer facing AI agents (e.g., support, outbound sales). Context: I spent last 6 months building phone agents for large publicly traded enterprises. The hardest part wasn't the LLM or voice quality - it was the AI agents stagnate or even regress as information becomes outdated. The only way for the agent not to keep making the same mistake / escalating the same request due to knowledge gap is not manually update the agents to see improvements. We built our internal system to do this and looking to productize this. Would love to get feedback from folks ❤️