WEBVTT 00:00:00.000 --> 00:00:02.339 So for the past few years, we've mostly interacted 00:00:02.339 --> 00:00:06.360 with AI through, well, a text box, right? Typing 00:00:06.360 --> 00:00:08.460 things in, getting text back. But I think it's 00:00:08.460 --> 00:00:11.880 time to stop typing because the really big shift 00:00:11.880 --> 00:00:15.640 happening like right now is AI moving, moving 00:00:15.640 --> 00:00:19.059 its brain out into the real world, having actual 00:00:19.059 --> 00:00:22.359 natural phone conversations in real time. Exactly. 00:00:22.460 --> 00:00:23.920 And that's what this deep dive is all about. 00:00:24.000 --> 00:00:25.899 There's this huge automation opportunity opening 00:00:25.899 --> 00:00:27.980 up. We're talking about how you can build functional. 00:00:29.379 --> 00:00:32.100 autonomous AI assistance, like starting today, 00:00:32.259 --> 00:00:34.799 really capitalize on this market that, well, 00:00:34.859 --> 00:00:37.579 we think is going to explode by 2026. Yeah, that's 00:00:37.579 --> 00:00:39.539 the mission here. We want to take you from maybe 00:00:39.539 --> 00:00:42.119 knowing nothing about this to really understanding 00:00:42.119 --> 00:00:45.840 the infrastructure of a voice agent, a functional 00:00:45.840 --> 00:00:47.979 one. built with no -code tools. We'll kind of 00:00:47.979 --> 00:00:49.960 unpack its technical anatomy, look at the tools 00:00:49.960 --> 00:00:52.280 you need, specifically VAPI and NAN, and then 00:00:52.280 --> 00:00:54.359 walk through the actual steps to build your first 00:00:54.359 --> 00:00:56.479 one. It's a massive skill gap out there, honestly. 00:00:56.640 --> 00:00:58.939 Huge demand, very lucrative. So yeah, let's start 00:00:58.939 --> 00:01:00.399 with the basics, the fundamental architecture. 00:01:00.799 --> 00:01:03.979 Okay, so the first thing is, how do you actually 00:01:03.979 --> 00:01:07.200 give... a large language model you know that 00:01:07.200 --> 00:01:09.780 smart text brain we've all used how do you give 00:01:09.780 --> 00:01:11.760 it a mouth and ears well the basic definition 00:01:11.760 --> 00:01:13.920 is pretty simple a voice agent is essentially 00:01:13.920 --> 00:01:17.379 a virtual assistant one that can hold like complex 00:01:17.379 --> 00:01:19.439 natural sounding conversations over the phone 00:01:19.439 --> 00:01:22.239 or even the web think of it like a chat bot on 00:01:22.239 --> 00:01:25.099 the inside but the interaction layer that's all 00:01:25.099 --> 00:01:27.540 speech so this isn't the same as when you call 00:01:27.540 --> 00:01:31.099 a company and get that clunky press One for sales, 00:01:31.260 --> 00:01:33.319 press two for service kind of thing. The thing 00:01:33.319 --> 00:01:35.260 everyone hates. Oh, absolutely not. That's the 00:01:35.260 --> 00:01:37.920 key difference. These agents, they can make and 00:01:37.920 --> 00:01:41.000 receive calls 247. They can maintain a genuinely 00:01:41.000 --> 00:01:44.359 human -like conversational flow. And this is 00:01:44.359 --> 00:01:46.379 critical. They integrate with your business systems, 00:01:46.480 --> 00:01:48.560 your CRM, your calendar, whatever. Okay. And 00:01:48.560 --> 00:01:51.180 the way it works, it relies on three core pieces 00:01:51.180 --> 00:01:53.299 working together, right? You called it the LLM 00:01:53.299 --> 00:01:55.000 with ears and a mouth. That's pretty much it. 00:01:55.099 --> 00:01:58.709 The ear is step one, speech to text or STT. That 00:01:58.709 --> 00:02:00.769 takes what the person says, their spoken words, 00:02:00.930 --> 00:02:03.209 and turns it into written text the AI can understand. 