WEBVTT 00:00:00.200 --> 00:00:03.300 With a lot of hype and negative headlines around Chad CPT and 00:00:03.300 --> 00:00:04.200 other AI tools. 00:00:04.200 --> 00:00:07.100 We started wondering whether there's actual value in the technology 00:00:07.100 --> 00:00:10.000 for companies on this episode of today and take her to find 00:00:10.000 --> 00:00:11.900 out what businesses are really doing today. 00:00:18.600 --> 00:00:19.600 Hi everyone, I'm cute. 00:00:19.600 --> 00:00:20.700 Y'all. Welcome to today. 00:00:20.700 --> 00:00:21.800 And Tech is joining me on the show. 00:00:21.800 --> 00:00:24.100 Today is AJ Mejia and he is a principal of a, 00:00:24.100 --> 00:00:26.700 i in analytic app, Gemini, welcome AJ. 00:00:28.200 --> 00:00:30.900 Thanks. Happy to be on you guys today, 00:00:30.900 --> 00:00:35.700 and I research institute recently published a report about 00:00:35.700 --> 00:00:39.500 how companies are deploying Jenner Fai tools, 00:00:39.500 --> 00:00:40.900 such as chat gvt, others. 00:00:40.900 --> 00:00:45.500 Tell us about sort of the motivation behind the report and was 00:00:45.500 --> 00:00:47.600 there anything that surprise you about the results? 00:00:47.600 --> 00:00:51.500 Cuz when I went through the report there were a couple of surprising 00:00:51.500 --> 00:00:53.400 means to me but I want to hear what you guys thought about 00:00:53.400 --> 00:00:53.700 it. 00:00:55.100 --> 00:00:58.700 Yeah. So you're going to the motivation behind a reporter. 00:00:58.700 --> 00:01:03.700 I think it's fairly obvious especially since you know, 00:01:03.700 --> 00:01:08.700 probably around November or December of last year interest 00:01:08.700 --> 00:01:12.500 in dinner. Today I exploded on the market. 00:01:12.500 --> 00:01:16.400 You know, there is a number of tools that were released. 00:01:16.400 --> 00:01:21.900 It started the number of public facing our products or at least 00:01:21.900 --> 00:01:25.600 said you're really kind of captured a lot of people's imaginations 00:01:25.600 --> 00:01:30.800 and we found that a lot of our conversations were being driven 00:01:30.800 --> 00:01:43.500 by a part of the survey. 00:01:43.500 --> 00:01:48.100 We conducted a survey of 1,000 organizations across the globe. 00:01:48.100 --> 00:01:53.500 You know, we talked about their exact exact Industries. 00:01:55.100 --> 00:02:02.300 Retail Financial Service here is to get an industry-wide 00:02:02.300 --> 00:02:07.100 perspective on, you know, how these Executives. 00:02:07.100 --> 00:02:13.200 So, you know, obviously, you know, obviously, 00:02:13.200 --> 00:02:15.400 we'd love to work with them on that Journey. 00:02:15.400 --> 00:02:20.100 But really important just also showing some of the commonality 00:02:20.100 --> 00:02:32.000 feel like that there was a very high number of optimism 00:02:32.000 --> 00:02:35.200 around it, and sometimes you would think that their skepticism 00:02:35.200 --> 00:02:38.600 around to a new technology especially one that can be kind of 00:02:38.600 --> 00:02:40.800 polarizing with with Jenna with Jenna and I 00:02:42.100 --> 00:02:54.500 Yeah. So that is absolutely no, that's probably all of them 00:02:54.500 --> 00:02:57.200 are adding some level of conversation and there, 00:02:57.200 --> 00:03:01.600 you know, executive meetings or boarding rooms around nari, 00:03:01.600 --> 00:03:16.800 but your other point just the executives and their process, 00:03:16.800 --> 00:03:21.500 or or Revenue driver was really surprising. 00:03:21.500 --> 00:03:26.200 And, you know, usually when you talk about new approaches 00:03:26.200 --> 00:03:31.200 for AI machine learning, you don't get that sort of widespread 00:03:31.200 --> 00:03:33.100 in a positivity. 00:03:33.100 --> 00:03:37.400 Certainly, certain industries, might be more positive than just 00:03:37.400 --> 00:03:41.600 a tremendous amount of excitement. 00:03:42.000 --> 00:03:45.400 Is really the best way to put it in terms of how generative 00:03:45.400 --> 00:03:55.100 AI every aspect of the company's business card about sort of the 00:03:55.100 --> 00:03:59.200 overall part of the, the survey was that many of them felt that 00:03:59.200 --> 00:04:02.200 it was not going to be disruptive to their industry. 00:04:02.200 --> 00:04:06.600 And that made me think, well, it feels like everything 00:04:06.600 --> 00:04:09.500 that you read it is going to be disruptive. 00:04:09.500 --> 00:04:13.800 And so I'm I was wondering whether you felt from the responses 00:04:13.800 --> 00:04:16.700 that the executive might not have understood. 