Episode 4: The State of the Generative AI Market: Parallels to Speech Recognition and Enterprise Software

00:00:11.410 –> 00:00:30.809
Alex Olesen: Welcome back to Motivation AI Matters. I’m really excited for our guest today. I’m joined by Persado’s Senior Vice President of Go-To-Market Strategy, Vipul Vyas. It’s great to have you on today’s episode.

5
00:00:31.020 –> 00:00:42.000
Alex Olesen: Now I know you, and I work really closely together, but to start things off for the listeners, do you mind giving us a bit of background on who you are…

6
00:00:42.090 –> 00:00:45.650
Alex Olesen: …your career? And what led you to Persado?

7
00:00:46.670 –> 00:01:08.630
Vipul Vyas: Sure, I at Persado currently I, as you mentioned, head up our go-to-market strategy team that encompasses a few things. That’s an unusual title and role. But it involves our sales, engineers, solutions, and consultants that are synonymous with our organization, along with their product marketing organization or value engineering group, which really quantifies and helps us understand the business impact.

8
00:01:08.630 –> 00:01:19.639
Vipul Vyas: We have on our clients along with our education enablement team and our commercialization team. So, it’s a few different groups that are also under one umbrella…

9
00:01:19.640 –> 00:01:34.260
Vipul Vyas: …that we use to really provide our customers with. a high touch experience in terms of understanding and quantifying the impact that we have given the tools that need to be able to succeed on their own…

10
00:01:34.270 –> 00:01:40.570
Vipul Vyas: …etc., and then just helping them understand their solutions as well. In terms of career…

11
00:01:40.660 –> 00:01:57.310
Vipul Vyas: …I can go all the way back. I was a management consultant for many years until I got my first real job. So, I was at Price Waterhouse and Booz Allen. I did my MBA at Dartmouth undergraduate Virginia, and my first quote, unquote, real job was after I left…

12
00:01:57.390 –> 00:02:26.519
Vipul Vyas: Booz Allen. I went to a startup and had to actually be convinced to go because that was too risky, and this was in 1999. It was in the speech recognition space and went from speech recognition to voice by metrics. So, machines understanding what people say, to machines understanding who said it. And then I did a little bit of time in health care at another start-up that I founded, and then I found myself at Persado back, helping…

13
00:02:26.570 –> 00:02:48.620
Vipul Vyas: …make the robots smarter, as I was doing with speech recognition and voice by metrics. So, that’s how I found myself here at Persado. I’m getting back into the natural language processing world that started out almost 20 years ago. I am also an adjunct faculty member at the University of San Francisco, focusing on emerging technologies in the healthcare space.

14
00:02:49.160 –> 00:03:09.849
Alex Olesen: So, you sit in a very unique position at Persado, one in which I’m fortunate enough to have a front-row seat to participate in. We’ve had some great guests in the last couple of weeks talking about what Generative AI means to marketers. We had a guest on the last episode to talk about…

15
00:03:10.050 –> 00:03:14.540
Alex Olesen: …security and safety around implementing these technologies.

16
00:03:14.790 –> 00:03:35.629
Alex Olesen: Your focus in our day-to-day is around the business implications of using Generative AI. Talk to me a little bit about the efficiency and efficacy framework that you’ve been working on the past couple of months because I know it’s gotten a lot of attention from executives through Persado.

17
00:03:36.390 –> 00:03:48.249
Vipul Vyas: Sure. We look at it, as you just mentioned, in terms of AI; specifically, Generative AI can have an impact in terms of improving efficiency and also impacting efficacy…

18
00:03:48.380 –> 00:04:00.779
Vipul Vyas: …and so what that means is, it can. Obviously, you know, people have visually felt things like Chat GPT. If folks have used it, it makes doing a specific task much faster.

19
00:04:00.800 –> 00:04:10.750
Vipul Vyas: You know what would have a report that would have taken, or a blog post that would have taken, you know, hours to create happens now in minutes, and then can be edited and smoothed out…

20
00:04:10.770 –> 00:04:20.820
Vipul Vyas: …in no time at all. So, there’s a massive efficiency gain, you know. It’s like going from a push real lawn mower to a riding lawn. Or…

21
00:04:20.870 –> 00:04:39.180
Vipul Vyas: …you know, men with shovels or people shovels to backhoe. Right? It’s just that kind of just brute power brought into one person’s hands. And that’s pretty substantial. Now, in terms of impact efficacy…