00:02:03.829 --> 00:02:06.829 And that text then feeds into the brain, the 00:02:06.829 --> 00:02:09.009 LLM. That's where the AI figures out the right 00:02:09.009 --> 00:02:10.990 response, something conversational and nuanced. 00:02:11.370 --> 00:02:14.330 Yep. And then finally, the mouth. That's text 00:02:14.330 --> 00:02:18.080 -to -speech. TTS, it converts the AI's text answer 00:02:18.080 --> 00:02:21.219 back into really natural sounding spoken words 00:02:21.219 --> 00:02:23.319 for the person to hear. What's kind of interesting 00:02:23.319 --> 00:02:25.719 here is that the skills you might already have 00:02:25.719 --> 00:02:29.020 from working with text based LLMs like prompting, 00:02:29.020 --> 00:02:31.639 figuring out the logic, those still totally apply. 00:02:31.860 --> 00:02:33.500 It's like you already get the brain part. Exactly. 00:02:33.719 --> 00:02:35.759 Now we just need to layer on the logistics and 00:02:35.759 --> 00:02:39.060 crucially deal with this new factor, latency. 00:02:39.439 --> 00:02:41.819 Ah, right. Speed. Yeah. If you're building a 00:02:41.819 --> 00:02:43.979 chat bot, you know, a delay of a second or two 00:02:43.979 --> 00:02:47.419 might be OK. But with voice, even half a second 00:02:47.419 --> 00:02:49.960 delay feels weird. Like the agent's buffering 00:02:49.960 --> 00:02:52.800 or it's confused. Latency just destroys user 00:02:52.800 --> 00:02:55.479 trust in a real -time conversation. That's honestly 00:02:55.479 --> 00:02:57.379 why voice agents are fundamentally harder to 00:02:57.379 --> 00:02:59.199 get right than chatbots. That makes total sense. 00:02:59.280 --> 00:03:01.539 Voice really demands that immediate back -and 00:03:01.539 --> 00:03:04.020 -forth feeling. Yeah, speed is completely non 00:03:04.020 --> 00:03:05.900 -negotiable if you want people to actually... 00:03:06.139 --> 00:03:08.620 accept it and use it. Okay. So let's dig a bit 00:03:08.620 --> 00:03:11.460 deeper then into the structure, the anatomy of 00:03:11.460 --> 00:03:13.639 these agents. You're saying there are four essential 00:03:13.639 --> 00:03:16.330 parts we need to understand. Right. four key 00:03:16.330 --> 00:03:19.229 pieces. Part one is the LLM, the brain, that's 00:03:19.229 --> 00:03:21.650 the intelligence engine. And choosing the right 00:03:21.650 --> 00:03:24.870 model, maybe GPT -4, maybe clog, maybe something 00:03:24.870 --> 00:03:27.650 else optimized for speed, that choice is critical. 00:03:27.750 --> 00:03:29.310 You're always doing this balancing act between 00:03:29.310 --> 00:03:32.509 how smart it is, how sophisticated, versus that 00:03:32.509 --> 00:03:34.990 latency, the speed, and of course how much the 00:03:34.990 --> 00:03:37.680 call costs to run. And part two, you said this 00:03:37.680 --> 00:03:39.719 is maybe the most important part, the system 00:03:39.719 --> 00:03:42.099 prompt, the playbook. Definitely the most important, 00:03:42.199 --> 00:03:44.080 yeah. This is basically the agent's instruction 00:03:44.080 --> 00:03:46.360 manual. It defines everything. Its role, like 00:03:46.360 --> 00:03:48.780 you are a warm, patient customer support rep. 00:03:48.919 --> 00:03:51.979 Its personality, its style, and maybe most importantly, 00:03:52.099 --> 00:03:54.300 its rules and boundaries. This is where you