00:04:16.700 --> 00:04:21.900 What they were disruptive means, or do they just sort of put 00:04:21.900 --> 00:04:24.800 that in the corner seeing that like they see almost like the 00:04:24.800 --> 00:04:27.500 gold chest but they don't see the dragon next to it. 00:04:27.500 --> 00:04:31.300 Either that was sort of the analogy, 00:04:31.300 --> 00:04:35.400 I came up with it, are they just ignoring sort of maybe the disruption 00:04:35.400 --> 00:04:38.100 part or or do they not understand what the definition 00:04:38.100 --> 00:04:39.800 of disruptive meets cuz it could be that too. 00:04:40.800 --> 00:04:46.300 Yeah, I didn't think they're ignoring the concept of Destruction. 00:04:46.300 --> 00:04:55.300 I think that a lot of these Executives and then they look at it 00:04:55.300 --> 00:04:58.600 as enhancements to their existing business. 00:04:58.600 --> 00:05:06.800 How do you optimize existing operational blows a disruptor 00:05:06.800 --> 00:05:10.200 in the sense of replacing entire areas of business? 00:05:10.200 --> 00:05:17.000 But they do, you know, obviously, obviously the future that 00:05:17.000 --> 00:05:21.200 could happen in the future in certain industries and then what 00:05:21.200 --> 00:05:25.900 not? But really the key takeaway is there looking at it as sort of 00:05:25.900 --> 00:05:30.600 optimizers to their existing work clothes as opposed to a disrupter 00:05:30.600 --> 00:05:33.100 of bearer of their specific business? 00:05:40.800 --> 00:05:45.500 Dining in sort of the questions that you ask them in terms 00:05:45.500 --> 00:05:49.900 of what areas of generative AI are the most relevant to visit. 00:05:49.900 --> 00:05:52.500 So, tell me about the top answers where most businesses 00:05:52.500 --> 00:05:56.900 to see them using generative AI or GPT tools. 00:05:56.900 --> 00:06:08.300 Absolutely. So right now certified does the most sure I understood 00:06:08.300 --> 00:06:09.700 use cases of dinner today. 00:06:09.700 --> 00:06:16.700 I'll say is things like potentially potentially marketing 00:06:16.700 --> 00:06:21.500 marketing and customer insights and personalization 00:06:21.500 --> 00:06:24.500 around the generating content. 00:06:24.500 --> 00:06:28.800 A lot of clients are interested in areas like customer service. 00:06:28.800 --> 00:06:32.700 Offered ample having you do more, Elgin virtual assistant, 00:06:41.200 --> 00:06:42.300 I climb inside. 00:06:42.300 --> 00:06:55.700 Yeah, and I certainly the whole area code where a lot of product 00:06:55.700 --> 00:07:01.700 companies are development, companies are interested in the workload, 00:07:01.700 --> 00:07:07.500 their developers. When you look at the initial scope of use cases, 00:07:07.500 --> 00:07:23.600 experience for customer experience Marketing in a lot of cases 00:07:23.600 --> 00:07:28.400 of murdering Knowledge Management, which is, which is essentially 00:07:28.400 --> 00:07:32.300 a graffiti, the content across organization. 00:07:32.300 --> 00:07:32.500 Yeah. 00:07:35.100 --> 00:07:38.700 Find podcast about digging through a different site or different 00:07:38.700 --> 00:07:44.800 different portals. And then there's this other sort of back 00:07:44.800 --> 00:07:50.300 and use case online coding and decoding optimization 00:07:50.300 --> 00:07:54.200 things like user acceptance, testing, that sort of thing. 00:07:54.200 --> 00:08:00.600 It was interesting in in the survey that coding generated 00:08:00.600 --> 00:08:01.300 about 50% 00:08:01.600 --> 00:08:04.900 of interest from those surveyed compared with higher numbers 00:08:04.900 --> 00:08:09.100 for things like chatbots data, text search that kind of sucked. 00:08:09.100 --> 00:08:13.200 It does feel like they're only after the low-hanging 00:08:13.200 --> 00:08:16.700 for right now and maybe coding, there are still some questions 00:08:16.700 --> 00:08:21.400 about accuracy and maybe time to develop a program like that. 00:08:21.400 --> 00:08:23.200 And a lot of companies will they say, 00:08:23.200 --> 00:08:27.300 well, it's easier to just grab the chatbot stuff quickly 00:08:27.300 --> 00:08:30.200 and then the search cuz there's so much out there. 00:08:30.200 --> 00:08:34.700 Where else it might be a longer-term project for coding and things 00:08:34.700 --> 00:08:37.200 like that. Or, or do you think that there might be some doubt 00:08:37.200 --> 00:08:41.900 from companies about its ability to sort to replace coders 00:08:41.900 --> 00:08:45.400 or not? Replace but yeah, it has coding capabilities. 00:08:47.000 --> 00:08:50.700 Find it. There's I think there's always going to be awesome 00:08:50.700 --> 00:08:52.700 when you're talkin about mission-critical applications 00:08:52.700 --> 00:08:54.200 of your business, right? 00:08:54.200 --> 00:08:59.500 You're a product company for example of your you're obviously 00:08:59.500 --> 00:09:08.300 you want to be very careful about ensuring you know you not taking 00:09:08.300 --> 00:09:10.900 any shortcuts that sort of thing. 