22
00:04:39.190 –> 00:04:48.160
Vipul Vyas: …that really, we think of in terms of performance. So is what is being produced. Also, getting you a…

23
00:04:48.290 –> 00:04:51.739
Vipul Vyas: …increase in the business metric. It’s trying to move…

24
00:04:51.910 –> 00:05:09.910
Vipul Vyas: …and so, what I mean by that is, the blog post that I alluded to before. Yes, you could write it faster now, certainly. But did it get more eyeballs…more traffic to it, and more thorough traffic to whatever potential product you were trying to sell or promote? That is where efficacy plays a role…

25
00:05:10.280 –> 00:05:17.249
Vipul Vyas: …and in most cases, these technologies, generally speaking, have not started to impact…

26
00:05:17.310 –> 00:05:21.059
Vipul Vyas: …that space. They do so in a tangential way in that…

27
00:05:21.280 –> 00:05:42.800
Vipul Vyas: …now you’re able to do more things and get to more things that before we’ve never gotten to just because you couldn’t. And so, having lifted the constraint of limited labor, you now are able to do more things, and you’re becoming effective through that mechanism. But at a discrete sort of unit level, most of the tools are still on…

28
00:05:43.020 –> 00:05:47.080
Vipul Vyas: …the trajectory of improving efficiency…

29
00:05:47.180 –> 00:05:52.700
Vipul Vyas: …now. When you do both, when you can improve both efficiency and efficacy….

30
00:05:53.130 –> 00:05:55.270
Vipul Vyas: …that’s when things sort of get…

31
00:05:55.380 –> 00:05:59.209
Vipul Vyas: …warped or you enter a sort of totally different world…

32
00:05:59.240 –> 00:06:14.829
Vipul Vyas: …and you can do many more magical things, the constraints that you were living within in the past have completely disappeared. And so, things can have fundamental shifts and paradigms. Can really, you know, can move. So…

33
00:06:15.240 –> 00:06:24.880
Vipul Vyas: …that’s where I think we’re heading very soon is where we’re going beyond just efficiency gains and also impacting efficacy. And the combination of…

34
00:06:24.930 –> 00:06:38.470
Vipul Vyas: …those two, which is the ultimate, better, faster, cheaper, is where there’s going to be even more profound impact. So, I can give you some examples very quickly, you know, in the legal space.

35
00:06:38.580 –> 00:06:44.040
Vipul Vyas: You were, you’re now able to efficiently review…

36
00:06:44.200 –> 00:07:13.090
Vipul Vyas: …dozens and dozens, if not thousands, of contracts, and look for compliance to a new indemnification or limitation of liability framework like how many of these are in line with our new standard? How many of these are not, and before you need, like a paralegal, or maybe in a lawyer, to wade through all of these, and that is arduous. Mistakes can happen. And now you can just do that much more efficiently. But you can also get some interesting efficacy gains where…

37
00:07:13.280 –> 00:07:20.109
Vipul Vyas: …now we have inbound complaints for consumers…

38
00:07:20.430 –> 00:07:26.500
Vipul Vyas: …inbound messages and communications that may flag for litigation, litigation risk…

39
00:07:26.610 –> 00:07:29.139
…and you just couldn’t ever get to them all…

40
00:07:29.300 –> 00:07:32.959
Vipul Vyas: …and so, you couldn’t flag which things were potentially sources of…

41
00:07:33.090 –> 00:07:47.760
Vipul Vyas: …legal exposure and which ones were not. Now you can actually sift through those and say, okay. The person who slipped on a…grocery section probably is more of a concern than someone who’s complaining about our prices….

42
00:07:47.850 –> 00:07:52.629
Vipul Vyas: …and so, you can wade through that and identify things that before….

43
00:07:52.980 –> 00:08:07.220
Vipul Vyas: …just because of lack of manpower, lack of just resources, you weren’t able to get to, one on the efficiency set, and then also being able to pattern, recognize and discern which things represent a risk…

44
00:08:07.420 –> 00:08:19.069
Vipul Vyas: …in a scalable way. So, that’s not in the marketing world, obviously, but that gives you a sense of kind of what’s going on in marketing. I think it’s a little bit more…

45
00:08:19.200 –> 00:08:31.589
Vipul Vyas: …direct in that did the language and the words and the message you created drive more of the activity? You wanted more of the behavior you wanted…

46
00:08:31.750 –> 00:08:43.099
Vipul Vyas: …than before. So, are you not just creating language faster, generating content faster? But are you generating better content that’s performing…

47
00:08:43.190 –> 00:08:45.670
Vipul Vyas: …in the way that you want it to perform?