explicitly 00:03:54.300 --> 00:03:57.360 state things like, under no circumstances give 00:03:57.360 --> 00:04:00.020 financial advice. You know, I have to admit something 00:04:00.020 --> 00:04:02.080 here. Even with all the time I've spent working 00:04:02.080 --> 00:04:05.469 with LLMs, Getting that prompt perfect on the 00:04:05.469 --> 00:04:08.110 first try. It's basically impossible. I still 00:04:08.110 --> 00:04:10.969 wrestle with prompt drift myself. Real conversation 00:04:10.969 --> 00:04:14.669 is just messy. You have to constantly test, tweak, 00:04:14.750 --> 00:04:16.829 iterate. That's honestly the hardest part of 00:04:16.829 --> 00:04:18.930 making these things work well in the real world. 00:04:19.149 --> 00:04:20.769 No, I appreciate you saying that. It's absolutely 00:04:20.769 --> 00:04:22.509 the truth when you actually deploy these things. 00:04:22.870 --> 00:04:25.410 Okay, so part three is the voice, the persona. 00:04:25.629 --> 00:04:28.350 This really impacts how users perceive the agent, 00:04:28.449 --> 00:04:30.980 whether they trust it. You can choose gender, 00:04:31.160 --> 00:04:35.560 age, accent from providers like 11 Labs. It brings 00:04:35.560 --> 00:04:38.060 the persona to life. And the last piece, number 00:04:38.060 --> 00:04:41.000 four, is the tools, the superpowers. This is 00:04:41.000 --> 00:04:43.199 the action layer, right? What lets it do more 00:04:43.199 --> 00:04:45.459 than just talk? Exactly. This is where it gets 00:04:45.459 --> 00:04:46.980 really powerful. We're talking about the agent 00:04:46.980 --> 00:04:49.220 being able to, say, view internal databases, 00:04:49.579 --> 00:04:52.160 create appointments directly in a calendar, process 00:04:52.160 --> 00:04:54.899 payments, or trigger literally any kind of automation 00:04:54.899 --> 00:04:58.009 through some connected service. So let's take 00:04:58.009 --> 00:05:01.149 that dentist office agent example. The brain 00:05:01.149 --> 00:05:04.610 might be a fast LLM like GPT -4. The playbook 00:05:04.610 --> 00:05:06.910 is super clear. You only schedule appointments, 00:05:07.269 --> 00:05:11.069 nothing else. The personas may be friendly, reassuring, 00:05:11.110 --> 00:05:13.889 and the superpowers. The superpowers would be 00:05:13.889 --> 00:05:16.470 check the dentist's calendar availability, book 00:05:16.470 --> 00:05:18.810 the appointment slot, and then maybe send a confirmation 00:05:18.810 --> 00:05:22.019 email or text. all done autonomously, instantly, 00:05:22.120 --> 00:05:24.279 during the call. And that system prompt, the 00:05:24.279 --> 00:05:26.060 playbook, is kind of like the agent's constitution. 00:05:26.379 --> 00:05:28.600 It sets all the rules. Exactly. It dictates everything 00:05:28.600 --> 00:05:30.839 it should and shouldn't do. Now, to build something 00:05:30.839 --> 00:05:33.139 professional like that, especially using no -code 00:05:33.139 --> 00:05:36.660 tools, we need specific platforms. The sources 00:05:36.660 --> 00:05:39.180 you looked at point towards a combination, VAPI 00:05:39.180 --> 00:05:41.939 and NAN. Yeah, VAPI is the specialized platform 00:05:41.939 --> 00:05:44.139 for the voice piece. It's designed for this. 00:05:44.199 --> 00:05:46.360 It handles all the complex voice logistics, the 00:05:46.360 --> 00:05:49.560 phone calls, the real -time STTTTS, managing... 