00:09:10.900 --> 00:09:25.900 There's always going to be that we don't want to just dive right 00:09:25.900 --> 00:09:28.600 into it because this is such a critical part of our operations 00:09:28.600 --> 00:09:38.100 here, point of their their relatively straightforward 00:09:38.100 --> 00:09:43.600 to be implemented for work but also but also there's already 00:09:43.600 --> 00:09:44.200 some 00:09:46.000 --> 00:09:50.600 Damonza, demonstrated applications of it in those areas. 00:09:50.600 --> 00:10:06.500 So that's why I think I can see that being more of a, 00:10:06.500 --> 00:10:09.800 just taking it a little bit slower to make sure that their 00:10:09.800 --> 00:10:12.300 primary business isn't getting the truck. 00:10:12.300 --> 00:10:16.600 You know, he's not disrupted but like influence negatively, 00:10:16.600 --> 00:10:24.700 by by just kind of throwing dinner to mail sample of something. 00:10:24.700 --> 00:10:25.800 It scored high on the list. 00:10:25.800 --> 00:10:29.200 In terms of what, what companies want to do it Saturday, 00:10:29.200 --> 00:10:32.500 I am that's in the generation of synthetic data. 00:10:32.500 --> 00:10:35.100 Can you explain what that that is? 00:10:35.100 --> 00:10:39.000 And as I was doing some more research on it and there was some 00:10:39.000 --> 00:10:41.000 other red flags that kind of popped up. 00:10:41.000 --> 00:10:43.700 I want to talk to you a little bit about that tell, 00:10:43.700 --> 00:10:45.900 tell me about what it is for AIG. 00:10:46.100 --> 00:10:49.200 Creating synthetic data cuz as soon as you tell someone that 00:10:49.200 --> 00:10:51.600 they got, that doesn't sound good. 00:10:51.600 --> 00:10:54.500 So tell me why there's such an interest there. 00:10:55.300 --> 00:10:59.300 Yeah, so the concept of synthetic data generation has actually 00:10:59.300 --> 00:11:03.800 been around for decades, it's a way to effectively training 00:11:03.800 --> 00:11:07.700 or tough models that you develop in order. 00:11:07.700 --> 00:11:15.400 And you know where there might not be a detective are classifying 00:11:15.400 --> 00:11:20.000 a certain data set of generating. 00:11:20.000 --> 00:11:31.600 Synthetic data has been around for a long time and you're 00:11:31.600 --> 00:11:39.800 using an area model to to generate data that's very similar 00:11:39.800 --> 00:11:46.700 to from like a structure and to actually do but not exactly 00:11:46.700 --> 00:11:55.000 the same. And once again that's usually a lot of people. 00:11:55.300 --> 00:12:01.200 Obviously getting much more data to be able to train a nester 00:12:01.200 --> 00:12:08.500 do models but also potentially hockey seedings real world as much 00:12:08.500 --> 00:12:11.300 as possible. But you don't exactly. 00:12:11.300 --> 00:12:22.700 Those are usually where we see the synthetic data generated 00:12:22.700 --> 00:12:36.000 imagery as well as structured and unstructured data. 00:12:36.000 --> 00:12:41.000 And so that's where that's where a lot of clients are also looking 00:12:41.000 --> 00:12:42.700 at it. I think they look at it from their perspective 00:12:42.700 --> 00:12:47.700 of. I don't have enough data for 4 actually building like an effective 00:12:47.700 --> 00:12:47.900 model. 00:12:55.300 --> 00:12:58.000 You're at the old fashion way it is there, 00:12:58.000 --> 00:13:01.700 a danger, though of companies that invest into synthetic 00:13:01.700 --> 00:13:05.500 data and then use that data to Dent rain there a II, 00:13:05.500 --> 00:13:08.700 and it gets a new situation where it's almost like you're 00:13:08.700 --> 00:13:12.400 photocopying a photocopy in a photocopy and it gets to the point 00:13:12.400 --> 00:13:13.600 where it gets muddled. 00:13:13.600 --> 00:13:16.800 There was an article in this week that I saw about that sort of 00:13:16.800 --> 00:13:18.200 a fact that has as a, 00:13:18.200 --> 00:13:21.100 i starts training itself on AI generated data. 00:13:21.100 --> 00:13:23.000 And this might be in the image space, 00:13:23.000 --> 00:13:25.400 or the video States space and eventually after like, 00:13:25.400 --> 00:13:28.500 5 or 6 in a relations, you just get something that is a blob 00:13:28.500 --> 00:13:33.900 of a mess, and I'm hoping that companies understand the proper 00:13:33.900 --> 00:13:35.700 use of synthetic K2, or is it just a? 00:13:35.700 --> 00:13:36.400 Is it though? 00:13:36.400 --> 00:13:40.700 Just a fact that synthetic means fake or means made up and then 00:13:40.700 --> 00:13:42.600 might be throwing some fear into this, 00:13:42.600 --> 00:13:44.700 this situation to set me on the right. 00:13:44.700 --> 00:13:45.400 Course, I guess. 