1 00:00:00,160 --> 00:00:02,240 Today on the Winning with Shopify podcast, it's. 2 00:00:02,760 --> 00:00:05,680 Beyond just putting me in a box that says hey I need an agent to 3 00:00:05,680 --> 00:00:08,280 do this for me. I just keep clicking stuff until 4 00:00:08,280 --> 00:00:11,720 it says speak to human. We don't love telling people the 5 00:00:11,720 --> 00:00:13,280 same thing over and over and over again. 6 00:00:13,320 --> 00:00:16,280 It has nothing to do with saving money, nothing to do with 7 00:00:16,280 --> 00:00:20,840 alleviating workload on agents. It's all about speed and 8 00:00:20,840 --> 00:00:23,120 offering a level of support that they otherwise. 9 00:00:23,120 --> 00:00:24,720 Couldn't this has to be a chat box? 10 00:00:24,720 --> 00:00:27,800 It's like first name and you're like Nick and then click send 11 00:00:27,800 --> 00:00:29,480 last name. You know that was too fast. 12 00:00:34,640 --> 00:00:37,080 Hello and welcome to another episode of the winning Shop of 13 00:00:37,080 --> 00:00:38,720 Fight podcast. For anyone that's not tuned in 14 00:00:38,720 --> 00:00:40,720 before, my name is Nick. I'm here in the British Isles, 15 00:00:40,880 --> 00:00:43,840 hence my very posh accent as you guys over in the States Tell me 16 00:00:44,160 --> 00:00:46,000 if you have been tuning in for a while and you haven't already, 17 00:00:46,000 --> 00:00:48,440 then please do hit subscribe on whatever channel you're on. 18 00:00:48,680 --> 00:00:51,520 Give us a like and also check out our TikTok and Instagram 19 00:00:51,520 --> 00:00:53,080 channels. We have been posting quite a lot 20 00:00:53,080 --> 00:00:54,680 on there and I've just checked my notes. 21 00:00:54,880 --> 00:00:57,560 Elliot's going to shoot me if I don't flag our YouTube channel 22 00:00:57,560 --> 00:00:58,960 as well. He's also been uploading lots of 23 00:00:58,960 --> 00:01:00,480 little shorts from previous previous episode. 24 00:01:00,480 --> 00:01:02,560 So if you love these long episodes, but you want to get 25 00:01:02,560 --> 00:01:04,480 some more bite sized content, then go and check out our 26 00:01:04,480 --> 00:01:07,160 Instagram, our TikTok and our YouTube channel. 27 00:01:07,400 --> 00:01:09,480 If you haven't already, check out the growth hub that is a 28 00:01:09,560 --> 00:01:10,960 free community of Shopify brands. 29 00:01:11,080 --> 00:01:14,280 We also have a few absolute legends from non Shopify brands 30 00:01:14,280 --> 00:01:17,160 that have come in there as well, and they are more than welcome. 31 00:01:17,160 --> 00:01:19,360 So if you're using WooCommerce magenta, anything else and 32 00:01:19,360 --> 00:01:21,040 you're sitting there listening to this podcast going, I'm 33 00:01:21,040 --> 00:01:23,040 crying inside because I want to be using Shopify. 34 00:01:23,320 --> 00:01:24,600 It's fine. It's a safe space. 35 00:01:24,800 --> 00:01:27,680 Come and chat with us. Now today you've probably seen 36 00:01:27,680 --> 00:01:29,440 in the title, we're talking about something you've probably 37 00:01:29,440 --> 00:01:31,120 never heard heard of before, which is AI. 38 00:01:31,400 --> 00:01:34,280 Obviously I'm joking, everyone knows what AI is and or 39 00:01:34,280 --> 00:01:36,480 certainly know a little bit about it, but we want to know a 40 00:01:36,480 --> 00:01:37,840 little bit more. So today we're going to be 41 00:01:37,840 --> 00:01:39,080 talking about AI and customer service. 42 00:01:39,080 --> 00:01:40,560 And joining me is a very special guest. 43 00:01:40,560 --> 00:01:42,200 His name is Jacob from a company called Tidio. 44 00:01:42,440 --> 00:01:44,120 Jacob, welcome to the show. How are you? 45 00:01:44,800 --> 00:01:46,240 I'm doing well, thanks for having me. 46 00:01:47,040 --> 00:01:49,040 Well, it's great to have you on the show, and as we do with all 47 00:01:49,040 --> 00:01:51,240 of our guests, when we first start off each episode, tell us 48 00:01:51,240 --> 00:01:54,160 a little bit about your background, tell us about Tidio, 49 00:01:54,520 --> 00:01:56,880 and then we'll crack on talking about some of the features of 50 00:01:56,880 --> 00:01:59,400 it, things it does, and how it can help some of our amazing 51 00:01:59,400 --> 00:02:02,080 listeners at home. Yeah, yeah, 100%. 52 00:02:02,080 --> 00:02:06,400 So I am an account executive at Tidio and I've been working in 53 00:02:07,080 --> 00:02:10,560 SAS sales, started as an SDR, kind of worked my way up into 54 00:02:10,680 --> 00:02:12,840 inside sales and then the AE position. 55 00:02:12,840 --> 00:02:16,560 So I'm I'm familiar with the space, but I've been working at 56 00:02:16,560 --> 00:02:20,760 Tidio, started as an AE, and I'm entirely on the inbound team. 57 00:02:20,800 --> 00:02:25,760 So 100% of my work day is spent meeting with businesses that are 58 00:02:25,760 --> 00:02:29,880 somewhat interested in some aspect of AI for their customer 59 00:02:29,880 --> 00:02:32,880 service and customer experience. Which is everyone right now, 60 00:02:32,920 --> 00:02:34,360 isn't it like. At this point in AI. 61 00:02:35,040 --> 00:02:35,960 Yep. Surely. 62 00:02:36,240 --> 00:02:38,400 Tell us a bit about Tidio then. What does Tidio actually do? 63 00:02:38,400 --> 00:02:39,760 We mentioned AI, mentioned customer service. 64 00:02:39,760 --> 00:02:42,280 Like if somebody's like, OK, I'm going to use Tidio, what are 65 00:02:42,280 --> 00:02:43,160 they actually getting from you guys? 66 00:02:43,160 --> 00:02:44,320 What service? What platform? 67 00:02:44,960 --> 00:02:46,560 Yeah, yeah. So it's, it's there's kind of 68 00:02:46,560 --> 00:02:48,880 three major pillars to the platform. 69 00:02:48,880 --> 00:02:50,640 It's a customer experience platform. 70 00:02:50,680 --> 00:02:54,680 We started 12 years ago as all live chat and ticketing. 71 00:02:54,720 --> 00:02:58,000 So in the early days, everything was human to human oriented, you 72 00:02:58,000 --> 00:02:59,080 could say. And that's kind of the 73 00:02:59,080 --> 00:03:01,240 foundation that we've been building upon. 74 00:03:02,000 --> 00:03:05,520 And then shortly after we got into the automation sphere and 75 00:03:05,520 --> 00:03:08,520 kind of built that on top of the help desk functionalities we 76 00:03:08,520 --> 00:03:11,520 have starting with flows. So still relatively simple, 77 00:03:11,520 --> 00:03:13,120 right? Decision trees you're you're 78 00:03:13,760 --> 00:03:17,240 your true chat bot. You could say a flow is all all 79 00:03:17,440 --> 00:03:20,640 if then decision making. And then most recently, really 80 00:03:20,640 --> 00:03:24,120 in the last yeah, two years, it's it's been a major focus on 81 00:03:24,120 --> 00:03:27,480 the conversational AI component, which is way different from 82 00:03:28,280 --> 00:03:31,520 these flow style chat bot automations because it's it's 83 00:03:31,520 --> 00:03:33,760 not pre scripted. I'm not saying what the next 84 00:03:33,760 --> 00:03:37,640 step is or connecting stages, right? 85 00:03:38,360 --> 00:03:40,400 And so the bit you're saying it's not is when you like you're 86 00:03:40,400 --> 00:03:43,080 talking to a chat bot, a live chat pops up and it says which 87 00:03:43,080 --> 00:03:45,240 of the four options you most interesting, like track my 88 00:03:45,280 --> 00:03:47,000 order, like get a refund, etcetera. 89 00:03:47,760 --> 00:03:49,680 Both sides right the the customer experience is 90 00:03:49,680 --> 00:03:52,240 completely different because it's not button driven and 91 00:03:52,240 --> 00:03:55,520 you're not dragging me through a series of steps where I can't 92 00:03:55,520 --> 00:03:58,560 see the end. But then also configuring it is 93 00:03:58,560 --> 00:04:01,040 completely different because I don't need to map out all of 94 00:04:01,040 --> 00:04:04,520 those steps. It's it's operating based off 95 00:04:04,520 --> 00:04:06,080 the data sources that I teach it. 96 00:04:06,080 --> 00:04:08,680 So I kind of, I control its brain and that's really the 97 00:04:08,680 --> 00:04:11,440 extent of how I manipulate the AI. 98 00:04:11,560 --> 00:04:14,160 So yeah. Which I think is really 99 00:04:14,160 --> 00:04:16,920 interesting because so many people, I think I read a study 100 00:04:16,920 --> 00:04:18,600 on this recently, I can't remember the number, but I 101 00:04:18,600 --> 00:04:20,959 remember being impressed by the how high the percentage was. 102 00:04:20,959 --> 00:04:23,560 The people that were like, I just keep clicking stuff until 103 00:04:23,560 --> 00:04:27,440 it says speak to human. And then I read the stats at the 104 00:04:27,440 --> 00:04:29,520 bottom of that saying. The irony is as well, all the 105 00:04:29,520 --> 00:04:31,720 people that were using that answered the survey were using 106 00:04:31,720 --> 00:04:33,640 the same platform. And about 40% of them actually 107 00:04:33,640 --> 00:04:36,280 were still talking to AI when they thought there was a human. 108 00:04:36,400 --> 00:04:38,000 Which we're going to talk about a little bit as we go through 109 00:04:38,000 --> 00:04:40,640 and kind of the ethics, the morals, and also the best 110 00:04:40,640 --> 00:04:42,920 approach to that. We're going to go back in in a 111 00:04:43,040 --> 00:04:46,600 bit more detail later, but how let's start off with like how 112 00:04:46,640 --> 00:04:49,120 AI, right? How is it transforming customer 113 00:04:49,120 --> 00:04:50,760 service for businesses? Like what are some of the 114 00:04:51,000 --> 00:04:54,360 fundamental things it's changing apart from, you know, you might, 115 00:04:54,360 --> 00:04:56,080 you might tell some of the obvious bits, but what are what 116 00:04:56,080 --> 00:04:58,880 are the less obvious things than just saving time, saving money, 117 00:04:58,880 --> 00:05:03,200 less humans on the ground? Yeah, I think a a big thing that 118 00:05:03,200 --> 00:05:06,920 I see from larger businesses, at least this is what they claim. 