'Voice AI is the quickest medium we have nowadays to communicate with technology.' Let's think about this quote from Jannis Moore (from the 7th Nov edition of The n8n Masterclass podcast: 'Used cases for deploying voice agents in SMB and enterprise'). Voice is certainly faster than typing (on laptops and on phones). Look at my Wispr Flow stats from the last 9 weeks. I have spoken 55k words into my computer and phone. Mostly interacting with Claude, Lovable and Supabase. But, on the other hand, reading is faster than listening. This is why we like the fact that our phones can transcribe voice mail or voice notes, so we don't have to spend 3 minutes listening. Voice input (speaking) also allows you to read, organise and digest information as you speak, which you cannot really do whilst typing, so there's a form of multi-tasking efficiency with voice input. For example you can interpret your notes and images as you speak. Your words then become a format that is more compatible with technology. Voice AI agents can have the same access to LLMs and business processes as chat AI agents. In many cases they will be a faster interface for getting what we need than typing or clicking. More people have access to phones than laptops more of the time. What are the use cases for this huge voice AI efficiency gain when applied at scale? Jannis Moore discusses four opportunities in this podcast: 1. Lead reactivation 2. Lead qualification 3. Customer support 4. Appointment setting I struggle with the first one. We don't pick up the phone to actual real people any more, sometimes people that we know and like, so I find it a real stretch to believe that we will pick up the phone to a voice agent. The podcast host Dylan Watkins does gently push back on this by asking Moore if there is value for the lead in this interaction. Lead qualification has some merit. If I'm the lead, I have an interest here so I might have enough to spend time with your agent to get to the information or opportunity I'm after. Customer Support has huge potential efficiencies and quality improvements through the use of Voice AI. By replacing IVR systems with Voice AI we can satisfy customers much more quickly and at far lower cost. Once we are confident in the guardrails and observability processes, agents can also have access to business processes via MCP servers and actually change things, transact and make customers lives better. As a customer I do not mind talking to an agent if they can actually help. I want the help. I don't want to talk to a person if an agent can solve my problem more quickly. I was surprised that privacy and security were not mentioned in the podcast. It is not possible to launch these services without a robust privacy and security framework. I am very much looking forward to trying out Voice AI in real business scenarios. Exciting times! The podcast can be found here (7th Nov edition) https://lnkd.in/eDZi9YVF