00:05:49.610 --> 00:05:51.949 that whole conversational interface. Think of 00:05:51.949 --> 00:05:53.850 it as the front end for the conversation. Okay. 00:05:53.949 --> 00:05:57.629 Voppy for voice. And then N8n. N8n is this really 00:05:57.629 --> 00:06:01.310 robust visual automation platform. It's for building 00:06:01.310 --> 00:06:03.730 out all the backend logic, the workflows. And 00:06:03.730 --> 00:06:06.029 the key thing is it connects to pretty much any 00:06:06.029 --> 00:06:08.550 service or API you can think of. It's the integration 00:06:08.550 --> 00:06:11.730 engine. So why put them together? What's the 00:06:11.730 --> 00:06:14.589 magic in combining Voppy and N8n? Well, it's 00:06:14.589 --> 00:06:17.069 like peanut butter and jelly, really. VAPI handles 00:06:17.069 --> 00:06:19.529 the conversation, the talking and listening, 00:06:19.709 --> 00:06:22.930 the mouth and ears. But NANN gives it unlimited 00:06:22.930 --> 00:06:25.129 tools, the ability to actually do things. Ah, 00:06:25.290 --> 00:06:28.470 okay. So VAPI is like the interface managing 00:06:28.470 --> 00:06:31.410 the user experience of the call. And NANN is... 00:06:31.600 --> 00:06:33.579 Everything happening behind the scenes, the business 00:06:33.579 --> 00:06:36.079 logic connecting to other systems. That's a really 00:06:36.079 --> 00:06:38.540 good way to put it. VAPI chats and it enacts. 00:06:38.600 --> 00:06:40.379 So the agent goes from just answering a question 00:06:40.379 --> 00:06:43.060 like, what's your return policy to actually taking 00:06:43.060 --> 00:06:45.500 action? Maybe checking inventory in one system, 00:06:45.639 --> 00:06:48.100 logging the call details in Salesforce and setting 00:06:48.100 --> 00:06:50.339 a follow up text message all within that single 00:06:50.339 --> 00:06:53.199 phone call. Got it. VAPI for the natural dialogue 00:06:53.199 --> 00:06:55.899 and ANN to let the agent interact with the messy 00:06:55.899 --> 00:06:58.709 real world stuff. Precisely. That's the power 00:06:58.709 --> 00:07:01.290 combo, mid -roll sponsor read provided separately. 00:07:01.850 --> 00:07:04.189 Okay, so we've covered the how. Now let's look 00:07:04.189 --> 00:07:06.449 at the two main ways these agents typically operate. 00:07:06.629 --> 00:07:08.970 You've basically got inbound and outbound. Right. 00:07:09.069 --> 00:07:11.930 Inbound agents, that's like your virtual receptionist, 00:07:11.949 --> 00:07:15.209 available 24 -7. This is where someone calls 00:07:15.209 --> 00:07:17.589 the agent's dedicated number and the agent picks 00:07:17.589 --> 00:07:20.209 up. Common uses are, you know, round -the -clock 00:07:20.209 --> 00:07:23.579 customer support. answering complex FAQs, maybe 00:07:23.579 --> 00:07:26.279 checking on an order status. The key is the customer 00:07:26.279 --> 00:07:28.639 initiates the call. Then you flip it and you 00:07:28.639 --> 00:07:30.980 have outbound agents. Think of these as your 00:07:30.980 --> 00:07:33.379 autonomous assistant. In this case, the agent 00:07:33.379 --> 00:07:36.540 proactively calls people for you. This area has 00:07:36.540 --> 00:07:39.490 huge ROI potential. Things like sales follow 00:07:39.490 --> 00:07:42.769 -ups, payment reminders or collections, and a 00:07:42.769 --> 00:07:44.430 really big one, appointment reminders. Yeah. 00:07:44.490 --> 00:07:46.850 Those drastically cut down on costly no -shows. 