119 00:05:06,920 --> 00:05:10,640 It has nothing to do with saving money, nothing to do with 120 00:05:10,640 --> 00:05:15,760 alleviating workload on agents and it's all about speed and 121 00:05:15,760 --> 00:05:18,480 offering a level of support that they otherwise couldn't. 122 00:05:19,240 --> 00:05:23,320 So I think that's a, a big thing is just the, the pace at which 123 00:05:23,320 --> 00:05:25,880 support operations are working at now. 124 00:05:26,200 --> 00:05:28,600 And they've, you know, we've kind of shifted from using 125 00:05:28,600 --> 00:05:31,920 automations think like chat bots, whereas really our 126 00:05:31,920 --> 00:05:34,160 expectation was just that they're going to help us sort 127 00:05:34,160 --> 00:05:36,480 out the volume, right? Things like a main menu, they 128 00:05:36,480 --> 00:05:39,560 take everyone in and then they help compartmentalize all these 129 00:05:39,560 --> 00:05:41,440 requests so that we can work through them faster. 130 00:05:41,520 --> 00:05:46,600 And now we're way beyond just sorting and organizing these 131 00:05:46,600 --> 00:05:48,160 requests so that humans can answer them. 132 00:05:48,160 --> 00:05:50,720 We're, we are answering them, right? 133 00:05:50,720 --> 00:05:55,160 We're trying to resolve tickets before they become tickets in 134 00:05:55,160 --> 00:05:58,160 the 1st place. So, yeah, maybe that's the more 135 00:05:58,160 --> 00:06:02,000 obvious the Level 0, but I think something maybe less talked 136 00:06:02,000 --> 00:06:05,120 about, it's just the general speed and customer expectations 137 00:06:05,120 --> 00:06:08,480 that go with that. Which I think is amazing because 138 00:06:09,000 --> 00:06:11,000 speed is not what I thought you were going to say, but I think 139 00:06:11,000 --> 00:06:13,160 it's such a valid point. It's such a logical point when 140 00:06:13,160 --> 00:07:00,360 you think about it. And certainly if you think about 141 00:07:00,360 --> 00:07:03,160 like, I don't know, big store, even if you had 1000 staff on 142 00:07:03,160 --> 00:07:06,920 the ground, but you had like 23000 requests coming in every 143 00:07:06,920 --> 00:07:10,120 20 minutes, every 30 minutes from from customers, which for 144 00:07:10,120 --> 00:07:12,560 some large brands would, would absolutely happen. 145 00:07:13,240 --> 00:07:16,000 But that speed being able to go right, 30% of them, we just 146 00:07:16,000 --> 00:07:18,880 dealt with it quickly. Or actually this, and I guess 147 00:07:19,760 --> 00:07:21,640 you kind of hinted at this. I don't know if this is kind of 148 00:07:21,640 --> 00:07:24,160 where, where you were going to go with this, but you could 149 00:07:24,160 --> 00:07:27,920 almost have like 1 department specializes in order and like 150 00:07:27,920 --> 00:07:30,320 confirmation emails, for example, if someone gets in 151 00:07:30,320 --> 00:07:32,600 touch to go, I just want to confirm like I've just ordered. 152 00:07:32,600 --> 00:07:34,280 Can I confirm my address or whatever. 153 00:07:34,520 --> 00:07:37,240 Having a team that's like we're logged into the order management 154 00:07:37,240 --> 00:07:39,480 system so we can see those details and change those 155 00:07:39,480 --> 00:07:41,280 quickly. We're logged into the return 156 00:07:41,280 --> 00:07:43,520 system over here. Means the chat box getting you 157 00:07:43,520 --> 00:07:45,480 to the right team, and the right team's still doing the job, 158 00:07:45,480 --> 00:07:47,600 isn't it? As opposed to AI replacing the 159 00:07:47,600 --> 00:07:49,520 humans as you said? Sure. 160 00:07:49,640 --> 00:07:52,360 Yeah, yeah, no, 100%. And then even doing some of 161 00:07:52,360 --> 00:07:55,680 those things for the agents, which is kind of at the 162 00:07:55,680 --> 00:07:58,840 forefront of what we're at now. It's beyond just putting me in a 163 00:07:58,840 --> 00:08:01,320 box that says, hey, I need an agent to do this for me. 164 00:08:01,320 --> 00:08:05,520 And it's hey, we've already done this for the customer, now we 165 00:08:05,520 --> 00:08:08,120 need the agent to do this final, last step. 166 00:08:08,200 --> 00:08:11,520 So yeah, it's just taking over more and more of the process, 167 00:08:11,560 --> 00:08:15,200 which isn't a bad thing, right? If you speak with people who are 168 00:08:15,200 --> 00:08:18,240 on the front lines, actually answering the support questions, 169 00:08:19,320 --> 00:08:23,320 we don't love telling people the same thing over and over and 170 00:08:23,320 --> 00:08:26,480 over again. So yeah, it's in everyone's best 171 00:08:26,480 --> 00:08:29,480 interest for the most part. Yeah, I think it's amazing. 172 00:08:29,480 --> 00:08:32,000 And also like I was talking to a friend of mine who works for a 173 00:08:32,000 --> 00:08:34,919 bank in their customer service team and I say what's one of the 174 00:08:34,919 --> 00:08:36,280 best bits of tech you've got there? 175 00:08:36,720 --> 00:08:40,600 And he he said my workload literally became 50% easier, 176 00:08:40,600 --> 00:08:43,159 which meant I could give customers more attention. 177 00:08:43,520 --> 00:08:46,400 When each customer comes in, they've already had 7 or 8 178 00:08:46,400 --> 00:08:48,720 messages backwards and forwards of what they think is me. 179 00:08:49,200 --> 00:08:51,120 But I'm actually just asking what's your date of birth, your 180 00:08:51,120 --> 00:08:54,360 name, your account number? And actually now I know who you 181 00:08:54,400 --> 00:08:55,560 are. So when I get to you in the 182 00:08:55,560 --> 00:08:57,560 queue and I've got, he said, I'll have three conversations, 183 00:08:57,560 --> 00:09:00,200 open a live chat, anyone time. All their information is just 184 00:09:00,200 --> 00:09:02,120 there. And I'm just like, hey, how can 185 00:09:02,120 --> 00:09:04,200 I help? And the AI bot will even, he 186 00:09:04,200 --> 00:09:06,560 said it will even summarize like what their main thing was. 187 00:09:06,800 --> 00:09:09,200 So if they said like, oh, I'm really unhappy and then my wife 188 00:09:09,200 --> 00:09:10,800 had this and then this happened to me. 189 00:09:11,040 --> 00:09:13,640 The AI bot will summarize all of that and say like, we keep 190 00:09:13,640 --> 00:09:16,520 losing our debit cards. Then the agent goes, would you 191 00:09:16,520 --> 00:09:17,560 like me to send you a new debit card? 192 00:09:17,640 --> 00:09:20,120 And it's like you just so that's his first interaction. 193 00:09:20,120 --> 00:09:22,440 But they've had actually quite a nice polite conversation and you 194 00:09:22,440 --> 00:09:24,080 say speed. It's been really quick. 195 00:09:24,560 --> 00:09:27,160 But I, he did tell me that with the speed, he was trying not to 196 00:09:27,160 --> 00:09:30,840 let the chat bot reply too quickly so that you were like, 197 00:09:30,880 --> 00:09:33,400 this has to be a chat box. It's like first name and you're 198 00:09:33,400 --> 00:09:36,200 like Nick and then click send last name and you're like, that 199 00:09:36,200 --> 00:09:38,640 was too fast. Like creating a delay in there. 200 00:09:39,080 --> 00:09:40,880 How much do you guys find? You have to do that with the 201 00:09:40,880 --> 00:09:43,520 tech as well, so that when it is sending those kind of initial 202 00:09:43,520 --> 00:09:46,080 messages or when you've got this, you have to to set delays 203 00:09:46,080 --> 00:09:47,400 just so the customer's expectations. 204 00:09:47,400 --> 00:09:49,520 Like it's quick, but it's not too quick. 205 00:09:49,520 --> 00:09:52,400 And there's still been no human usage for this stuff. 206 00:09:52,400 --> 00:09:55,320 Actually, that's quite mundane. No, I would say that's extremely 207 00:09:55,320 --> 00:09:58,240 relevant and those are the things that like a lot of the 208 00:09:58,240 --> 00:10:00,760 clients I meet with, they don't even think about that kind of 209 00:10:00,760 --> 00:10:03,840 stuff. Is that being too fast through 210 00:10:03,840 --> 00:10:09,120 an automation could be perceived not so well, which like I think 211 00:10:09,160 --> 00:10:12,000 all of that is going to change long term, but we are still 212 00:10:12,000 --> 00:10:14,680 absolutely dealing with that right now. 213 00:10:14,680 --> 00:10:19,440 Where if end users don't feel that it's human like that, then 214 00:10:19,440 --> 00:10:22,560 they're those same visitors who are in the chat saying human, 215 00:10:22,560 --> 00:10:26,200 human, human, human, right. So no, it's, it's very 216 00:10:26,200 --> 00:10:27,240 important. Yeah, yeah. 217 00:10:27,240 --> 00:10:30,240 And overlooked many times. Absolutely. 218 00:10:30,240 --> 00:10:32,960 And yeah, I can imagine their confidence lost when it's not 219 00:10:32,960 --> 00:10:34,880 done right. And so, I mean, we're kind of 220 00:10:34,880 --> 00:10:37,280 hinting at this anyway. Would would you say that 221 00:10:37,280 --> 00:10:40,840 actually AI is enabling humans to do their jobs better? 222 00:10:40,840 --> 00:10:42,960 Or is actually it starting to replace humans? 223 00:10:42,960 --> 00:10:48,360 Or is it a murky mix of the two? I would say it's a non murky mix 224 00:10:48,360 --> 00:10:50,920 of the two. Like it's and and yeah, I think 225 00:10:51,880 --> 00:10:54,560 I work in sales so I need to know my audience most of the 226 00:10:54,560 --> 00:10:57,400 time. And a lot of times you might 227 00:10:57,400 --> 00:11:00,560 have customer support directors and things like that where they 228 00:11:00,560 --> 00:11:04,520 don't want to hear anything about really replacing human 229 00:11:04,560 --> 00:11:06,960 agents, right. But then we might talk to the. 230 00:11:07,600 --> 00:11:10,640 Finance they have. Finance or even like these 231 00:11:10,640 --> 00:11:14,200 bigger organizations, they have teams that are like director of 232 00:11:14,200 --> 00:11:17,200 process improvement. And so look, that's all they 233 00:11:17,200 --> 00:11:20,680 want to talk about is how can we improve process. 234 00:11:20,680 --> 00:11:23,440 And if that's replacing people, that just means we're going to 235 00:11:23,440 --> 00:11:25,280 use those people to do something more valuable. 