00:13:45.400 --> 00:13:46.200 I'm asking you. 00:13:48.100 --> 00:13:53.400 No cursor. So you're in terms of what you pointed out, 00:13:53.400 --> 00:13:56.800 a copy of the photo copy Concepts. 00:13:56.800 --> 00:14:06.800 Absolutely no crap without the proper guardrail or word that 00:14:06.800 --> 00:14:11.100 can happen if you just sort of Leverage Center to be out of that 00:14:11.100 --> 00:14:12.100 understanding. 00:14:14.800 --> 00:14:20.000 I think, you know, there is, there is also a lot of concern 00:14:20.000 --> 00:14:21.200 around things like that. 00:14:21.200 --> 00:14:24.600 If you aren't using your generating synthetic database, 00:14:24.600 --> 00:14:29.700 on some real-world data, you have the rest of always pulling 00:14:29.700 --> 00:14:36.900 in by assholes, pulling in pulling in like normal, 00:14:36.900 --> 00:14:44.000 but that there is a lot of sort of days at checkpoints 00:14:44.000 --> 00:14:52.100 that you have to consider when generating synthetic where you're 00:14:52.100 --> 00:14:54.000 in a potentially looking at, you know, 00:14:54.000 --> 00:14:57.600 something a little bit more objective with processor like a, 00:14:57.600 --> 00:15:00.500 you know, you're measuring Bowl. 00:15:00.500 --> 00:15:07.300 Li coyote measurements are to fill your mechanical Parts in example. 00:15:07.300 --> 00:15:10.500 Yep. But then there's other places where they are looking 00:15:10.500 --> 00:15:14.300 at either subjective data or more make fun of Misha. 00:15:14.800 --> 00:15:21.500 Applications of those would be, you would be careful about 00:15:21.500 --> 00:15:26.400 everything that I paid him, because he can absorb, 00:15:26.400 --> 00:15:33.000 you know, potentially, those Notions or any other. 00:15:33.000 --> 00:15:39.900 I can just pass it off as, 00:15:39.900 --> 00:15:41.700 as a normal aeration update. 00:15:41.700 --> 00:15:46.200 Yeah, yeah, you know, you mentioned risk a couple seconds 00:15:46.200 --> 00:15:49.200 ago and there's another part of the survey that I wanted to talk 00:15:49.200 --> 00:15:53.500 about was how it looks to me as if the companies were down playing 00:15:53.500 --> 00:15:55.500 a lot of the risks associated with Jen. 00:15:55.500 --> 00:15:58.800 I one of the questions was that 74% 00:15:59.100 --> 00:16:01.100 said, they believed the benefits of of a, 00:16:01.100 --> 00:16:04.100 i outweigh the rest and I wanted to ask you, 00:16:04.100 --> 00:16:07.800 is this a case where companies might not understand a bunch 00:16:07.800 --> 00:16:08.400 of the risks? 00:16:08.400 --> 00:16:11.300 Like blood things, you just brought up with bias and potential 00:16:11.300 --> 00:16:14.400 love of data, getting more fit into this weird. 00:16:14.800 --> 00:16:19.400 Table type thing or do you feel that companies might be there 00:16:19.400 --> 00:16:23.400 still confident that the risks will be minimize were solved 00:16:23.400 --> 00:16:25.100 down the road or at least, you know, 00:16:25.100 --> 00:16:28.500 they understand the risks and they're just confident of the benefits. 00:16:28.500 --> 00:16:30.900 You know what I'm saying? 00:16:30.900 --> 00:16:32.200 It's is it, is it one or the other, 00:16:32.200 --> 00:16:33.400 or is it a mixture of both? 00:16:34.600 --> 00:16:36.500 Yeah, I think it's a little bit of the way, 00:16:36.500 --> 00:16:39.800 it's more of the latter, you know, 00:16:39.800 --> 00:16:50.200 how our process after the Beast be aware of and when track 00:16:50.200 --> 00:16:53.600 for I think most companies and executives are aware. 00:16:53.600 --> 00:17:00.400 Otherwise it would just be throwing a lot of I think the general 00:17:00.400 --> 00:17:04.000 sense is positivity that you know, it'll get mitigated 00:17:04.000 --> 00:17:13.000 as not just me but also as work closable as they're able to more 00:17:13.000 --> 00:17:15.700 effectively leverage, this in our organization. 00:17:17.200 --> 00:17:20.900 I think they're counting on actually implementing those guardrails 00:17:20.900 --> 00:17:22.600 implementing those stuff, the copter. 00:17:22.600 --> 00:17:24.900 And so it's one of those issues where, 00:17:24.900 --> 00:17:42.800 you know, it's it's a it's a new technology in general, 00:17:42.800 --> 00:17:44.000 but also 10 a.m. 00:17:44.000 --> 00:17:50.100 and it's expecting that as it evolved over the next several 00:17:50.100 --> 00:17:56.100 years. Those are going to get mitigated by Hyderabad model. 00:17:56.100 --> 00:18:00.500 And another point is a survey that I thought was very interesting 00:18:00.500 --> 00:18:05.200 was companies are recognizing the Jhene. 00:18:05.200 --> 00:18:09.100 I might bump up against other efforts in the company, 00:18:09.100 --> 00:18:13.200 especially around sustainability, because AI projects have such 00:18:13.200 --> 00:18:14.400 a big carbon footprint. 