48
00:08:46.680 –> 00:09:00.049
Alex Olesen: And you know this is a topic that we discuss with clients all the time. But for the listeners, you know, Persado is a very good example of a technology which provides both. We do provide more…

49
00:09:00.090 –> 00:09:10.540
Alex Olesen: …effective language generating an incremental amount of conversions. And we, you know, constantly aspire to be a net time saver…

50
00:09:10.710 –> 00:09:14.019
Alex Olesen: …for the marketing teams that we work with.

51
00:09:14.400 –> 00:09:18.510
Alex Olesen: On the prior episode, we had a very…

52
00:09:19.040 –> 00:09:21.739
Alex Olesen: …interesting conversation around…

53
00:09:21.910 –> 00:09:36.930
Alex Olesen: …Generative AI’s role in augmenting jobs in the workplace. We discuss the Goldman report, which says that about 300 million jobs are going to be displaced by the end of this decade.

54
00:09:37.150 –> 00:09:47.889
Alex Olesen: What’s your take on Goldman’s findings and the potential for Generative AI to either pivot some people’s jobs in the workplace or potentially replace them?

55
00:09:48.390 –> 00:09:53.170
Vipul Vyas: There will be dislocation of different degrees…

56
00:09:53.450 –> 00:10:02.220
Vipul Vyas: …and that transition, change that any new technology… creates is typically…

57
00:10:02.770 –> 00:10:06.359
Vipul Vyas: …there’s some discomfort associated with that.

58
00:10:06.480 –> 00:10:15.740
Vipul Vyas: Now, you can do that smoothly, relatively smoothly, over a long arc of time that can be absorbed by society and just human beings and their ability to adapt…

59
00:10:15.940 –> 00:10:18.560
Vipul Vyas: …and so if you look at…

60
00:10:18.570 –> 00:10:25.420
Vipul Vyas: …the US labor force in the 1890s, early, 1900s, 80…

61
00:10:25.440 –> 00:10:35.799
Vipul Vyas: …percent of folks were employed in the agricultural space. You know…farm-related occupations. And today it’s two percent, I mean, very few people know a farmer.

62
00:10:36.320 –> 00:10:47.660
Vipul Vyas: So, two percent of the population produces food for the 98 percent of the people who aren’t in that space. When before, it was almost not quite the inverse, but close.

63
00:10:47.710 –> 00:10:57.200
Vipul Vyas: Pretty much everyone was involved in food production. And so that happened, if you look at graphs of how that transition occurred. It was slow and linear.

64
00:10:57.210 –> 00:11:05.779
Vipul Vyas: Fairly, you know, steadily over the course of a century, and so we absorbed that relatively well.

65
00:11:06.080 –> 00:11:33.949
Vipul Vyas: The dislocation of increased urbanization and the implications of people getting off the farms were managed in a way that didn’t create this much social strife. There are some, of course, you know, the sort of tale of the family farm and the, you know, aggregation of agriculture and fewer hands. There are no issues that people are concerned about there, but in general, from a social dislocation perspective, it was relatively…

66
00:11:33.980 –> 00:11:41.949
Vipul Vyas: …well absorbed. If you look at manufacturing in contrast, it really started to decline in the late sixties and seventies.

67
00:11:42.060 –> 00:12:01.250
Vipul Vyas: And there are some specific reasons as to why people theorize that it happened the way it did. You saw a precipitous decline from, I think, in 27- ish, maybe a little over 30 percent of the labor force employed in manufacturing, and it is now close to at least in the 7% range today.

68
00:12:01.340 –> 00:12:03.619
Vipul Vyas: it may even have drifted a little bit lower.

69
00:12:04.280 –> 00:12:17.389
Vipul Vyas: More recently. So, that happened in the course of less than 50 years and really accelerated in a couple of phases, the eighties, and then the two thousands. And it happened fast…

70
00:12:17.440 –> 00:12:24.430
Vipul Vyas: …and that’s you know what you have as a result of the rust belt. You know, the cities that were impacted, like Detroit…

71
00:12:24.450 –> 00:12:30.950
Vipul Vyas: …and that didn’t happen smoothly. And that doesn’t mean that can’t be mitigated.