236 00:11:25,280 --> 00:11:27,920 Doesn't mean we're just going in sweeping through the 237 00:11:27,920 --> 00:11:29,600 organization and firing everyone, right. 238 00:11:30,440 --> 00:11:35,040 But I think that, yeah, as far as helping human agents, we 239 00:11:35,040 --> 00:11:38,800 think about things like Copilot where it's the AI is already 240 00:11:38,800 --> 00:11:43,800 tapped into the conversation and I'm not having the AI take over 241 00:11:43,800 --> 00:11:46,040 the conversation entirely. But there are definitely 242 00:11:46,040 --> 00:11:49,360 instances where I can say, wow, this is what the AI would have 243 00:11:49,360 --> 00:11:51,720 said to this question that the customer just asked me. 244 00:11:51,720 --> 00:11:54,120 I could sit here and retype this whole thing, or I could just 245 00:11:54,120 --> 00:11:56,360 edit or send what the AI has already typed. 246 00:11:56,360 --> 00:12:00,560 So in many ways, speed, like it's, it's helping the agents 247 00:12:00,560 --> 00:12:03,400 work faster. It's helping the agents avoid 248 00:12:04,280 --> 00:12:06,320 those questions that make their eyes roll. 249 00:12:06,400 --> 00:12:08,080 You know, and when you're dealing, when you're a live 250 00:12:08,080 --> 00:12:10,960 agent and you're handling, you know, maybe 100 chats, 50 plus 251 00:12:10,960 --> 00:12:15,360 chats a day, you have plenty of those eye roll moments. 252 00:12:15,360 --> 00:12:18,240 It's those questions that are on the website made like there are 253 00:12:18,240 --> 00:12:21,000 things that are on the homepage that visitors will ask your live 254 00:12:21,000 --> 00:12:23,000 agents. So I think in that respect, we 255 00:12:23,000 --> 00:12:27,400 clean that up. It is just completely false 256 00:12:27,440 --> 00:12:30,920 though to say that like AI is not taking anyone's job. 257 00:12:31,680 --> 00:12:34,560 And this isn't just in the respect of support, right? 258 00:12:34,560 --> 00:12:38,680 If, if I have a team of 100 people with 0 automation in 259 00:12:38,680 --> 00:12:44,360 place and in a matter of 30 to 60 days, I can have 50% of all 260 00:12:44,360 --> 00:12:48,000 of our previous chats resolved in a couple of you know, back 261 00:12:48,000 --> 00:12:52,280 and forth messages with the AII need my human agents who were 262 00:12:52,280 --> 00:12:55,600 responsible for that 50% and I need to give them something 263 00:12:55,600 --> 00:12:58,360 else. So it's yeah, not that they're 264 00:12:58,360 --> 00:13:00,640 disappearing, but like Tedio went through this, we went 265 00:13:00,640 --> 00:13:02,680 through this ourselves. We didn't used to have AI 266 00:13:02,680 --> 00:13:05,200 support in our chat. We have full live team of human 267 00:13:05,200 --> 00:13:08,200 agents and we still do. But those live agents are not 268 00:13:08,200 --> 00:13:11,280 working on level 0, they're working on level 1, Level 2, 269 00:13:11,280 --> 00:13:15,400 They're working on multi day tickets that require external 270 00:13:15,400 --> 00:13:17,720 teams from the support team to help with that kind of stuff. 271 00:13:17,800 --> 00:13:21,280 So everything's shifting up. I mean, you see this in coding, 272 00:13:21,400 --> 00:13:22,880 You see this everywhere. So yeah. 273 00:13:22,880 --> 00:13:24,360 Yeah, and it's. That's a long answer. 274 00:13:24,520 --> 00:13:26,400 Yeah, yeah, no, no, I really appreciate it. 275 00:13:26,400 --> 00:13:28,240 And I think it's the same with that. 276 00:13:28,240 --> 00:13:30,320 I used to run an IT company and we used to provide a lot of 277 00:13:30,320 --> 00:13:32,520 phone systems and we had like level 1, level 2, Level 3 278 00:13:32,520 --> 00:13:35,640 support, and level 1 support was like, my phone's not working. 279 00:13:35,640 --> 00:13:37,720 Have you turned it off and on? No, do that. 280 00:13:37,840 --> 00:13:40,280 Oh, it's working now. That's like level 1, level 2. 281 00:13:40,280 --> 00:13:42,080 Then it's like, OK, let's reinstall some software and 282 00:13:42,080 --> 00:13:45,400 stuff, but you, the end user are going to do it, tell you how to 283 00:13:45,400 --> 00:13:47,400 do it and you're going to do it. And we obviously need a more 284 00:13:47,400 --> 00:13:49,920 senior person that Level 3 is like something's broken in the 285 00:13:49,920 --> 00:13:52,040 data center and that we need to fix that. 286 00:13:52,040 --> 00:13:53,880 And that's Level 3. And you're absolutely right. 287 00:13:53,880 --> 00:13:57,040 If we just had everybody like maybe one person at level 1, one 288 00:13:57,040 --> 00:14:00,280 person level 2 and then 50 people at Level 3, that's a good 289 00:14:00,280 --> 00:14:02,120 way of breaking down those levels. 290 00:14:02,520 --> 00:14:04,440 And I think it's really interesting what you say 291 00:14:04,480 --> 00:14:07,640 actually about the and I think it's applies to any role in 292 00:14:07,640 --> 00:14:09,880 sales. And actually to go on a very 293 00:14:09,880 --> 00:14:13,320 quick tangent, everyone listening, marketing is most the 294 00:14:13,320 --> 00:14:14,880 most important thing in marketing is the person you 295 00:14:14,960 --> 00:14:16,240 marketing to. And I love what you just said 296 00:14:16,240 --> 00:14:19,120 about there different people you find yourself in front of from a 297 00:14:19,120 --> 00:14:21,240 sales perspective because they've all got their own 298 00:14:21,240 --> 00:14:24,120 different drivers and you need to get those things to align. 299 00:14:24,640 --> 00:14:26,240 And the hardest thing is, especially if you've got 300 00:14:26,240 --> 00:14:29,440 multiple people, it's the conversations they have between 301 00:14:29,440 --> 00:14:31,400 themselves that you're not in because obviously you're 302 00:14:31,400 --> 00:14:33,360 external. And it's when they're like that 303 00:14:33,360 --> 00:14:36,480 Tedio talk, he didn't manage to answer this very, he said he 304 00:14:36,480 --> 00:14:38,400 can't do that. That's a real sticking point. 305 00:14:38,560 --> 00:14:40,400 And you need someone else to advocate for you to go. 306 00:14:40,600 --> 00:14:43,080 But we don't need that. That's not a feature we need. 307 00:14:43,080 --> 00:14:45,320 Or actually, even if it doesn't have that, we're still saving 308 00:14:45,320 --> 00:14:47,240 40% efficiency. That's. 309 00:14:47,240 --> 00:14:51,080 The thing more accurate, yeah, yeah, yeah, no, just to your 310 00:14:51,080 --> 00:14:54,880 point there, people are are so fixated, some stakeholders, they 311 00:14:54,880 --> 00:14:58,040 get fixated on the problem today. 312 00:14:58,080 --> 00:15:01,080 And and then you start going back and forth about really 313 00:15:01,080 --> 00:15:04,080 trivial things like oh, what about the extra click that we're 314 00:15:04,080 --> 00:15:06,200 going to have to face on this platform, whereas our old 315 00:15:06,200 --> 00:15:08,280 platform, our agents don't have to do that extra click. 316 00:15:08,720 --> 00:15:10,320 And so it's exactly what you just said. 317 00:15:10,520 --> 00:15:12,960 What about the 40% resolution so. 318 00:15:13,560 --> 00:15:15,960 We, I mean, we have it with some clients and yeah, just any 319 00:15:15,960 --> 00:15:18,720 clients listening right now or yeah, any potential clients 320 00:15:18,720 --> 00:15:21,040 we're going to pitch to in the future, like please hear this in 321 00:15:21,040 --> 00:15:22,600 the nicest, politest possible way. 322 00:15:22,600 --> 00:15:23,640 I'm doing this for your own good. 323 00:15:24,080 --> 00:15:26,080 Sometimes we show data to people. 324 00:15:26,240 --> 00:15:28,880 And I've just done a talk in London when this episode goes 325 00:15:28,880 --> 00:15:31,760 out and where I'm specifically talking about the assumptions we 326 00:15:31,760 --> 00:15:34,720 make in marketing because we haven't checked the data. 327 00:15:34,880 --> 00:15:36,520 And once we've made the assumption and deployed a 328 00:15:36,520 --> 00:15:38,320 change, we still don't check the data. 329 00:15:38,600 --> 00:15:40,920 And I'm sharing some case studies, which I'll be the 330 00:15:40,920 --> 00:15:43,480 probably the first agencies ever stood on the stage and go look, 331 00:15:43,560 --> 00:15:45,720 look at these negative results on some case studies. 332 00:15:45,720 --> 00:15:48,240 And it's where we've all like us, including we've made the 333 00:15:48,240 --> 00:15:50,480 wrong assumptions. We've gone, no one reads all 334 00:15:50,480 --> 00:15:52,600 this chunky text, put it behind a read more button. 335 00:15:52,600 --> 00:15:54,280 So we did and conversion rate dropped. 336 00:15:54,520 --> 00:15:57,360 People wanted the text and we just made an assumption and 337 00:15:57,360 --> 00:15:58,800 didn't test it. And then when we did test it 338 00:15:58,800 --> 00:16:01,440 retrospectively, Oh my gosh, it made such a difference. 339 00:16:01,800 --> 00:16:05,360 But I think it's, as I say, it's just so important when we look 340 00:16:05,400 --> 00:16:07,760 at who we're selling to and that's marketing. 341 00:16:07,960 --> 00:16:10,680 But also when you look at the your system and actually the 342 00:16:10,680 --> 00:16:14,240 people in customer service, the more I think from my 343 00:16:14,240 --> 00:16:16,880 perspective, the more AI can unlock some of the key 344 00:16:16,880 --> 00:16:19,800 information about the customer you're speaking to, the quicker 345 00:16:19,800 --> 00:16:22,280 you're going to get the answer. You mentioned something I think 346 00:16:22,280 --> 00:16:24,120 is really, really valuable. And I talk about this quite a 347 00:16:24,120 --> 00:16:26,080 lot on the show. So if you've heard this before, 348 00:16:26,200 --> 00:16:27,600 click boil on the kettle quickly. 