00:18:14.400 --> 00:18:16.100 It almost feels like there's going to be 00:18:17.200 --> 00:18:20.000 Bumping heads, where you do. 00:18:20.000 --> 00:18:23.300 I j i project my go up against the sustainability 00:18:23.300 --> 00:18:28.000 project in. Can those two factors work together or is it going 00:18:28.000 --> 00:18:29.600 to be a case of War, he goes, 00:18:29.600 --> 00:18:32.800 well, we were getting so much benefit from this gen II stuff 00:18:32.800 --> 00:18:37.700 that we may have to reduce our our climate goals or resg goals or 00:18:37.700 --> 00:18:41.000 anything like that for sustainability that that might go off 00:18:41.000 --> 00:18:41.500 to the side. 00:18:43.300 --> 00:18:45.900 No. But they have to work together. 00:18:45.900 --> 00:18:58.500 The, when we were talking about a whole Suite of tools and solutions 00:18:58.500 --> 00:19:03.500 that are out there on the market and there's there's many ways. 00:19:03.500 --> 00:19:06.600 Can you try and reduce the carbon footprint? 00:19:06.600 --> 00:19:08.500 Apple, Sam's Barber. 00:19:08.500 --> 00:19:13.600 It's using you no more of a foundational model and training. 00:19:13.600 --> 00:19:17.600 On a smaller data sides and using less competition powers 00:19:17.600 --> 00:19:18.500 of that. Sort of thing. 00:19:18.500 --> 00:19:23.600 There is a lot of weight you can you can help mitigate the cost 00:19:23.600 --> 00:19:25.800 by creating more optimize Center. 00:19:25.800 --> 00:19:26.800 They are models. 00:19:29.000 --> 00:19:33.100 Do you know it's not her sense of it so it's one better today. 00:19:33.100 --> 00:19:41.300 Are they somewhere? 00:19:41.300 --> 00:19:46.400 You have a suite of tools to use your other different products 00:19:46.400 --> 00:19:49.900 and different companies and you have to pick and choose the right 00:19:49.900 --> 00:19:54.300 one for your specific space and I can come down. 00:19:54.300 --> 00:19:57.100 You obviously cost is going to be a huge consideration. 00:19:57.100 --> 00:20:00.700 But also, you know, things like arguing, 00:20:00.700 --> 00:20:05.000 you know, whether you need a real talking to me where you want 00:20:05.000 --> 00:20:07.800 to happen. There's a lot of different considerations 00:20:07.800 --> 00:20:12.300 along bags that you can absolutely be. 00:20:12.300 --> 00:20:18.200 You'll figure it into a cigarette into your system ability 00:20:18.200 --> 00:20:21.600 goals or at least minimizing the footprint of generator. 00:20:21.600 --> 00:20:23.600 They are model without the saying all. 00:20:23.600 --> 00:20:24.800 Well if I'm not willing to 00:20:25.600 --> 00:20:27.000 I can't use dinner today. 00:20:27.000 --> 00:20:30.400 I'm not willing to sacrifice my sustainability. 00:20:30.400 --> 00:20:37.000 Don't think it's a it's a black-and-white very impressed 00:20:37.000 --> 00:20:41.000 with the fact that you guys ask that question or that there 00:20:41.000 --> 00:20:46.800 was enough of an awareness out there you know the cost of a some 00:20:46.800 --> 00:20:49.700 of these Jhene Aiko projects that that that would have an impact 00:20:49.700 --> 00:20:53.600 height. I think sometimes maybe companies might ignore some 00:20:53.600 --> 00:20:57.100 of that it in the in the sense of what we're still going to 00:20:57.100 --> 00:21:02.200 do our sustainability efforts just ignore the cost of the carbon 00:21:02.200 --> 00:21:03.000 footprint over here. 00:21:03.000 --> 00:21:07.000 Then a lot of companies are at least recognizing that I want to 00:21:07.000 --> 00:21:09.600 switch gears a little bit and ask you about you. 00:21:09.600 --> 00:21:12.200 You've done a lot of work at capgemini with a lot of these 00:21:12.200 --> 00:21:16.000 companies employing some of these AI projects without I'm not 00:21:16.000 --> 00:21:18.500 going to ask you specific projects or product names are companies 00:21:18.500 --> 00:21:20.600 like that. But can you give me a sense of some of the examples 00:21:20.600 --> 00:21:23.300 of where companies are seeing some success? 00:21:23.300 --> 00:21:25.300 Give me some examples of projects that 00:21:25.600 --> 00:21:27.000 Yeah, that are actually working. 00:21:28.100 --> 00:21:33.200 Yeah. So you know exactly speaking of broccoli received 00:21:33.200 --> 00:21:36.900 areas were received no interest in these companies, 00:21:36.900 --> 00:21:41.700 as well as where they're investing some Knowledge Management. 00:21:41.700 --> 00:21:46.100 Like you mentioned has a huge area where you know when we know 00:21:46.100 --> 00:21:47.800 a lot of companies are working on, 00:21:47.800 --> 00:21:54.000 it could be one specific business. 