72
00:12:31.060 –> 00:12:34.979
Vipul Vyas: If you look at Germany and Japan…

73
00:12:35.220 –> 00:12:51.709
Vipul Vyas: …they actually really worked hard to make sure their manufacturing bases were at least somewhat secure and tempered against some of the dislocation that was occurring in the US. Being the manufacturing powerhouse coming out of World War II did not feel such…

74
00:12:52.220 –> 00:13:03.710
Vipul Vyas: …protection or such mindfulness of that was necessary. We kind of took it… for granted that we were that manufacturing powerhouse, and so…

75
00:13:03.940 –> 00:13:08.440
Vipul Vyas: …that decline was never buffered. So I say that to…

76
00:13:08.470 –> 00:13:16.260
Vipul Vyas: …point to the fact that AI probably will create this dislocation. But it can be managed just like in the past. We’ve…

77
00:13:16.570 –> 00:13:28.099
Vipul Vyas: ..you know, there are examples of where forces that can lead to a severe dislocation can be mitigated. You just have to understand the threat, understand…

78
00:13:28.240 –> 00:13:32.170
Vipul Vyas: …the social issues it can create and then…

79
00:13:32.470 –> 00:13:35.660
Vipul Vyas: …make sure to invest to mitigate them.

80
00:13:36.790 –> 00:13:41.199
Alex Olesen: So, I think that’s a good segue into another part of your career.

81
00:13:41.230 –> 00:13:45.690
Alex Olesen: You are an adjunct professor…

82
00:13:45.790 –> 00:13:51.019
Alex Olesen: …at the University of San Francisco focusing on emerging technologies.

83
00:13:51.400 –> 00:14:03.249
Alex Olesen: What’s the discourse in your classroom with your students? What are they? What’s top of mind for them as they contemplate entering or re-entering the workforce?

84
00:14:03.450 –> 00:14:09.079
Alex Olesen: And what are some of the debates that you have among your students?

85
00:14:10.080 –> 00:14:12.689
Vipul Vyas: I think many of them want to understand, you know…

86
00:14:13.720 –> 00:14:28.309
Vipul Vyas: …to the spirit of your last question, what’s the job market going to be for them? What’s it gonna look like? How’s it going to be different than when they first contemplated going to school? Because things change pretty rapidly in the public, you know, Psyche and Zeitgeist…

87
00:14:28.730 –> 00:14:33.139
Vipul Vyas: …as of September 2022, when Chat GPT emerged on the scene.

88
00:14:33.300 –> 00:14:37.109
Vipul Vyas: So a lot of them are thinking about it. What are the implications for them personally?

89
00:14:37.480 –> 00:14:40.849
Vipul Vyas: I think this is one of those technologies that…

90
00:14:41.060 –> 00:14:42.190
Vipul Vyas: …will…

91
00:14:42.240 –> 00:14:46.879
Vipul Vyas: …Our CEOs alluded to having this sort of hype cycle in the short term…

92
00:14:47.010 –> 00:14:59.269
Vipul Vyas: …and then have more profound effects over the long term, you know. Travel agents got put out of business overnight in the nineties with the emergence of the internet. There was sort of a sputtering along for a while until…

93
00:14:59.510 –> 00:15:04.530
Vipul Vyas: …it just didn’t make sense anymore, right? And I think a lot of…

94
00:15:04.540 –> 00:15:11.729
Vipul Vyas: …industries that are going to be disrupted by Generative AI. It’s not going to happen all of a sudden. It’ll be slow. And then all of a sudden…

95
00:15:11.970 –> 00:15:19.990
Vipul Vyas: …I think the real discussion is, how do people position themselves? Well…

96
00:15:20.070 –> 00:15:24.490
Vipul Vyas: …to take advantage of the technology versus potentially get run over by it…

97
00:15:24.910 –> 00:15:36.700
Vipul Vyas: …and I think that just as in manufacturing, which where automation actually had a huge impact on the rate of employment. If you could run the machines, you had a job…

98
00:15:37.000 –> 00:15:49.660
Vipul Vyas: …if you were the machine. So, there’s gonna be a need for human augmentation for at least some period of time into the foreseeable future. And so, people who can…

99
00:15:50.030 –> 00:15:54.549
Vipul Vyas: …get the machine to do what we need to do.

100
00:15:54.670 –> 00:15:57.100
Vipul Vyas: And actually, it’s more going to be about…

101
00:15:57.360 –> 00:16:02.479
Vipul Vyas: …framing the problem. So people are going to be moving from production…

102
00:16:02.530 –> 00:16:07.799
Vipul Vyas: …to direction in terms of they’re going to be directors versus actors, if you will.