00:07:46.949 --> 00:07:49.329 Here, the agent initiates. And there's that hybrid 00:07:49.329 --> 00:07:52.089 model too, right? The WebSo widget agent. Yeah, 00:07:52.149 --> 00:07:53.490 that's a neat one. It's like a little chat bubble 00:07:53.490 --> 00:07:55.850 on your website, but instead of typing, the user 00:07:55.850 --> 00:07:58.449 clicks a button and it starts an instant voice 00:07:58.449 --> 00:08:00.949 call right through their browser. Super convenient. 00:08:01.670 --> 00:08:04.490 Bypasses needing a phone number entirely. So 00:08:04.490 --> 00:08:06.310 if a small business is looking for the quickest 00:08:06.310 --> 00:08:09.750 win, the clearest ROI, which type usually delivers 00:08:09.750 --> 00:08:13.230 that? Generally, inbound support for off hours 00:08:13.230 --> 00:08:16.250 or overflow and outbound appointment reminders. 00:08:16.730 --> 00:08:19.470 Those tend to show immediate measurable financial 00:08:19.470 --> 00:08:21.889 value pretty quickly. We've laid out that the 00:08:21.889 --> 00:08:24.889 tech is here. It's functional. But maybe let's 00:08:24.889 --> 00:08:26.970 zoom out for a second. Why should someone listening 00:08:26.970 --> 00:08:29.949 right now really focus on learning about voice 00:08:29.949 --> 00:08:33.200 agents now? Because this whole space. Voice AI 00:08:33.200 --> 00:08:35.700 agents, it feels like it's right at the tipping 00:08:35.700 --> 00:08:38.279 point for massive mainstream adoption by businesses. 00:08:38.580 --> 00:08:41.080 Probably around 2026 is when we'll see it everywhere. 00:08:41.159 --> 00:08:43.879 It's not some far -off future thing. The technical 00:08:43.879 --> 00:08:46.039 hurdles, especially latency, they've largely 00:08:46.039 --> 00:08:47.820 been overcome recently. We're at an inflection 00:08:47.820 --> 00:08:50.600 point. And the value proposition for businesses 00:08:50.600 --> 00:08:53.080 seems incredibly straightforward. It's easy for 00:08:53.080 --> 00:08:56.100 them to grasp. You save potentially a lot of 00:08:56.100 --> 00:08:59.200 money like that $50 receptionist salary example 00:08:59.200 --> 00:09:02.519 while also getting 24 -7 service. It's just pure 00:09:02.519 --> 00:09:05.460 efficiency. Absolutely. And that clear value 00:09:05.460 --> 00:09:07.799 creates this perfect storm situation for anyone 00:09:07.799 --> 00:09:10.340 learning this now. You have extremely high demand 00:09:10.340 --> 00:09:12.600 from businesses desperate to automate things 00:09:12.600 --> 00:09:15.379 like call centers. But you have very low competition 00:09:15.379 --> 00:09:17.720 because, frankly, not many people know how to. 00:09:17.740 --> 00:09:19.779 build these things properly yet using the right 00:09:19.779 --> 00:09:22.580 tools like VAPI and A8N. It really gets interesting 00:09:22.580 --> 00:09:24.299 when you think about the kinds of problems this 00:09:24.299 --> 00:09:26.100 solves. These aren't trivial things. They're 00:09:26.100 --> 00:09:29.799 real, expensive, often painful business bottlenecks, 00:09:29.820 --> 00:09:32.000 scheduling nightmares, rising customer service 00:09:32.000 --> 00:09:35.080 costs. Whoa. Yeah, just imagine scaling that 00:09:35.080 --> 00:09:37.639 one agent. You build it once, refine it, and 00:09:37.639 --> 00:09:39.679 suddenly it can handle, I don't know, a billion 00:09:39.679 --> 00:09:42.500 routine customer interactions a year. Perfectly. 