349 00:16:27,960 --> 00:16:31,560 And we talk a lot about talking to the customer service team to 350 00:16:31,560 --> 00:16:35,320 find out what people are asking. And we as an SEO agency, we put 351 00:16:35,320 --> 00:16:37,880 that into content. So if everyone's phoning up 352 00:16:37,880 --> 00:16:42,000 going, I have not purchased yet, I am pre purchased, I'm going to 353 00:16:42,000 --> 00:16:45,480 the extent of a phone call to find out when I order this item 354 00:16:45,480 --> 00:16:47,800 that weighs 400 kilograms. I don't know what that is in 355 00:16:47,800 --> 00:16:50,360 pounds for you guys in the US, it's heavy like several, like 356 00:16:50,360 --> 00:16:53,600 probably about 7 or 8 humans worth who work out and aren't 357 00:16:53,600 --> 00:16:56,680 too tall. So something's really heavy. 358 00:16:56,680 --> 00:16:58,640 You phone customer service, say, is it going to be delivered by 359 00:16:58,640 --> 00:17:01,640 two people and they go, oh, the contents on the website that is 360 00:17:01,760 --> 00:17:04,839 SEO gold dust because they're just the ones that phoned up. 361 00:17:05,160 --> 00:17:07,800 The other people would go, well, this company clearly just Chuck 362 00:17:07,800 --> 00:17:10,280 it in a van and someone wheels it to your front door and that's 363 00:17:10,280 --> 00:17:13,359 the end of the process. So question to you, and this is 364 00:17:13,359 --> 00:17:15,400 slightly off pieced a little bit talking about the AI, but I'm 365 00:17:15,400 --> 00:17:18,599 sure the AI can support this. How much reporting did your 366 00:17:18,599 --> 00:17:21,680 customers get from TIDO and how much, how many questions they 367 00:17:21,680 --> 00:17:23,800 start asking once they've got a better piece of software like 368 00:17:23,800 --> 00:17:27,480 you guys to then go this, this is our marketing strategies 369 00:17:27,480 --> 00:17:29,840 breaking down all these pain points to make the sale. 370 00:17:30,760 --> 00:17:35,960 Yeah, I would say the analytics are more than half the battle to 371 00:17:35,960 --> 00:17:38,640 making the AI perform well anyways. 372 00:17:39,280 --> 00:17:41,760 So it's it's going to serve you on multiple fronts. 373 00:17:41,760 --> 00:17:44,600 But no matter how much data you have today, even if you've been 374 00:17:44,600 --> 00:17:47,840 running your support operation for 15 years and you've got all 375 00:17:47,840 --> 00:17:51,160 the transcripts, you're only going to make your AIX percent 376 00:17:51,160 --> 00:17:53,440 smart using the content that you have today. 377 00:17:53,480 --> 00:17:57,760 And then the rest is going to come from real interactions and 378 00:17:57,760 --> 00:18:00,560 people that went and asked the AI that unique question for the 379 00:18:00,560 --> 00:18:02,840 first time. So the TEDIO analytics, 380 00:18:03,160 --> 00:18:07,200 obviously the Gold Star for us is resolution rate, which is the 381 00:18:07,200 --> 00:18:09,360 percentage of chats that are resolved without human 382 00:18:09,360 --> 00:18:11,960 interaction. But what's really cool is 383 00:18:11,960 --> 00:18:15,680 looking into conversations by intent, which is basically the 384 00:18:15,680 --> 00:18:17,640 topic of these of these conversations. 385 00:18:18,000 --> 00:18:20,840 So beyond knowing what people are asking, we can even group 386 00:18:20,840 --> 00:18:24,880 them into kind of assigning A magnitude to the topic so you 387 00:18:24,880 --> 00:18:28,800 can see exactly how popular certain topics were inside of 388 00:18:28,800 --> 00:18:31,680 your AI. And like the major intention 389 00:18:32,640 --> 00:18:34,800 when we built that analytic, it was for the AI. 390 00:18:34,800 --> 00:18:37,560 It definitely wasn't to influence marketing teams. 391 00:18:37,640 --> 00:18:39,600 Data is good for everyone regardless. 392 00:18:39,960 --> 00:18:43,440 We were using it to understand where should we focus the most 393 00:18:43,440 --> 00:18:47,920 time in regards to ongoing training of the AI agent and the 394 00:18:47,920 --> 00:18:50,360 data sources. Because if yeah, if you get 395 00:18:50,360 --> 00:18:54,080 10,000 people into the chat every month and 6000 people are 396 00:18:54,080 --> 00:18:58,000 talking about the process of your subscription service and 397 00:18:58,280 --> 00:19:00,480 that's what we knew to focus on content wise. 398 00:19:00,800 --> 00:19:01,960 Right. Absolutely. 399 00:19:02,120 --> 00:19:04,200 I do. I, I, I love doing SEO because 400 00:19:04,520 --> 00:19:06,880 if you get the right marketing director or someone like the CEO 401 00:19:06,880 --> 00:19:09,520 senior who gets behind it and gets the strategy. 402 00:19:09,520 --> 00:19:12,040 It's like we're now chatting to the PR teams, the customer 403 00:19:12,040 --> 00:19:15,120 service, the sales teams. We're talking to customers to 404 00:19:15,120 --> 00:19:16,480 find out what they like, what they don't like. 405 00:19:16,640 --> 00:19:18,720 We're talking to the tech teams about all the analytics like 406 00:19:19,040 --> 00:19:21,720 e-mail, like no one is off limits to us because it all 407 00:19:21,720 --> 00:19:24,520 influences the content on the front end for indexing on Google 408 00:19:24,520 --> 00:19:26,840 to provide a good experience. And then we proved that good 409 00:19:26,840 --> 00:19:29,240 experience. So the only other thing I'll 410 00:19:29,240 --> 00:19:33,160 just wrap up on my SEO point about a customer service would 411 00:19:33,160 --> 00:19:35,680 be you mentioned like how many resolved thing, you know, how 412 00:19:35,680 --> 00:19:38,360 many things were resolved? It's, it's one thing for I think 413 00:19:38,360 --> 00:19:40,960 a website just to say like excellent customer service. 414 00:19:41,320 --> 00:19:44,040 And we actually ran a test on a site where we put excellent 415 00:19:44,040 --> 00:19:46,400 customer service because it was and people spoke about it in the 416 00:19:46,400 --> 00:19:48,760 reviews and we made an assumption that we tested. 417 00:19:48,760 --> 00:19:51,160 And the assumption again, it's going to become a big part of 418 00:19:51,160 --> 00:19:52,480 the podcast guys, is assumptions. 419 00:19:52,760 --> 00:19:55,720 The assumption that we tested was people talk about it in 420 00:19:55,720 --> 00:19:57,200 reviews, how great the customer service is. 421 00:19:57,200 --> 00:20:00,240 And everyone's so lovely when they reach out to us as a brand 422 00:20:00,240 --> 00:20:02,520 with my my client as a brand and talk to them. 423 00:20:02,520 --> 00:20:04,880 We thought people cared. We put excellent customer 424 00:20:04,880 --> 00:20:08,000 service in there, like header of the site, like excellent, but we 425 00:20:08,000 --> 00:20:10,520 didn't prove it or anything. And we tried five star customer 426 00:20:10,520 --> 00:20:12,520 service had no influence on conversion rate. 427 00:20:12,760 --> 00:20:14,240 We did not get more people buying. 428 00:20:14,560 --> 00:20:16,800 And then we took a step back and tried hypothesizing why? 429 00:20:17,000 --> 00:20:18,800 Why? We not get more sales and one of 430 00:20:18,800 --> 00:20:21,240 the things we think, and it's always the thing you think the 431 00:20:21,240 --> 00:20:24,920 data tells you what's happened, not why or or or how the the 432 00:20:24,920 --> 00:20:27,840 thing we think now is if we have to reassure people pre sale that 433 00:20:27,840 --> 00:20:30,320 we've got good customer service, something's gone wrong. 434 00:20:30,800 --> 00:20:33,240 So we're still in that grey box at the moment with that 435 00:20:33,240 --> 00:20:36,200 particular client of going what's because like, it's not 436 00:20:36,200 --> 00:20:38,240 it's not an expensive set of products they've got. 437 00:20:38,240 --> 00:20:40,560 It's not something you would save up for loads. 438 00:20:40,560 --> 00:20:42,680 Their Klana doesn't get used very often for the like buy now, 439 00:20:42,680 --> 00:20:44,320 pay later. So we're still in that little 440 00:20:44,320 --> 00:20:46,200 black box with the customer service one of going, what 441 00:20:46,200 --> 00:20:48,080 content do we need to put in just to basically say we're 442 00:20:48,080 --> 00:20:50,920 really nice people will will be reasonable. 443 00:20:50,920 --> 00:20:53,800 And if something goes wrong, we'll fix it without going 444 00:20:54,040 --> 00:20:56,280 things go wrong. So when they do, we'll fix it 445 00:20:56,360 --> 00:20:57,840 because that's not a reassuring message. 446 00:20:58,120 --> 00:21:01,160 So I think that that analytics engine and which is where AI is 447 00:21:01,160 --> 00:21:03,320 so powerful, right? We give it a spreadsheet. 448 00:21:03,440 --> 00:21:06,400 These are all the recent transcript chats to do with 449 00:21:06,400 --> 00:21:08,760 delivery information. Tell me what we're missing or 450 00:21:08,760 --> 00:21:10,640 write me some web copy that kind of summarizes all this to 451 00:21:10,640 --> 00:21:13,120 reassure people that like our delivery is really good. 452 00:21:14,200 --> 00:21:16,760 And, and the AI we, we put it to ChatGPT and said like these, 453 00:21:16,760 --> 00:21:19,240 these are just 50 out of like 50,000 customers that have 454 00:21:19,480 --> 00:21:21,400 complained about delivery. And that was it in that period 455 00:21:21,400 --> 00:21:24,120 where we had 50,000 customers and the chat, ChatGPT came back 456 00:21:24,120 --> 00:21:27,480 with all these like tag lines of like, you know, X percent, you 457 00:21:27,480 --> 00:21:30,880 know, 99.9% customer satisfaction with our delivery 458 00:21:30,880 --> 00:21:33,120 service. And I was like, OK, cool, people 459 00:21:33,120 --> 00:21:34,280 are buying now. And I think it's really 460 00:21:34,280 --> 00:21:35,280 interesting to kind of do that bit. 461 00:21:35,280 --> 00:21:37,640 But AI is so good for that analytical bit, isn't it? 462 00:21:38,160 --> 00:21:40,720 Sure. Well even like the teams I meet 463 00:21:40,720 --> 00:21:43,080 with that don't have any chat in place. 464 00:21:43,080 --> 00:21:46,520 It's also even interesting just to see not from the customer 465 00:21:46,520 --> 00:21:49,720 support side of things, but from the pre sales side of things. 