103
00:16:08.130 –> 00:16:19.380
Vipul Vyas: Movie-making analogy perspective. So, instead of being the orchestrator, the actual final leg of work, they’re going to be saying, this is the problem I want to solve.

104
00:16:19.490 –> 00:16:29.100
Vipul Vyas: This is how we want to solve it and then evaluate the solution. And so I think that is going to be a big change in people’s skills, that it’s going to be about defining problems…

105
00:16:29.160 –> 00:16:33.510
Vipul Vyas: …well, in an articulate manner…

106
00:16:33.570 –> 00:16:40.169
Vipul Vyas: …versus the brute force solution that’s maybe where they’ve had focused in the past.

107
00:16:41.050 –> 00:16:43.169
Alex Olesen: I think that’s correct, and…

108
00:16:44.130 –> 00:17:00.079
Alex Olesen: …I’m observing that too in the market when I speak to clients. We, you know, at Persado, we’ve made a push to make part of our product offering more self-service, less reliant on client first-party data…

109
00:17:00.290 –> 00:17:07.099
Alex Olesen: …and I even heard a couple of our clients refer to Persado as an extra member of the creative team.

110
00:17:07.430 –> 00:17:12.060
Alex Olesen: And I really do think that the dynamic that we see at play…

111
00:17:12.109 –> 00:17:17.329
Alex Olesen: …with enterprises who are implementing this technology effectively…

112
00:17:17.450 –> 00:17:18.780
Alex Olesen: …and efficiently…

113
00:17:19.619 –> 00:17:23.539
Alex Olesen: …it’s not necessarily replacing human beings today…

114
00:17:24.040 –> 00:17:27.339
Alex Olesen: …but rather helping undo writer’s block…

115
00:17:27.530 –> 00:17:30.799
Alex Olesen: …or help alleviate the blank page problem.

116
00:17:31.250 –> 00:17:41.959
Alex Olesen: And then when you take that and supplement it with the macro-economic data that Persado provides around hundreds of millions of consumer interactions…

117
00:17:41.990 –> 00:17:46.170
Alex Olesen: …that’s where you add a statistical element to…

118
00:17:46.270 –> 00:17:47.670
Alex Olesen: human intuition.

119
00:17:48.770 –> 00:18:06.830
Alex Olesen: So, in your observations, companies that are getting this right today: What are they doing, and what is your advice to an executive who’s contemplating making their first investment into Generative AI for the enterprise?

120
00:18:08.720 –> 00:18:13.969
Vipul Vyas: It’s an interesting, it’s a tough question because there’s…

121
00:18:14.020 –> 00:18:17.579
Vipul Vyas: …a few things are happening.

122
00:18:18.140 –> 00:18:26.249
Vipul Vyas: One, the core technology providers, such as Open AI, which has got a significant amount of funding from Microsoft, so it is almost sort of a commingled entity….

123
00:18:26.340 –> 00:18:32.739
Vipul Vyas: …obviously came on the scene with Chat GPT several months ago, almost…

124
00:18:33.650 –> 00:18:41.769
Vipul Vyas: …a little bit over half a year ago, I suppose? That’s now freely available to everyone…

125
00:18:41.810 –> 00:18:47.590
Vipul Vyas: …and the core technology is also available. So, you have as a function of that…

126
00:18:47.710 –> 00:18:50.970
Vipul Vyas: …the proliferation of dozens…

127
00:18:51.060 –> 00:19:02.530
Vipul Vyas: …probably more like hundreds of small companies that are all building point solutions on top of the core technology. Because now the barrier to entry…

128
00:19:03.110 –> 00:19:05.890
Vipul Vyas: …to be innovative, to create something…

129
00:19:06.020 –> 00:19:19.339
Vipul Vyas: …is much lower because someone’s taking the hit on that big development effort, the big training model development transforms all development machine learning efforts, all that big lift.

130
00:19:20.250 –> 00:19:27.299
Vipul Vyas: And now other people can leverage it to get in the market very quickly, and so they are. And so, it’s almost like a…

131
00:19:27.760 –> 00:19:33.199
Vipul Vyas: …brush fire came through. And now there’s a meadow full of flowers, just a lot of them to choose from.

132
00:19:33.410 –> 00:19:37.399
Vipul Vyas: But I think in terms of executives picking…

133
00:19:37.860 –> 00:19:39.879
Vipul Vyas: …the solutions that make sense…

134
00:19:40.960 –> 00:19:42.810
Vipul Vyas: …the generalist…

135
00:19:42.970 –> 00:19:50.570
Vipul Vyas: …technologies are not going to solve specific business problems because they’re focused on being good at being general…

136
00:19:50.830 –> 00:19:53.620
Vipul Vyas: …at offering a generic capability.