00:09:42.500 --> 00:09:44.840 Every time. No sick days, no vacations. Being 00:09:44.840 --> 00:09:46.620 able to build solutions for those kinds of expensive 00:09:46.620 --> 00:09:49.519 problems, that positions you as an expert really 00:09:49.519 --> 00:09:51.639 early on in a field that's about to take off. 00:09:51.860 --> 00:09:53.919 Okay, so if the opportunity is really that big, 00:09:54.080 --> 00:09:56.539 what's the main hurdle? What's stopping more 00:09:56.539 --> 00:09:58.500 people from jumping in and building these right 00:09:58.500 --> 00:10:01.500 now? Honestly, it mostly comes down to the lack 00:10:01.500 --> 00:10:04.039 of specialized knowledge. People aren't familiar 00:10:04.039 --> 00:10:06.320 with the specific tools and techniques needed 00:10:06.320 --> 00:10:09.340 to connect that conversational part, the voice, 00:10:09.480 --> 00:10:13.080 to the action part. the back end systems so let's 00:10:13.080 --> 00:10:15.279 try and fix that right now let's get practical 00:10:15.279 --> 00:10:17.639 how about we walk through building a very basic 00:10:17.639 --> 00:10:20.679 customer support agent one that just uses a knowledge 00:10:20.679 --> 00:10:24.059 base all inside boppy okay sounds good so step 00:10:24.059 --> 00:10:27.120 one and two are just set up right Get a VAPI 00:10:27.120 --> 00:10:29.240 account, grab the free credits they offer for 00:10:29.240 --> 00:10:31.419 testing, then just poke around the default agent 00:10:31.419 --> 00:10:33.500 settings, see where you set the LLM model, the 00:10:33.500 --> 00:10:36.080 first message the caller hears. Exactly. And 00:10:36.080 --> 00:10:37.919 pay close attention to the system prompt structure 00:10:37.919 --> 00:10:40.379 in that default agent. You'll see those key parts 00:10:40.379 --> 00:10:43.059 we talked about. The role, the identity, the 00:10:43.059 --> 00:10:45.279 voice characteristics, and that step -by -step 00:10:45.279 --> 00:10:48.799 conversational flow. That flow often uses a kind 00:10:48.799 --> 00:10:51.320 of if this, then that logic to try and guide 00:10:51.320 --> 00:10:53.179 the conversation, keep it on track, even when 00:10:53.179 --> 00:10:55.759 the user says unexpected things. Then step three, 00:10:55.820 --> 00:10:58.700 this is where it gets useful. Adding a knowledge 00:10:58.700 --> 00:11:01.940 base, like the agent's cheat sheet. Yep. Inside 00:11:01.940 --> 00:11:04.179 Boppy, there's usually a file section. You just 00:11:04.179 --> 00:11:07.500 upload your company documents, FAQs, PDFs, text 00:11:07.500 --> 00:11:09.919 files with product info, warranty details, whatever. 00:11:10.279 --> 00:11:12.600 Then you connect those files to your agent. And 00:11:12.600 --> 00:11:15.159 just like that, the agent can now reference that 00:11:15.159 --> 00:11:18.279 specific information during a call. Like if someone 00:11:18.279 --> 00:11:21.019 asks about the shipping policy. Instantly. It 00:11:21.019 --> 00:11:23.840 basically gives the agent an internal search 00:11:23.840 --> 00:11:27.039 engine for your company info. So instead of putting 00:11:27.039 --> 00:11:28.919 the caller on hold while a human looks it up, 00:11:29.019 --> 00:11:32.039 the AI can find and relay the info right away. 00:11:32.220 --> 00:11:35.029 Okay. Step four is obviously testing. talking 00:11:35.029 --> 00:11:37.730 to it absolutely crucial use the talk to assistant 00:11:37.730 --> 00:11:40.090 button voppy provides have a real conversation 00:11:40.090 --> 00:11:42.570 and importantly watch the live transcription 00:11:42.570 --> 00:11:45.090 logs as you talk you can literally see the agent 00:11:45.090 --> 00:11:47.629 thinking see when it decides to query that knowledge 00:11:47.629 --> 00:11:49.990 -based tool