466 00:21:49,960 --> 00:21:53,480 What's the first question one of your website visitors has when 467 00:21:53,480 --> 00:21:55,600 they land on the site? And if you don't have a really 468 00:21:55,600 --> 00:21:59,240 low friction channel for them to ask that question, you're not 469 00:21:59,240 --> 00:22:01,600 going to know. But if you create, I've seen 470 00:22:01,600 --> 00:22:03,160 strategies where people just say, hey, if you have any 471 00:22:03,160 --> 00:22:05,000 questions, let us know. Type them right here in a chat. 472 00:22:05,320 --> 00:22:07,920 And then, I mean, you have direct feedback from what these 473 00:22:07,920 --> 00:22:10,880 people are actually curious and super, yeah, super cool. 474 00:22:11,120 --> 00:22:14,040 And so I mean, using someone like TIDO or, you know, or an AI 475 00:22:14,040 --> 00:22:16,800 bot in that situation to help make that sale to provide the 476 00:22:16,800 --> 00:22:19,520 right information. How hard is it? 477 00:22:19,520 --> 00:22:22,280 And this is a really loaded question that I'm sure guests 478 00:22:22,280 --> 00:22:24,000 want me to ask. Do you ever find that in your 479 00:22:24,000 --> 00:22:26,960 warehouse you have dead stock, you can't sell it, no one wants 480 00:22:26,960 --> 00:22:28,280 to buy it and you can't get a refund on it. 481 00:22:28,520 --> 00:22:31,600 Or alternatively, you're missing inventory in your warehouse that 482 00:22:31,600 --> 00:22:33,760 you need, you've got sales coming in and you're low on 483 00:22:33,760 --> 00:22:36,880 stock and you haven't ordered more stock in time, All these 484 00:22:36,880 --> 00:22:38,480 things hit profit. And I've been there. 485 00:23:08,840 --> 00:23:13,480 How hard is it to get an AI bot to a chat bot or something to 486 00:23:13,480 --> 00:23:16,240 provide the information quickly on a web page? 487 00:23:16,240 --> 00:23:18,280 Like say all the customer service agents are away and it's 488 00:23:18,280 --> 00:23:19,720 like, you know, talk to our chat bot, we can help you. 489 00:23:20,080 --> 00:23:22,240 How hard is it to get that balance of like factual 490 00:23:22,240 --> 00:23:26,360 information, being salesy enough like you or I would in a pitch 491 00:23:26,680 --> 00:23:27,880 to get someone to buy that product? 492 00:23:28,080 --> 00:23:31,400 How do you how do you train the AI bot to to know enough of that 493 00:23:31,400 --> 00:23:34,680 to be effective whilst maintaining that kind of brand 494 00:23:34,840 --> 00:23:36,640 stability? Sure. 495 00:23:36,680 --> 00:23:40,080 I some people are definitely in a better spot than others and 496 00:23:40,080 --> 00:23:41,800 you've been talking a lot about SEO. 497 00:23:41,800 --> 00:23:46,920 The sites that I click on and train the AI on that have a nice 498 00:23:47,120 --> 00:23:50,120 effort put into the SEO. They really talk about their 499 00:23:50,120 --> 00:23:52,440 services and why they are good at what they do. 500 00:23:52,440 --> 00:23:56,240 They build credibility and talking about current use cases 501 00:23:56,240 --> 00:23:58,600 or clients or testimonials and things like that. 502 00:23:59,120 --> 00:24:02,640 Many times for like a well written, well built website, 503 00:24:02,800 --> 00:24:05,680 that's enough. That is you, you've already put 504 00:24:05,680 --> 00:24:08,440 in all the time and effort into figuring out how do we say what 505 00:24:08,440 --> 00:24:10,280 we want to say that is compelling. 506 00:24:10,520 --> 00:24:15,200 Most websites already have that. So the AI is just taking from 507 00:24:15,600 --> 00:24:19,640 like the main influence on Lyro AI through Tidio at least is 508 00:24:19,640 --> 00:24:21,560 about the data sources, what you train it on. 509 00:24:21,560 --> 00:24:25,360 So if I answer every question with a credibility building 510 00:24:25,360 --> 00:24:27,640 statement at the end of it, generally the AI is going to 511 00:24:27,640 --> 00:24:29,720 pick up on that and and do that same kind of stuff. 512 00:24:29,760 --> 00:24:33,080 So in that case, minutes, right, you've already done all the hard 513 00:24:33,080 --> 00:24:35,240 work. Yeah, yeah, I think it's really 514 00:24:35,240 --> 00:24:37,600 powerful as well. Just saying to AI like crawl the 515 00:24:37,600 --> 00:24:39,840 site, grab any content you want from this site. 516 00:24:39,840 --> 00:24:42,840 And but tell me if you think anything's like a bit odd or 517 00:24:42,840 --> 00:24:44,400 doesn't align with the rest of the site. 518 00:24:44,400 --> 00:24:47,080 Because we all, we've all got a page in our website somewhere 519 00:24:47,080 --> 00:24:49,400 that says, hey, we give out free ice cream or something, but we 520 00:24:49,400 --> 00:24:50,960 don't anymore. But someone's not deleted the 521 00:24:50,960 --> 00:24:52,920 page. And but I mean, all of this, 522 00:24:52,920 --> 00:24:55,000 we're talking a lot about the chat AI side. 523 00:24:55,280 --> 00:24:58,200 How are customer expectations changing as a result of all 524 00:24:58,200 --> 00:25:01,240 these changes and AI becoming more of a feature in the buying, 525 00:25:01,520 --> 00:25:02,920 the buying journey, the buying cycle? 526 00:25:03,160 --> 00:25:06,560 Sure. And I a lot, it's changing a lot 527 00:25:06,560 --> 00:25:09,680 and I think it's a lot it's definitely due to more than just 528 00:25:09,680 --> 00:25:13,480 how businesses are building their support stacks. 529 00:25:14,680 --> 00:25:17,760 It's everywhere, right? Like students in university, 530 00:25:17,760 --> 00:25:21,480 they're all using AI for their note taking, etcetera, etcetera. 531 00:25:23,040 --> 00:25:26,400 A lot of people who find our business when I asked how'd you 532 00:25:26,400 --> 00:25:27,920 find us, by the way, what were you looking at? 533 00:25:27,920 --> 00:25:31,720 Well, I was talking to AI. So I think that beyond just 534 00:25:33,000 --> 00:25:35,920 what's on your website and how you tailor your support team to 535 00:25:35,920 --> 00:25:39,560 work with AI, they're already inevitably going to use it in 536 00:25:39,640 --> 00:25:43,120 their Google search when it pops up that AI summary or when they 537 00:25:43,120 --> 00:25:47,440 go to ChatGPT to find it. So the expectation of speed is 538 00:25:47,440 --> 00:25:51,160 already like, well planted in their heads. 539 00:25:52,240 --> 00:25:55,240 And I think that that's the big thing is going to be speed. 540 00:25:55,240 --> 00:25:56,360 How's it? Yeah. 541 00:25:56,360 --> 00:25:59,000 How's it changing the landscape? It's making everything way 542 00:25:59,000 --> 00:26:01,680 faster. And the expectation is that you 543 00:26:01,680 --> 00:26:04,600 should be fast. And maybe I'm the worst, like 544 00:26:04,600 --> 00:26:08,400 I'm the worst customer because I know what is possible with AI. 545 00:26:08,720 --> 00:26:12,880 So if I'm waiting for if you keep me waiting longer than 5 546 00:26:12,880 --> 00:26:14,840 minutes and I'm trying to do something like change my 547 00:26:14,840 --> 00:26:17,880 shipping address. Like I said, I'm the worst. 548 00:26:17,880 --> 00:26:20,960 But I'm fuming because I'm just like, come on, you, you guys are 549 00:26:21,240 --> 00:26:24,800 you do maybe millions in revenue every year and you couldn't put 550 00:26:24,800 --> 00:26:28,400 in the basic, the most basic of tools to help me help myself at 551 00:26:29,080 --> 00:26:31,600 midnight or whenever I'm on my phone doing doing my and. 552 00:26:31,640 --> 00:26:34,280 The danger of all of that, that's the unquantifiable 553 00:26:34,280 --> 00:26:37,760 annoyance of customers as I like to call it or I don't always 554 00:26:37,760 --> 00:26:39,320 call it exactly that, but something similar. 555 00:26:39,320 --> 00:26:41,160 But you get the idea that actually that that is 556 00:26:41,160 --> 00:26:46,160 frustrating customers like you, but like there's no analytics is 557 00:26:46,160 --> 00:26:48,240 going to tell you like, you know, Jacob's sitting at home 558 00:26:48,240 --> 00:26:51,000 frustrated right now. So sometimes, and this is going 559 00:26:51,000 --> 00:26:53,600 to go against the whole grain of my agency and what we do about 560 00:26:53,600 --> 00:26:56,360 testing and data mining. There are some things you just 561 00:26:56,360 --> 00:26:58,440 don't know. I sometimes say to clients, like 562 00:26:58,840 --> 00:27:01,760 there, there is such a thing as just good and bad layouts, good 563 00:27:01,760 --> 00:27:04,200 and bad design. And you can start to prove it 564 00:27:04,200 --> 00:27:07,600 with heat maps and other stuff. So I think with that, you could 565 00:27:07,600 --> 00:27:10,160 see like the exit rates of people that are waiting over 2 566 00:27:10,160 --> 00:27:12,480 minutes for a response and go, all right, well, that becomes a 567 00:27:12,480 --> 00:27:14,040 KPI. Our customer service team needs 568 00:27:14,040 --> 00:27:18,040 to respond within two minutes. But all of this I just realised 569 00:27:18,040 --> 00:27:20,280 one of my notes from an earlier question we didn't cover, but I 570 00:27:20,280 --> 00:27:22,640 really do want to ask you about this because this is something 571 00:27:22,640 --> 00:27:25,960 that we, we had a round table in the winning shop like growth hub 572 00:27:25,960 --> 00:27:27,800 the other day. As I mentioned in my intro, 573 00:27:27,800 --> 00:27:30,240 guys, go join it if you want. We were talking about AI. 574 00:27:30,240 --> 00:27:32,240 We were actually supposed to be talking about e-mail content and 575 00:27:32,240 --> 00:27:34,280 we end up talking a little about AI because AI creates a lot of 576 00:27:34,280 --> 00:27:34,840 e-mail content. Now. 577 00:27:34,840 --> 00:27:37,480 Then we had a separate session on AI with a separate group of 578 00:27:37,480 --> 00:27:40,240 people as well, a separate group of brands, and both groups 579 00:27:40,440 --> 00:27:44,680 shared a concern about privacy and giving AI access to 580 00:27:44,680 --> 00:27:47,760 information. You guys, obviously to do what 581 00:27:47,760 --> 00:27:51,280 you do, you are giving the AI and personal customer data 582 00:27:51,280 --> 00:27:53,720 information, you're giving it sales information. 