137
00:19:53.770 –> 00:20:03.299
Vipul Vyas: And they want ecosystems to point solutions because it helps them cheaply. Figure out what’s relevant in the market. Other people can go off and explore what use cases are…

138
00:20:03.620 –> 00:20:06.120
Vipul Vyas: …use or pertinent and…

139
00:20:06.290 –> 00:20:09.370
Vipul Vyas: …do all that heavy lifting for them…

140
00:20:09.660 –> 00:20:14.239
Vipul Vyas: …and they can sit back. You know, it’s the whole gold rush analogy of…

141
00:20:14.400 –> 00:20:28.660
Vipul Vyas: …the people who supplied the pick. Axes were the ones that made the money, not the miners, necessarily themselves. And so, I think the general technologies will still focus on the general technology, general tech because there’s so much still to do there. There’s so much green field still…

142
00:20:28.990 –> 00:20:42.000
Vipul Vyas: …to conquer and so point solutions will be the things that emerge that people have to evaluate because they’re focused on specific problems by their nature point solutions are attempting to solve specific issues or problems or challenges.

143
00:20:42.010 –> 00:21:02.950
Vipul Vyas: And so, then it’s a matter of what point solutions are the right ones to pick. And so, ones that actually are intimately familiar with the business problem that you’re trying to solve to some level of expertise of the processor space and that are focused on moving specific KPIs because, at the end of the day, you’re not buying Generative AI, you’re buying a business outcome.

144
00:21:03.360 –> 00:21:06.329
Vipul Vyas: It’s ideally what you’re buying. You’re buying a result.

145
00:21:06.500 –> 00:21:19.869
Vipul Vyas: And so looking back at other examples or other industries or other similar technologies where there’s a core technology provider set that has come on the scene that enabled a lot of other people to proliferate.

146
00:21:20.100 –> 00:21:24.870
Vipul Vyas: You see a few themes. One is that these point solutions have four key characteristics.

147
00:21:24.940 –> 00:21:31.580
Vipul Vyas: One is that they have optimized UI/UX for the specific…

148
00:21:31.760 –> 00:21:33.930
Vipul Vyas: …problem they’re trying to solve…

149
00:21:33.990 –> 00:21:43.230
Vipul Vyas: …and that’s an artful effort, though not insurmountable. So, they have to…

150
00:21:43.490 –> 00:21:54.029
Vipul Vyas: …make it as easy as possible for the end user to accomplish…tasks. And so then you at least get user buy-in. And the second thing is workflow integration.

151
00:21:54.100 –> 00:22:04.760
Vipul Vyas: It’s similar to UI/UX. Does this seamlessly flow into existing or adjacent processes, such that it’s not hyper-disruptive to existing…

152
00:22:05.360 –> 00:22:33.159
Vipul Vyas: …workflows, such that it’s hard to absorb? That may change slowly over time. But initially, you want to fit into a construct where people can get their heads around it so that the technology can actually be successfully absorbed. So, again, one is UI/UX. Two is a workflow integration. Three is reporting. And that’s really the way that the vendor justifies their invoice. But it’s also a way that you understand. That is this solution actually moving…

153
00:22:33.280 –> 00:22:36.699
Vipul Vyas: …the metric, that it says it was going to move, that it was going to move.

154
00:22:37.040 –> 00:22:53.449
Vipul Vyas: And so reporting and robustness of reporting and the insights you can give are gonna be big, and those are solutions specific, not general or generic. You know that that typically won’t come from a generic provider. And the fourth thing is that they’ve extended the technology. So, in this case, they’ve extended language models…

155
00:22:53.590 –> 00:22:58.360
…there’s a litany of previous examples that are illustrative of this…

156
00:22:58.450 –> 00:23:25.069
Vipul Vyas: …from speech to recognition technology that I’m familiar with. Where, if you want to do driving directions…the UI/UX had to be optimized for maps, workflow integration, and how to integrate with GPS ships and whatnot, which is the generic speech. Recognition applications would not bother with like, why would they? They had to actually report in terms of accuracy and get that feedback loop going, and then a fourth. They had to do novel things such as disambiguate…

157
00:23:25.100 –> 00:23:39.529
Vipul Vyas: …words like Marquette Street and Market Street M.A.R.Q.U.E.T.T.E. versus M.A.R.K.E.T. and they have to use traffic data to figure out probabilistically if they both sound similar when someone says them.