00:21:54.000 --> 00:22:00.100 I'm just giving an example like looking at actually like a user 00:22:00.100 --> 00:22:04.300 portal, like me looking at benefits are looking at some other 00:22:04.300 --> 00:22:06.300 section all the way through that more, 00:22:06.300 --> 00:22:17.500 you know, kind of that war led to the broader area that client 00:22:17.500 --> 00:22:20.300 for busting in things. 00:22:20.300 --> 00:22:26.000 Like, it tastes like content creation and marketing, 00:22:26.000 --> 00:22:26.200 or 00:22:28.100 --> 00:22:33.600 That's another area where we see a lot of an appliance 00:22:33.600 --> 00:22:37.100 kind of expressing interest given the rapid, 00:22:37.100 --> 00:22:45.600 you know, the source of your customized content for our customers. 00:22:45.600 --> 00:22:59.000 A lot of clients who are even though they're taking your point of 00:22:59.000 --> 00:23:07.500 measured approach generation or automated, seems like you'd be 00:23:07.500 --> 00:23:13.800 testing automation so I can help them with their help them 00:23:13.800 --> 00:23:14.400 with their work. 00:23:14.400 --> 00:23:15.000 So 00:23:15.900 --> 00:23:19.200 You know, not getting into a specific client detail. 00:23:19.200 --> 00:23:24.100 Usually we are seeing a lot of traction in those spaces and that's 00:23:24.100 --> 00:23:25.500 how a lot of clients you know, 00:23:25.500 --> 00:23:29.200 that this was reflective of the broader industry but those 00:23:29.200 --> 00:23:31.800 are the ones that they're really honing in on a Ford NAA. 00:23:31.800 --> 00:23:36.800 I saying I'll take down and you know a potentially speed up the 00:23:36.800 --> 00:23:41.000 process of things like Knowledge Management or order, 00:23:41.000 --> 00:23:44.500 you a generating generating some insights for me. 00:23:44.500 --> 00:23:47.500 Otherwise, I would have to poke around for like six different 00:23:47.500 --> 00:23:57.400 sources. In a lot of interests and generating insights, 00:23:57.400 --> 00:23:58.200 like if I had 00:23:59.200 --> 00:24:05.500 Six reports. Then I asked the AI to sort to generate a summary 00:24:05.500 --> 00:24:10.900 of those six reports or what it would be more like find find 00:24:10.900 --> 00:24:12.000 an Insight within this. 00:24:12.000 --> 00:24:15.200 Like how do you find you find that it does a good job of 00:24:15.200 --> 00:24:18.500 it or it is it because we or does it just sort of give 00:24:18.500 --> 00:24:20.700 you a generic kind of bland response. 00:24:20.700 --> 00:24:24.000 I've I've tried it with with summarizing certain articles 00:24:24.000 --> 00:24:27.700 in and work creative types of things and then I don't find that 00:24:27.700 --> 00:24:28.800 it's very creative. 00:24:28.800 --> 00:24:32.800 But on the summary part I feel like it does to a relatively 00:24:32.800 --> 00:24:35.500 good job of summarizing large report. 00:24:36.600 --> 00:24:40.000 Tibet some really great question and I don't want to be very 00:24:40.000 --> 00:24:50.600 clear on the differentiating, like Which models that have been 00:24:50.600 --> 00:24:53.100 released in the last 8 months or so. 00:24:53.100 --> 00:25:02.500 They are not computacional model, they are not inherently 00:25:02.500 --> 00:25:14.100 in. And of themselves, do complex calculations or a piece of calculator 00:25:14.100 --> 00:25:17.100 in do a better job with that mathematics and they can. 00:25:17.100 --> 00:25:26.900 Yeah. So you have to understand either to help predict the next 00:25:26.900 --> 00:25:28.700 word, so you give an input. 00:25:28.700 --> 00:25:31.600 And it says, okay, based on the same quote statistically, 00:25:31.600 --> 00:25:34.900 here's here's what the hell play, what? 00:25:34.900 --> 00:25:36.300 What they're looking for is now. 00:25:36.600 --> 00:25:39.900 So that is definitely a different. 00:25:39.900 --> 00:25:45.700 She wear generative, AI doesn't replace more traditional, 00:25:45.700 --> 00:25:48.600 AI machine learning computational model. 00:25:48.600 --> 00:25:53.100 Absolutely have to still be in play. 00:25:54.100 --> 00:26:07.000 Because it's because it's not just me being orchestrated 00:26:07.000 --> 00:26:08.800 up. So, what's your point? 00:26:08.800 --> 00:26:12.400 It's more. It's not creating those insights. 00:26:12.400 --> 00:26:15.700 But potentially either summarizing those insides, or knowing 00:26:15.700 --> 00:26:18.900 where to go to the information, which, 00:26:18.900 --> 00:26:21.500 you know, which data storage that the information is already, 00:26:21.500 --> 00:26:25.900 there. Opens up a lot of possibilities, 00:26:25.900 --> 00:26:28.600 so, you know, it speeds up people's work clothes. 00:26:28.600 --> 00:26:33.800 So, I might be looking at three or four or five dashboard 00:26:33.800 --> 00:26:35.000 been trying to piece together information. 