583 00:27:53,960 --> 00:27:57,240 There's a lot of data in there. So how do you, I mean, to ask a 584 00:27:57,240 --> 00:27:59,680 question about the tool? How do you, A, keep it compliant 585 00:27:59,680 --> 00:28:02,200 and B, like there are obviously risks around that. 586 00:28:02,200 --> 00:28:06,080 Like what should people be aware of when using AI and unleashing 587 00:28:06,080 --> 00:28:08,560 it to their e-mail database, their Shopify data, and then 588 00:28:08,560 --> 00:28:10,520 other similar tools? Sure. 589 00:28:10,520 --> 00:28:13,720 I think a, a big thing that I always remind people is that, 590 00:28:14,240 --> 00:28:18,240 and I can only speak to Lyro and and Tedio's, Tedio's product, 591 00:28:18,600 --> 00:28:22,160 but it's never inventing or creating, right. 592 00:28:22,160 --> 00:28:25,560 So it really only regurgitates what we tell. 593 00:28:25,560 --> 00:28:29,360 It is the truth and it, it kind of looks at the whole truth as 594 00:28:29,720 --> 00:28:32,040 what we've given it inside of the data sources. 595 00:28:33,760 --> 00:28:38,600 I think I run into more barriers like customer customer barriers 596 00:28:38,600 --> 00:28:44,160 around like trust and just preconceived ideas about AI and 597 00:28:44,160 --> 00:28:48,040 what it can do. Can AI take a piece of private 598 00:28:48,040 --> 00:28:51,560 information and do crazy malicious things with it? 599 00:28:51,560 --> 00:28:56,040 Like, yeah, if someone, if someone built it to do that 100% 600 00:28:56,040 --> 00:28:58,800 that that's possible. But in my experience, just on 601 00:28:58,800 --> 00:29:02,080 the, the sales side of things, like we're, we're held to the 602 00:29:02,080 --> 00:29:05,560 same standards as like maybe your, your CRM would be as far 603 00:29:05,560 --> 00:29:09,280 as how compliant are you? How do you protect data and 604 00:29:09,280 --> 00:29:12,440 privacy just as much as your CRM does, right? 605 00:29:12,440 --> 00:29:16,960 And so I mean, and it's all SAS and all the SAS is so connected 606 00:29:16,960 --> 00:29:19,200 together. So I think even tools that are 607 00:29:20,160 --> 00:29:23,720 not directly exposed to customer data, they also people want to 608 00:29:23,720 --> 00:29:26,720 see your SoC compliance, they want to know that you're, they, 609 00:29:26,840 --> 00:29:29,400 they're asking us where our data is stored and etcetera, 610 00:29:29,400 --> 00:29:33,440 etcetera. So no, it it, it feels extremely 611 00:29:34,280 --> 00:29:36,200 direct, but it should be. I mean, yeah. 612 00:29:36,600 --> 00:29:39,440 Absolutely and I think that one of the dangers that one of the 613 00:29:39,440 --> 00:29:42,680 fears that came up as a danger from one of the brands and I 614 00:29:42,680 --> 00:29:44,920 won't I won't show your names. It was a nice clothes safe space 615 00:29:44,920 --> 00:29:46,320 and I get a lot of people say I couldn't make it. 616 00:29:46,320 --> 00:29:47,760 Was it recorded? It's like no, no. 617 00:29:47,760 --> 00:29:50,400 The idea of a round table was you have a seat at the table and 618 00:29:50,400 --> 00:29:52,880 you share as much as you listen, which is fine. 619 00:29:52,920 --> 00:29:55,840 Like, yeah, so there were, there were enough of them were running 620 00:29:55,840 --> 00:29:58,960 another one just for those guys. But the the question, sorry, a 621 00:29:58,960 --> 00:30:00,680 complete tangent, but the question was actually around 622 00:30:00,680 --> 00:30:05,120 like, I'm nervous that if I had an AI bot rather than a customer 623 00:30:05,120 --> 00:30:08,360 service Rep and someone said, Hey, I'm, you know, like I know 624 00:30:08,360 --> 00:30:10,880 the neighbors 6 doors down and I know they order from this. 625 00:30:11,080 --> 00:30:13,120 I'm going to pretend to be them and say I've lost my order 626 00:30:13,120 --> 00:30:15,080 number. Like, how does, how do you make 627 00:30:15,080 --> 00:30:18,880 sure that AI like doesn't share someone else's information in 628 00:30:18,880 --> 00:30:20,960 the same way? I think it was Samsung, wasn't 629 00:30:20,960 --> 00:30:22,320 it? They'd put some confidential 630 00:30:22,320 --> 00:30:25,800 information into an open AI and then found out that like if you 631 00:30:25,800 --> 00:30:29,040 just said, what did Samsung, Samsung ask yesterday about 632 00:30:29,040 --> 00:30:31,400 this, this and this? It shared the IT said, oh, we're 633 00:30:31,400 --> 00:30:33,240 asking about this and like shared the information because I 634 00:30:33,240 --> 00:30:34,720 think it was chat GBC or open AI. 635 00:30:34,720 --> 00:30:37,280 So it didn't know like not to share it. 636 00:30:37,280 --> 00:30:39,960 It didn't know it was confidential, which obviously it 637 00:30:39,960 --> 00:30:41,440 hit the headlines all across the world. 638 00:30:41,440 --> 00:30:44,440 So like it's changed. But yeah, so just there is that 639 00:30:44,440 --> 00:30:47,760 level of scarcity around AI. Isn't it that like how will it 640 00:30:47,760 --> 00:30:49,320 know to share the correct information? 641 00:30:49,520 --> 00:30:51,200 But you can build all those parameters in, can't you? 642 00:30:51,200 --> 00:30:57,240 A. 100% Yeah, yeah, yeah. And that's it's more, it's more 643 00:30:57,240 --> 00:30:59,480 important than it even giving right answers. 644 00:30:59,720 --> 00:31:01,680 I would say like when you, when you think about how we're 645 00:31:01,680 --> 00:31:04,840 building it and what we're cautious about, I would say 646 00:31:04,840 --> 00:31:08,000 there's more of a focus in ensuring number one, that like 647 00:31:08,000 --> 00:31:10,480 data is completely siloed. A lot of customer data too. 648 00:31:10,480 --> 00:31:13,440 It's, it's only referenced temporarily by the AI. 649 00:31:14,360 --> 00:31:17,840 So it's not that it always has access to this value. 650 00:31:17,840 --> 00:31:22,240 It's when it's right, right. It's like, oh, I can find that 651 00:31:22,240 --> 00:31:25,040 for you. It finds it and then we have to 652 00:31:25,040 --> 00:31:28,520 program in those, double check the e-mail, double check the 653 00:31:28,560 --> 00:31:30,840 order ID, but we emailed you a code. 654 00:31:30,840 --> 00:31:34,400 Can you use that kind of thing? So I mean it's yeah, similar 655 00:31:34,400 --> 00:31:39,600 similar to any, I would say any customer facing automation or 656 00:31:39,760 --> 00:31:41,320 platform or something like that, yeah. 657 00:31:41,640 --> 00:31:43,360 I did. I had a funny issue with AI 658 00:31:43,360 --> 00:31:46,560 recently and this is me using it like new chat every time I ask 659 00:31:46,560 --> 00:31:49,120 it anything just like a classic AI kind of amateur. 660 00:31:49,520 --> 00:31:52,360 And I asked the other day we run meet ups and we've in fact, 661 00:31:52,360 --> 00:31:54,280 we've actually got a meet up in September if anyone's to come. 662 00:31:54,280 --> 00:31:57,320 Molly are hosting us in London and Shopify brands. 663 00:31:57,320 --> 00:31:58,360 If you're not unsure if you still come. 664 00:31:58,600 --> 00:32:01,440 So we're running that and I said right, right, AI. 665 00:32:01,560 --> 00:32:03,160 Here's all the information of the key points. 666 00:32:03,400 --> 00:32:06,680 Write me an event description. And I copy and pasted it. 667 00:32:06,680 --> 00:32:09,040 One of my teams, I said, Nick, you know, you like playing with 668 00:32:09,120 --> 00:32:11,400 AI. Yeah, they said you've been 669 00:32:11,400 --> 00:32:13,880 playing with AI again. It's like, yeah, why? 670 00:32:14,160 --> 00:32:15,680 And they say you've obviously used it to write the 671 00:32:15,680 --> 00:32:16,840 description. It's like, how do you know that? 672 00:32:16,840 --> 00:32:19,200 And it's like, because it's on like, I think it's the 4th of 673 00:32:19,200 --> 00:32:21,080 September, but it says the 6th everywhere. 674 00:32:21,560 --> 00:32:22,960 I was like, where's it got the 6th from? 675 00:32:22,960 --> 00:32:24,560 And I checked all the information I gave it went back 676 00:32:24,560 --> 00:32:26,400 over the chair. I never told it the 6th. 677 00:32:26,440 --> 00:32:29,680 I said it was the 4th and shared like part of the e-mail to say, 678 00:32:29,680 --> 00:32:32,160 like, you know, look at this e-mail from Molly. 679 00:32:32,400 --> 00:32:34,400 These will be information, information about the cocktails 680 00:32:34,400 --> 00:32:36,840 that are going to be there and other stuff handled this 681 00:32:36,840 --> 00:32:39,560 information in and it did and it even it said there was going to 682 00:32:39,560 --> 00:32:41,280 be. I think it was like free hot 683 00:32:41,280 --> 00:32:43,560 dogs and pizza. I was like, we never said hot 684 00:32:43,560 --> 00:32:45,560 dogs and pizza. So like it had and I think it 685 00:32:45,560 --> 00:32:48,480 was ChatGPT or manus I was using and it started making stuff up. 686 00:32:48,760 --> 00:32:51,040 So I think I think the good thing that you guys are doing is 687 00:32:51,040 --> 00:32:53,320 you're using a locked as you say, like it just does one 688 00:32:53,320 --> 00:32:56,040 request and goes right, find this order in this list, grab 689 00:32:56,040 --> 00:32:58,960 this information, then double check that, then pass it 690 00:32:58,960 --> 00:33:00,800 through. Whereas I was asking to do an 691 00:33:00,840 --> 00:33:03,360 open question of like write me a nice thing about this and it 692 00:33:03,360 --> 00:33:05,320 gone. Oh, the 6th of September must be 693 00:33:05,320 --> 00:33:08,440 nicer and pizza and and hot dogs sounds nicer because one of the 694 00:33:08,440 --> 00:33:10,640 guys at Molly emailed me and said to Nick, we're not doing, 695 00:33:10,640 --> 00:33:11,960 we haven't agreed what food we're doing yet. 696 00:33:11,960 --> 00:33:14,560 And it's like, oh, if you could do pizza and hot dogs, it would 697 00:33:14,680 --> 00:33:16,720 cheer up the AI bots because that's who said it. 698 00:33:17,040 --> 00:33:20,800 So it is important, I think with AI that a lot of people go, if I 699 00:33:20,800 --> 00:33:22,800 just get it say like, you know, write me a story, it will give 700 00:33:22,800 --> 00:33:24,360 me an amazing story exactly as I want. 701 00:33:24,360 --> 00:33:26,520 It's like, but you haven't given it any information. 