158
00:23:39.600 –> 00:23:50.599
Vipul Vyas: …which one is more likely going to be to disambiguate with me, too? So, that to extend the core technology to address the specific problem, such as an example, you know, in the case of…

159
00:23:50.620 –> 00:24:04.989
Vipul Vyas: …Persado, we’ve had to actually take language and actually inject emotion into it. and then couple that with, you know, a clear description of a call to action, to then a couple of motions with motivation to actually drive…

160
00:24:05.280 –> 00:24:18.460
Vipul Vyas: …action. And the behaviors we want. So Persado is another just case study of solution in the space with generic capabilities and generic technologies where the point solution has to extend…

161
00:24:18.570 –> 00:24:24.800
Vipul Vyas: …to what the core solution the core technology does to solve a business problem, in this case, driving conversions…

162
00:24:24.830 –> 00:24:29.960
Vipul Vyas: …and the behaviors we want from consumers. So…

163
00:24:30.230 –> 00:24:34.820
Vipul Vyas: …that’s the theme. You can see that with voice-by-metrics…

164
00:24:34.920 –> 00:24:41.430
Vipul Vyas: …as well in this another form of gender value people don’t have to think about is a path planning with…

165
00:24:41.590 –> 00:24:53.290
Vipul Vyas: …actual autonomous vehicles. The path plan is actually a form of Generative AI, and you have different applications in different scenarios from the 18-wheeler versus an autonomous…

166
00:24:53.510 –> 00:25:04.789
Vipul Vyas: …bus system that they’re now rolling out in different cities. So yeah, I think the executives are going to think about making bets on technologies. The good thing is that you know.…

167
00:25:04.950 –> 00:25:08.630
Vipul Vyas: …these bets aren’t sealed in cement…

168
00:25:08.750 –> 00:25:12.690
Vipul Vyas: …but you look for some hallmark characteristics…

169
00:25:12.750 –> 00:25:16.060
Vipul Vyas: …of winners. We’re likely winners…

170
00:25:16.260 –> 00:25:21.829
Vipul Vyas: …based on some of the attributes I just mentioned. And that’s based on my…

171
00:25:22.010 –> 00:25:35.169
Vipul Vyas: …you know, sort of set of characteristics I went through is just based on history or the last 20 years. Seeing similar trends play out where a general technology came on the scene. And then a bunch of point solutions emerged.

172
00:25:35.360 –> 00:25:39.360
Vipul Vyas: There’s a whole cacophony of activity and then a few…

173
00:25:39.670 –> 00:25:41.700
Vipul Vyas: …remain standing at the end…

174
00:25:41.850 –> 00:25:46.190
Vipul Vyas: …and it was the ones that have those four characteristics I mentioned more often than not…

175
00:25:46.290 –> 00:25:51.520
Alex Olesen: It’s very reminiscent of Vipple if you wind the clock back about 10-15 years.

176
00:25:51.640 –> 00:26:02.470
Alex Olesen: Large enterprises were either an SAP shop or an Oracle shop, both of which had HR functionality, finance, functionality, and procurement.

177
00:26:02.730 –> 00:26:12.460
Alex Olesen: …and then, with the advent of software as a service and cloud providers, you had what was referred to back in the day as best of breed…

178
00:26:12.780 –> 00:26:17.810
Alex Olesen: …and I remember in at the beginning of my career, selling into these companies…

179
00:26:17.870 –> 00:26:30.620
Alex Olesen: …there was a philosophy of, you’re either a loyalist to one of SAP or Oracle, and you buy every department of functionality or best of breed…

180
00:26:30.680 –> 00:26:40.659
Alex Olesen: …and you have the core ERP that would API into cloud solutions, which back in 2008, 2010…

181
00:26:40.900 –> 00:26:45.990
Alex Olesen: …didn’t necessarily have their own analytics and reporting functionality…

182
00:26:46.170 –> 00:26:50.939
Alex Olesen: …had limited rep in terms of their capabilities…

183
00:26:51.260 –> 00:27:04.679
Alex Olesen: …fast forward to today. A lot of those point solutions, such as work day, are notable examples. Those run entire large organizations. And I think it’s actually a pretty interesting parallel…

184
00:27:04.880 –> 00:27:12.119
Alex Olesen: …between an open AI and SAP general technology that can be used to satisfy…

185
00:27:12.200 –> 00:27:14.300
Alex Olesen: …repeatable business functions…

186
00:27:14.420 –> 00:27:18.720
Alex Olesen: …and then more bespoke LLMs will be built either on top of…

187
00:27:18.830 –> 00:27:30.830
Alex Olesen: …those language models or in conjunction with them. and I do think that it will follow a very similar technology adoption arc that the ERPs and best-of-breed solutions followed about 15 years ago.