00:26:35.000 --> 00:26:37.400 At least if you ask a question, 00:26:37.400 --> 00:26:39.900 you know dinner today, I can go and say, 00:26:39.900 --> 00:26:42.600 okay this is the dashboard without information I pulled out. 00:26:42.600 --> 00:26:48.700 So that's really the key to my experience with working with 00:26:48.700 --> 00:26:51.900 a lot of these companies are, are they still taking a a good 00:26:51.900 --> 00:26:53.700 job? Are they doing a good job of working? 00:26:54.100 --> 00:26:58.200 With the technology or does it still feel like the wild west 00:26:58.200 --> 00:27:01.100 or just throw something against the wall and see what sticks 00:27:01.100 --> 00:27:03.900 do you feel like companies are? 00:27:03.900 --> 00:27:06.400 Do they know what they're doing at this point or are they still 00:27:06.400 --> 00:27:08.200 in that very early experimental phase? 00:27:08.200 --> 00:27:22.400 I think it's still experimental on the science of, 00:27:22.400 --> 00:27:27.300 you know, certain companies have already adopted Jenny isolutions, 00:27:27.300 --> 00:27:30.300 but a lot of them are figuring out where the, 00:27:30.300 --> 00:27:43.200 where the best value proposition for it is in Wheeling knows 00:27:43.200 --> 00:27:45.900 they're working. You know, things like that when you're working 00:27:45.900 --> 00:27:48.900 now with management to look at 1, 00:27:54.200 --> 00:27:56.800 We build an effective model, the actor, 00:27:56.800 --> 00:27:59.800 either back to start dating or multiple data set for how do we 00:27:59.800 --> 00:28:03.100 orchestrate multiple models together, that's just an example. 00:28:03.100 --> 00:28:09.600 But that's that is that is because you're still trying to build the 00:28:09.600 --> 00:28:11.000 most optimal version. 00:28:11.000 --> 00:28:16.700 But I think a lot of a lot of companies are customers so that 00:28:16.700 --> 00:28:19.600 we see two are doing it in a measured way. 00:28:19.600 --> 00:28:22.300 They want to make sure they're not growing to something 00:28:22.300 --> 00:28:25.100 at a wall and seeing what they look like they're going through 00:28:25.100 --> 00:28:28.700 the business process of identifying a potentially where dinner 00:28:28.700 --> 00:28:32.600 today all to be helpful. 00:28:32.600 --> 00:28:43.000 That will process of identifying those business is very still 00:28:43.000 --> 00:28:47.000 and don't want to throw like a ton of money at something 00:28:47.000 --> 00:28:52.700 to have no idea what it's good for leaning more towards internal 00:28:52.700 --> 00:28:59.600 facing project. Applications or more public facing external, 00:28:59.600 --> 00:29:02.600 is it? Is it 50/50? 00:29:03.600 --> 00:29:10.800 Are we see both of the marketing and personalization 00:29:10.800 --> 00:29:12.500 is very much customer-facing. 00:29:12.500 --> 00:29:22.600 Lot of clients are interested in like close to speak to customers 00:29:22.600 --> 00:29:25.500 were naturally and understand or their history. 00:29:25.500 --> 00:29:29.300 But also want to eat outside and I'll really, 00:29:29.300 --> 00:29:43.900 I don't, you know, they're there for the employees to better. 00:29:43.900 --> 00:29:47.100 Enhance their work was obviously things like the food generation. 00:29:47.100 --> 00:29:56.400 Yeah. I'm interested in both areas eventually. 00:29:56.400 --> 00:29:57.700 How can I freeze things? 00:29:57.700 --> 00:30:01.800 Like sales retention customer satisfaction on the internal 00:30:01.800 --> 00:30:04.600 side you're really seeing you know how Racial process. 00:30:04.600 --> 00:30:11.200 Make our people more efficient with their work companies 00:30:11.200 --> 00:30:16.300 currently looking at Roi of their projects, 00:30:16.300 --> 00:30:20.600 or is it still too early to start wondering whether all of these 00:30:20.600 --> 00:30:22.900 efforts are going to start paying off at some point? 00:30:23.900 --> 00:30:27.900 I think you saw the phase where Roi isn't, 00:30:27.900 --> 00:30:32.800 you know, it is completely clearly measured. 00:30:32.800 --> 00:30:37.100 It started like a chicken or the egg. 00:30:37.100 --> 00:30:41.600 You have to be able to build some of these experiments 00:30:41.600 --> 00:30:45.800 to measure ryberry, do more than like these experiments. 00:30:45.800 --> 00:30:52.600 If that's not that, that's a, that's a feature of the process. 00:30:52.600 --> 00:31:00.300 But in terms of very clear perception, 00:31:00.300 --> 00:31:04.700 and I think the report covers a lot of potential sales up lives, 00:31:04.700 --> 00:31:08.700 in all that said, how and what the actual numbers are going 00:31:08.700 --> 00:31:15.500 to be in what I can only happen after I started using this in, 00:31:15.500 --> 00:31:17.000 in their day-to-day work? 00:31:17.000 --> 00:31:17.500 Yeah. 00:31:23.900 --> 00:31:26.900 Group a little bit more leeway than maybe they would give other 00:31:26.900 --> 00:31:30.100 Technologies or other projects at this point. 