702 00:33:26,560 --> 00:33:28,640 I think that's where the process kind of comes in, doesn't it? 703 00:33:29,400 --> 00:33:33,560 Yeah, yeah, people also like there's, it's so much deeper 704 00:33:33,560 --> 00:33:36,880 than just a like a ChatGPT script as well, right. 705 00:33:36,880 --> 00:33:40,120 And without getting like way into it, you know, they've done 706 00:33:40,120 --> 00:33:41,840 all the studies. What's the best way to prompt 707 00:33:41,960 --> 00:33:44,880 AI? And one of the big ways is by 708 00:33:44,880 --> 00:33:49,320 creating kind of autonomously thinking agents that work 709 00:33:49,320 --> 00:33:51,240 together, right? And you'll, and I don't know if 710 00:33:51,240 --> 00:33:53,640 you've seen stuff like that, but we're, rather than writing a 711 00:33:53,640 --> 00:33:56,360 script and telling the AI to act with a certain persona, you 712 00:33:56,360 --> 00:33:58,640 create almost a board of directors, right? 713 00:33:58,920 --> 00:34:02,000 So this one's responsible for finding the truth and having 714 00:34:02,000 --> 00:34:04,440 data backed evidence for why it found the answer. 715 00:34:04,720 --> 00:34:08,199 But Tideo has some that are just in charge of follow up 716 00:34:08,199 --> 00:34:11,840 questions. And like the the e-mail AI, 717 00:34:11,840 --> 00:34:14,520 there's an extra kind of additional mechanism that's 718 00:34:14,520 --> 00:34:19,600 looking at is this, is this likely to end up in a multi 719 00:34:19,800 --> 00:34:22,960 message conversation? And if yes, you know, then it, 720 00:34:23,120 --> 00:34:26,800 it acts completely different. So yeah, I hear a lot of people 721 00:34:26,800 --> 00:34:29,920 through that ask me that, you know, what's stopping me from 722 00:34:29,920 --> 00:34:32,840 building my own AI? Like I use ChatGPT, I use 723 00:34:32,840 --> 00:34:36,120 Claude, what's stopping me from going and programming in my own? 724 00:34:36,120 --> 00:34:39,239 It doesn't seem that difficult. And I think that like your 725 00:34:39,239 --> 00:34:42,600 experience that you just shared about how maybe there's one 726 00:34:42,600 --> 00:34:44,960 extra line you could have wrote in that prompt to avoid that 727 00:34:44,960 --> 00:34:48,000 hallucination. Or maybe there's you could have 728 00:34:48,000 --> 00:34:52,600 5X to the length of that prompt to create additional personas in 729 00:34:52,600 --> 00:34:55,040 that prompt to make sure that there was no hallucination. 730 00:34:55,880 --> 00:34:59,200 That's what really, really high tech people, I guess, think 731 00:34:59,200 --> 00:35:02,360 about versus like me and you, you know, the average end user 732 00:35:02,360 --> 00:35:04,440 of AI. I'm not, I've played around with 733 00:35:04,440 --> 00:35:06,160 image, not to go on a tangent, I've played around with image 734 00:35:06,160 --> 00:35:08,360 generation and I just can't ever get it to work. 735 00:35:08,360 --> 00:35:11,840 But it's because I don't know how to communicate effectively 736 00:35:11,840 --> 00:35:14,720 to yeah, the model that's generating my image. 737 00:35:14,800 --> 00:35:15,440 So. Yeah. 738 00:35:15,440 --> 00:35:19,360 So, so that talk that that talk I did a couple of weeks ago, I, 739 00:35:19,760 --> 00:35:21,560 I wanted to share a lot of information about a client, but 740 00:35:21,560 --> 00:35:23,040 I was like, I know they won't let me use their name and I 741 00:35:23,040 --> 00:35:24,520 don't want to use their name because I want to be really 742 00:35:24,520 --> 00:35:26,120 honest about some of the conversations we had and the 743 00:35:26,120 --> 00:35:28,240 assumptions we made. And actually some of their like 744 00:35:28,280 --> 00:35:31,360 both the amazing great uplifts in revenue we found where we 745 00:35:31,360 --> 00:35:33,760 didn't expect to and actually found out our customers are 746 00:35:33,840 --> 00:35:36,560 really boring. They care about warranty, they 747 00:35:36,560 --> 00:35:39,920 care about delivery prices. And and what they don't care 748 00:35:39,920 --> 00:35:42,280 about is what we thought was like this beautiful website 749 00:35:42,280 --> 00:35:45,320 design, like the more beautiful we made the website, the worse 750 00:35:45,320 --> 00:35:47,920 the conversion rate was. They were like, I just, I want a 751 00:35:47,920 --> 00:35:50,720 really good solid item. And it when we asked customers 752 00:35:50,720 --> 00:35:53,000 what they said, when I look at the image, I can just see it. 753 00:35:53,600 --> 00:35:55,760 And again, in that talk, because I wanted to share this 754 00:35:55,760 --> 00:35:59,160 information about this client of ours, I just invented Janet's 755 00:35:59,160 --> 00:36:01,920 Jumper Shop and just asked an AI give me an image of Janet's 756 00:36:01,920 --> 00:36:04,120 Jumper shop. I was like, that guy's got three 757 00:36:04,120 --> 00:36:06,160 arms. And this lady, she looks great, 758 00:36:06,160 --> 00:36:08,600 but her shoulder is like inside this guy's chest. 759 00:36:09,080 --> 00:36:12,240 And what have you done? And then I said, and now give me 760 00:36:12,240 --> 00:36:14,520 a shot of the same guy. He's doing some number crunching 761 00:36:15,120 --> 00:36:17,000 and they were just number crunching, but like suddenly 762 00:36:17,000 --> 00:36:19,360 they were in the warehouse instead of the office and there 763 00:36:19,360 --> 00:36:21,600 were balloons everywhere. And I was like, the balloons are 764 00:36:21,600 --> 00:36:24,120 cool. So I used this like humorously 765 00:36:24,120 --> 00:36:27,000 in the talk to say like, and look, this is what AI think 766 00:36:27,000 --> 00:36:30,040 Janet's jumper shop looks like. How how much fun are these guys 767 00:36:30,040 --> 00:36:33,120 having with their extra limbs? And this guy's got no chest, but 768 00:36:33,120 --> 00:36:34,880 you know, and here they are doing the number crunching. 769 00:36:34,880 --> 00:36:36,560 So you've got a number crunch guys, you need balloons. 770 00:36:36,720 --> 00:36:39,000 And it was like just kind of as I went through the talk making 771 00:36:39,000 --> 00:36:41,400 three points about this nameless client. 772 00:36:41,640 --> 00:36:43,680 And it was quite fun actually, just kind of adding those bits 773 00:36:43,680 --> 00:36:45,360 in. And, but I mean, AI, the guy's 774 00:36:45,360 --> 00:36:47,800 moving at a rate of knots. Like as you say, like even what 775 00:36:47,800 --> 00:36:49,760 you've just told me here, I'm sure everybody listening now is 776 00:36:49,880 --> 00:36:53,160 going to go, I'm going to get it to Fact Check itself and 5X the 777 00:36:53,160 --> 00:36:55,480 way it does tasks and stuff. That's important. 778 00:36:55,480 --> 00:36:57,240 But where's AI going in your opinion? 779 00:36:57,240 --> 00:36:59,480 And, and from a Tedio perspective, like 5 years, 10 780 00:36:59,480 --> 00:37:02,840 years, I know it's so far away and that AI kind of, you know, 781 00:37:02,840 --> 00:37:05,240 it's like a second lifetime of AI almost. 782 00:37:06,280 --> 00:37:12,080 Yeah, that's a big question and I can, I can only guess based on 783 00:37:12,080 --> 00:37:14,440 what I what I see. Get you back in five years to 784 00:37:14,440 --> 00:37:15,680 see if you're right. So no pressure but. 785 00:37:16,280 --> 00:37:18,520 Sure. Yeah, Yeah, I think speed in 786 00:37:18,520 --> 00:37:23,000 general and like a big push that we're seeing is maybe at first 787 00:37:23,000 --> 00:37:26,880 it was revolutionary if AI could answer a customer's question, 788 00:37:26,880 --> 00:37:31,880 generic repeatable questions. And a couple weeks ago, we've 789 00:37:31,880 --> 00:37:35,800 kind of pushed live globally the ability for the AI to 790 00:37:35,960 --> 00:37:39,280 conversationally reference API endpoints. 791 00:37:39,520 --> 00:37:42,160 And so there's just a fancy way of saying instead of telling me, 792 00:37:42,640 --> 00:37:45,320 hey, to check your order status, step #1 log into your account, 793 00:37:45,320 --> 00:37:48,720 step #2 do this where it is resolving my issue and telling 794 00:37:48,720 --> 00:37:51,440 me how to help myself, Now it just checks it for me, right? 795 00:37:51,480 --> 00:37:55,320 It just, it'll, it'll use that endpoint in order to check that 796 00:37:55,320 --> 00:37:58,160 status. Yeah, immediately. 797 00:37:58,160 --> 00:38:02,240 And I think that in the next 5 years also to beyond having the 798 00:38:02,400 --> 00:38:06,520 AI connected to your systems, the systems are more connected 799 00:38:06,520 --> 00:38:10,920 than ever as well. So I've seen what I would have, 800 00:38:11,200 --> 00:38:14,680 you know, a year ago, I've seen use cases where, you know, a 801 00:38:14,680 --> 00:38:17,160 year ago, yeah, I would have said that's, look, AI is really 802 00:38:17,160 --> 00:38:19,160 cool and it's super powerful. But if I can't do things like 803 00:38:19,160 --> 00:38:21,480 that. And now I'll give you an example 804 00:38:21,480 --> 00:38:24,880 like a retention based use case where we are checking the 805 00:38:24,920 --> 00:38:27,640 average customer value, we're checking the subscription tier 806 00:38:27,640 --> 00:38:30,720 and we're checking how long they've been a customer in order 807 00:38:30,720 --> 00:38:35,720 to determine how the AI is going to process churn attempt, right? 808 00:38:36,000 --> 00:38:40,080 Whereas not that long ago we could not do, we could not do 809 00:38:40,080 --> 00:38:42,080 things like that. So now it can can check like, 810 00:38:42,080 --> 00:38:44,560 hey, this is a medium value customer. 811 00:38:44,560 --> 00:38:46,320 They've been with our service for a year. 812 00:38:46,480 --> 00:38:48,480 They're already on a discounted subscription. 