to find an answer and step five which 00:11:49.990 --> 00:11:51.629 you mentioned is where the real work happens 00:11:51.629 --> 00:11:56.289 iterating on the prompt yes 90 of the value comes 00:11:56.289 --> 00:11:59.490 from this loop the first version will rarely 00:11:59.490 --> 00:12:03.139 be perfect maybe it sounds too robotic You go 00:12:03.139 --> 00:12:05.220 back to the prompt, add instructions like be 00:12:05.220 --> 00:12:07.500 friendly and conversational. Maybe it talks too 00:12:07.500 --> 00:12:09.860 much. Add keep your answers concise and direct. 00:12:10.159 --> 00:12:13.120 So it's this constant cycle. Test it. See what's 00:12:13.120 --> 00:12:16.039 wrong. Document the issues. Refine the system 00:12:16.039 --> 00:12:18.580 prompt and test it again. That's where the human 00:12:18.580 --> 00:12:21.679 really shapes the AI's performance. Exactly. 00:12:21.679 --> 00:12:23.600 Now, the agent we just described, it can talk. 00:12:23.639 --> 00:12:25.500 It can reference files from the knowledge base. 00:12:26.440 --> 00:12:28.620 But it still has one big limitation at this stage. 00:12:28.720 --> 00:12:30.460 Right. It can't actually do anything yet, can 00:12:30.460 --> 00:12:33.000 it? It can't take real action in the world. Precisely. 00:12:33.000 --> 00:12:34.720 It's still just talking and looking things up. 00:12:34.970 --> 00:12:37.210 Okay, so that kind of brings us back to the start, 00:12:37.289 --> 00:12:39.990 but with a tangible result. We've outlined how 00:12:39.990 --> 00:12:42.669 to build a functional AI receptionist, one that 00:12:42.669 --> 00:12:45.330 can answer calls and use company info, all without 00:12:45.330 --> 00:12:48.110 writing code. We went beyond just the LLM as 00:12:48.110 --> 00:12:50.309 a text box. We looked at its anatomy, the brain, 00:12:50.450 --> 00:12:52.649 the playbook, the persona, and the potential 00:12:52.649 --> 00:12:55.549 for superpowers using tools. Yeah, we proved 00:12:55.549 --> 00:12:57.409 it can handle answering calls based on documents. 00:12:57.870 --> 00:13:01.370 But the real game changer, the automation powerhouse. 00:13:01.950 --> 00:13:04.149 That gets unlocked when we give it those superpowers. 00:13:04.250 --> 00:13:06.950 Yeah. Likely using something like N8N. Just think 00:13:06.950 --> 00:13:09.029 about that difference for a moment. An agent 00:13:09.029 --> 00:13:11.710 that can only talk versus one that can all on 00:13:11.710 --> 00:13:14.750 the same call, simultaneously check a real -time 00:13:14.750 --> 00:13:17.269 calendar, book the appointment right into the 00:13:17.269 --> 00:13:20.490 system, log the interaction in a CRM or spreadsheet, 00:13:20.809 --> 00:13:23.769 and send out the confirmation email, all while 00:13:23.769 --> 00:13:26.440 the person is still on the line. That's the moment 00:13:26.440 --> 00:13:28.720 the agent transforms from just an answering machine 00:13:28.720 --> 00:13:31.779 into a truly autonomous employee. And that is 00:13:31.779 --> 00:13:34.399 the enormous opportunity sitting there for anyone 00:13:34.399 --> 00:13:36.919 who decides to learn these tools right now. So 00:13:36.919 --> 00:13:38.620 we really encourage you to take what we've discussed 00:13:38.620 --> 00:13:40.720 today, start exploring, and see the potential 00:13:40.720 --> 00:13:42.899 of voice automation for yourself. Absolutely. 00:13:43.000 --> 00:13:46.100 Until the next deep dive, keep learning, keep 00:13:46.100 --> 00:13:46.460 building.