00:31:31.400 --> 00:31:38.000 The general. So yes, I mean it the way it happens is, 00:31:38.000 --> 00:31:40.200 you know, a lot of clients. 00:31:40.200 --> 00:31:43.400 So I guess there's always a lot of job posting yelling on out 00:31:43.400 --> 00:31:49.200 there or Jenny i m Century people understand how these models 00:31:49.200 --> 00:31:55.200 operate and how it takes or how to implement it in their organization. 00:31:55.200 --> 00:31:58.500 The Baptist indicative of a lot of times, 00:31:58.500 --> 00:32:01.300 don't want a lot of companies don't want to be left behind 00:32:01.300 --> 00:32:04.900 in the process so they're investing up front in at least some 00:32:04.900 --> 00:32:07.300 leadership or a like, 00:32:08.900 --> 00:32:14.800 Executive presence and then Dale, you know the idea they're 00:32:14.800 --> 00:32:20.300 still invest a little more rapidly than some of the other traditional 00:32:20.300 --> 00:32:25.500 software Solutions because the perception is both traditional 00:32:25.500 --> 00:32:28.400 solution for more of like a, this is going to fix a specific 00:32:28.400 --> 00:32:31.500 problem. So, you know. 00:32:31.500 --> 00:32:35.300 Business function or invest in that particular product, 00:32:35.300 --> 00:32:38.800 where is having a lot of Enterprises are taking like a bowl 00:32:38.800 --> 00:32:40.700 of Enterprise approach to this like dinner today? 00:32:40.700 --> 00:32:41.900 I can be applied everywhere. 00:32:41.900 --> 00:32:47.700 Yeah, so from that perspective, they're looking at it as an investment. 00:32:47.700 --> 00:32:53.200 I think they do they are looking at it from there. 00:32:53.200 --> 00:32:58.400 Either from their original Partners or Bender or from their 00:32:58.400 --> 00:33:00.700 internal leadership to the Dr. 00:33:00.700 --> 00:33:07.700 Newspaper within the next six months to a year. 00:33:07.700 --> 00:33:08.500 Where do you see 00:33:08.900 --> 00:33:10.900 A lot of companies taking this approach. 00:33:10.900 --> 00:33:16.400 Has it continued experimentation, do you start seeing more deployments 00:33:16.400 --> 00:33:20.600 and new projects or just getting to that point where to see whether 00:33:20.600 --> 00:33:22.800 this stuff is actually going to work in sick? 00:33:23.700 --> 00:33:28.400 I think if you're talkin about Clydesdales in like a year or more, 00:33:28.400 --> 00:33:32.900 I mean you kind of have to see deployments otherwise there's going 00:33:32.900 --> 00:33:38.700 to be, you know, Jenny I Brite no company is going to invest, 00:33:38.700 --> 00:33:42.300 y'all just millions of dollars without any tangible outcome 00:33:42.300 --> 00:33:43.400 long. Of time. 00:33:46.600 --> 00:33:53.300 It is certainly were in that phase where there's like a space, 00:33:53.300 --> 00:33:56.400 but the reality is the list of use cases, 00:33:56.400 --> 00:33:58.000 not or potential. 00:33:58.000 --> 00:34:01.000 Clients are exploring his way up there, 00:34:01.000 --> 00:34:06.500 but in the long-term, it's likely going to mature and settle 00:34:06.500 --> 00:34:11.400 into, you know, I don't like a smaller subset of use cases 00:34:11.400 --> 00:34:14.300 that truly revolutionary. 00:34:14.300 --> 00:34:26.100 Or you start applying applying the hammer to it. 00:34:26.100 --> 00:34:32.200 So, that's what I had and they have the big deal. 00:34:32.200 --> 00:34:34.700 That expertise in-house working partner. 00:34:34.700 --> 00:34:37.500 They're going to start understanding that. 00:34:37.500 --> 00:34:40.400 Okay, well we thought originally business case would be super 00:34:40.400 --> 00:34:43.100 valuable but the reality is it's this other use case. 00:34:43.500 --> 00:34:45.700 And they're going to throw more money at. 00:34:45.700 --> 00:34:52.700 But yeah, I mean, I didn't get any more enterprise-scale, 00:34:52.700 --> 00:34:58.100 deployments a better today, on business functions for the Enterprises 00:35:03.500 --> 00:35:18.400 Exactly. Yeah. It's like even within the subset of male is like, 00:35:18.400 --> 00:35:20.100 his Are there specific types of nails? 00:35:20.100 --> 00:35:28.800 Which are better suited for our business guy always screws 00:35:28.800 --> 00:35:35.200 are. All right, thanks AJ for joining us on the the show today. 00:35:36.200 --> 00:35:39.600 My place would be great. 00:35:39.600 --> 00:35:43.400 So don't forget to like the video. 00:35:43.400 --> 00:35:46.100 Subscribe to the channel and add any comments that you have 00:35:46.100 --> 00:35:48.600 a low during this every week for new episodes of today. 00:35:48.600 --> 00:35:49.700 And I'm keep try. 00:35:49.700 --> 00:35:50.300 Thanks for watching.