813 00:38:48,480 --> 00:38:51,720 So don't offer a discount, don't offer this program because it's 814 00:38:51,720 --> 00:38:53,880 not relevant to their use case, etcetera, etcetera. 815 00:38:53,880 --> 00:38:57,240 And then it's going down different resolution paths based 816 00:38:57,240 --> 00:39:01,520 on, yeah, the, the information that's housed inside of the 817 00:39:01,520 --> 00:39:04,760 subscription platform itself. So not even in the data sources 818 00:39:04,760 --> 00:39:08,120 of the AI. And so I think that's that's 819 00:39:08,120 --> 00:39:10,320 where all of this is going. Five years, though, would you 820 00:39:10,320 --> 00:39:12,400 say 10 years? I mean, people aren't going to 821 00:39:12,400 --> 00:39:15,680 type anymore. I can tell you that like it, the 822 00:39:15,840 --> 00:39:19,320 AI is so fast that it'll probably take me twice as long 823 00:39:19,320 --> 00:39:22,320 to type my question than it will for the AI to go reference a 824 00:39:22,320 --> 00:39:24,480 bunch of values in the internal system, change my shipping 825 00:39:24,480 --> 00:39:26,760 address, and then let me know it's been completed. 826 00:39:26,760 --> 00:39:29,880 So I don't know where we get with that, right? 827 00:39:29,880 --> 00:39:32,920 I see people doing cool stuff with like virtual agents. 828 00:39:33,440 --> 00:39:35,920 A lot of people are interested in AI voice and things like 829 00:39:35,920 --> 00:39:37,200 that. Yeah, yeah. 830 00:39:37,440 --> 00:39:40,400 I'm less interested in. My car it's built into Audi now 831 00:39:40,480 --> 00:39:42,400 so I just in my car. So if anyone was in the car 832 00:39:42,400 --> 00:39:45,040 right now this might work. Hey, Audi, subscribe to winning 833 00:39:45,040 --> 00:39:48,480 Shopify podcast and then ChatGPT might go and subscribe to the 834 00:39:48,480 --> 00:39:51,000 podcast for you. And but yeah, I asked I asked it 835 00:39:51,000 --> 00:39:53,800 a question today to show someone how rubbish the Audi AI like 836 00:39:53,800 --> 00:39:57,000 chat feature was. It's rubbish like Siri, Alexa 837 00:39:57,000 --> 00:39:59,520 and Google. So to show them and it came back 838 00:39:59,520 --> 00:40:02,280 saying according to and the way it talks is rubbish still. 839 00:40:02,280 --> 00:40:05,200 It was like very broken up, but it said according to ChatGPT. 840 00:40:05,440 --> 00:40:08,200 I think we're asking what a macchiato was in a drive through 841 00:40:08,200 --> 00:40:09,920 on the way to a client meeting at Starbucks. 842 00:40:10,200 --> 00:40:12,600 So it was like, yeah, someone was like, what's a macchiato? 843 00:40:12,600 --> 00:40:14,080 Like I see, I don't actually know. 844 00:40:14,080 --> 00:40:14,920 I see Everest. We asked. 845 00:40:14,920 --> 00:40:16,720 The AI was like 2 macchiatos please. 846 00:40:16,880 --> 00:40:19,200 This sounds great. And it was, yeah, just asking a 847 00:40:19,200 --> 00:40:21,480 little things like that to suddenly have it in the car. 848 00:40:21,640 --> 00:40:24,920 But I've been chatting to ChatGPT a little bit and it, it 849 00:40:24,920 --> 00:40:26,720 asks you way more questions when you talk to it. 850 00:40:26,720 --> 00:40:29,560 When you type, you ask something and it gives it to you when you 851 00:40:29,560 --> 00:40:31,080 talk. I don't know if it's ChatGPT 852 00:40:31,080 --> 00:40:33,160 actually or Grok or manners. They're the three I've been 853 00:40:33,160 --> 00:40:34,600 using the most. But I started having a 854 00:40:34,600 --> 00:40:36,680 conversation with it. I'm pretty sure it's ChatGPT and 855 00:40:36,680 --> 00:40:38,640 it start, I sort of said, you know, I need to get an outfit 856 00:40:38,640 --> 00:40:40,880 ready for a wedding. And it was like, OK, what's your 857 00:40:40,880 --> 00:40:42,760 relation to this person? What's the colour scheme of the 858 00:40:42,760 --> 00:40:43,880 wedding? And started asking questions 859 00:40:43,880 --> 00:40:46,480 like this is cool. Like, you know, asking, whereas 860 00:40:46,800 --> 00:40:49,760 previously I would have to give it all of that, say all right 861 00:40:49,760 --> 00:40:52,520 now based on it being a lilac themed wedding, what is the 862 00:40:52,840 --> 00:40:55,880 colour code for that? I'm saying if my wife listens to 863 00:40:55,880 --> 00:40:57,480 this, I'm saying really high expectations for this wedding. 864 00:40:57,480 --> 00:41:00,360 We're going to in a few weeks and we've got a little, a little 865 00:41:00,360 --> 00:41:02,560 baby who's only going to be sort of three or four months old. 866 00:41:02,600 --> 00:41:03,920 That should be, yeah, three months. 867 00:41:03,920 --> 00:41:06,160 So she's probably going to be sick on whatever I'm wearing. 868 00:41:06,160 --> 00:41:09,680 So that's also a consideration. Something that doesn't absorb 869 00:41:09,680 --> 00:41:12,120 sick would be amazing. But anyway, look, it's been 870 00:41:12,120 --> 00:41:15,040 great to have you on. As I said right at the start, 871 00:41:15,040 --> 00:41:20,800 guys, if you want to get 20% off, go to tidiotidio.com/WWS. 872 00:41:21,000 --> 00:41:22,520 These guys are going to be around for a little while. 873 00:41:22,880 --> 00:41:25,400 Jacob, I want to say a massive thank you for your time. 874 00:41:25,600 --> 00:41:28,640 If people want to follow you guys on social or see what's 875 00:41:28,640 --> 00:41:30,960 going along, and rather than just going to that landing page, 876 00:41:30,960 --> 00:41:32,880 what are your social handles? How can people reach out to you? 877 00:41:33,560 --> 00:41:38,040 Yeah, TIDIO, find us on LinkedIn, find us on YouTube if 878 00:41:38,040 --> 00:41:40,680 you want to see back end of the actual platform. 879 00:41:41,640 --> 00:41:44,760 And then I would recommend, yeah, everyone should go to the 880 00:41:44,760 --> 00:41:48,840 forward slash WWS and actually see what it does with your own 881 00:41:48,840 --> 00:41:52,880 information because everything's hypothetical until you plug in 882 00:41:52,880 --> 00:41:54,760 your own information. And that's where you're going to 883 00:41:54,760 --> 00:41:57,960 have the most aha moments. And that's, I should have said 884 00:41:57,960 --> 00:41:59,920 right at the start as well, guys, that's, that's the part of 885 00:41:59,920 --> 00:42:02,800 the free bit as well, isn't it? So as a free demo, they can plug 886 00:42:02,800 --> 00:42:05,960 their information in and it can start telling them, I mean, tell 887 00:42:05,960 --> 00:42:08,680 me more about it, what it can do and and how that kind of 888 00:42:08,680 --> 00:42:11,240 process, that initial step, that free bit actually works. 889 00:42:11,920 --> 00:42:13,560 Yeah, yeah. So it's, it's essentially you're 890 00:42:13,560 --> 00:42:17,360 plugging in just your homepage URL or any URL that that you 891 00:42:17,360 --> 00:42:19,720 want to really. And since customers don't read, 892 00:42:19,880 --> 00:42:23,240 what the AI is doing is it's reading all of these sub URLs 893 00:42:23,240 --> 00:42:26,120 and crawling the website for any text based content. 894 00:42:26,120 --> 00:42:29,160 It is reading it. And then it's creating as many 895 00:42:29,160 --> 00:42:32,560 QA pairs as it can possibly think of about that information 896 00:42:32,880 --> 00:42:37,080 and allowing you to go into a playground and and test exactly 897 00:42:37,080 --> 00:42:39,680 what that AI speaks like. So I would encourage anyone to 898 00:42:40,920 --> 00:42:43,080 build it and then go try and break it, right. 899 00:42:43,080 --> 00:42:45,720 If if you're worried about it hallucinating or answering 900 00:42:45,720 --> 00:42:48,800 questions in a certain way, bring it on your own data and 901 00:42:48,800 --> 00:42:51,240 and go see it for yourself. Yeah. 902 00:42:52,440 --> 00:42:54,440 Cool, I'm certainly going to do that with a couple of my clients 903 00:42:54,440 --> 00:42:56,440 sites and see what happens over the next next few days. 904 00:42:56,440 --> 00:42:58,560 So if you want to see how that goes, guys, drop me an e-mail 905 00:42:58,560 --> 00:43:00,560 and I'll let I'll let you know. But yeah, definitely check that 906 00:43:00,560 --> 00:43:01,760 out. And again, you get 20% off 907 00:43:01,760 --> 00:43:03,240 through us, which is really generous to these guys. 908 00:43:03,240 --> 00:43:07,600 So titio.com/WWS and Jacob, a massive thank you for your time, 909 00:43:07,600 --> 00:43:10,280 your energy, all your answers and giving up so much to come 910 00:43:10,280 --> 00:43:13,200 and join us today. So thank you so much. 100% thank 911 00:43:13,200 --> 00:43:15,400 you truly it was fun. Amazing. 912 00:43:15,680 --> 00:43:18,200 And for everybody listening at home, hit that subscribe button. 913 00:43:18,560 --> 00:43:20,400 Leave us a review. Five stars. 914 00:43:20,400 --> 00:43:23,040 If you're not happy, reach out. I'd love to get some feedback. 915 00:43:23,040 --> 00:43:25,640 Haters are going to hate and some of them do e-mail me from 916 00:43:25,640 --> 00:43:27,600 time to time. The most recent one, which I 917 00:43:27,600 --> 00:43:30,600 thought was really, really funny is it said stop making up well 918 00:43:30,600 --> 00:43:33,160 known sayings. So I asked for a few and sent 919 00:43:33,160 --> 00:43:36,400 them back some Google links to say this is not my saying but 920 00:43:36,400 --> 00:43:38,560 that one is and I will claim that from here on. 921 00:43:38,560 --> 00:43:41,280 So yeah, if you've got any negative feedback, bring it on. 922 00:43:41,680 --> 00:43:43,520 I've got plenty of spare time at the moment, awake in the middle 923 00:43:43,520 --> 00:43:44,920 of night feeding my lovely little daughter. 924 00:43:45,160 --> 00:43:48,120 So if you haven't tuned in before, I hope you've enjoyed 925 00:43:48,120 --> 00:43:50,000 today's show. We post every single Friday and 926 00:43:50,000 --> 00:43:51,640 we hope you can join us again next week. 927 00:43:51,920 --> 00:43:52,600 Take care everybody.