188
00:27:31.640 –> 00:27:32.709
Vipul Vyas: You’re probably right.

189
00:27:33.400 –> 00:27:36.809
Vipul Vyas: Alright. I think that the other thing that will…

190
00:27:37.660 –> 00:27:38.820
Vipul Vyas: …follow…

191
00:27:39.010 –> 00:27:43.600
Vipul Vyas: …I suspect is going to be along the lines of…

192
00:27:43.770 –> 00:27:50.609
Vipul Vyas: …you’re gonna have Generative AI everywhere, especially from a marketing perspective. You’re gonna have Generative AI at all the point solutions…

193
00:27:50.690 –> 00:28:01.450
Vipul Vyas: …or at all the different points, channels, specifically. And so, you’ll have an SMS Generative AI solution. You’ll have one for email, one for web…

194
00:28:01.760 –> 00:28:12.640
Vipul Vyas: …and so you can have lots of different potential Generative AI solutions to pick from provided by your existing vendors for core messaging technologies…

195
00:28:12.900 –> 00:28:24.450
Vipul Vyas: …and so what may happen is that each one of these is going to have a different voice based on the people that are in charge for that particular channel within the enterprise…

196
00:28:24.620 –> 00:28:33.270
Vipul Vyas: …based on the vagaries of that particular Generative AI solution. And so, you’re going to be communicating different messages…

197
00:28:33.650 –> 00:28:37.019
Vipul Vyas: …to different, to people at different times…

198
00:28:37.130 –> 00:28:40.680
Vipul Vyas: …because you have this sort of…

199
00:28:41.290 –> 00:29:01.589
Vipul Vyas: …menagerie of different Generative AI solutions. This happened with CRM many, many years ago, where once people knew they could use SMS tools, they could use email, and they could use different ways to communicate with customers. Everyone kind of went nuts and started sending customers a bunch of SMS messages, a bunch of emails, a bunch of…

200
00:29:01.700 –> 00:29:05.529
Vipul Vyas: …even in some cases…

201
00:29:05.710 –> 00:29:13.279
Vipul Vyas: …printed material. And there was no one sort of coordinating what messaging was going to the customer…

202
00:29:13.560 –> 00:29:15.859
Vipul Vyas: …what made sense to prioritize? And what…

203
00:29:16.550 –> 00:29:29.900
Vipul Vyas: …ways we could, you know what mechanism we could use to keep someone from being overwhelmed so that we diluted our own messaging? That’s when CRM really came on the scene, right to basically be a coordination function. To say, I want to be an air traffic controller…

204
00:29:30.030 –> 00:29:40.499
Vipul Vyas: …to coordinate what is sent, to whom, and when? And how often? And that was basically an attempt to rein in the madness.

205
00:29:41.050 –> 00:29:59.629
Vipul Vyas: I literally, at one point, many, many years ago, got an SMS, a paper letter, and an email from the same company on the same day, wanting to do three different things. And so, this is again exactly what CRM was trying to solve. And I think nowadays what we’ll want to head off at the past is…

206
00:29:59.860 –> 00:30:01.650
Vipul Vyas: …different language…

207
00:30:01.660 –> 00:30:06.980
Vipul Vyas: …being used for a person in different places…

208
00:30:07.460 –> 00:30:16.989
Vipul Vyas: …and a lack of consistency and a dilution of brand identity. and also…

209
00:30:17.050 –> 00:30:22.810
Vipul Vyas: …you know, some people call it brand voice, but also just the ability to…

210
00:30:24.140 –> 00:30:30.169
Vipul Vyas: …really distinguish the company and also orient the customer. Here’s what they can expect…

211
00:30:30.180 –> 00:30:32.159
…in terms of how we communicate…

212
00:30:32.480 –> 00:30:35.590
Vipul Vyas: …because there is a single point…

213
00:30:36.050 –> 00:30:41.399
Vipul Vyas: …of coordination. And Persado can play that role of essentially a language hub…

214
00:30:41.430 –> 00:30:56.230
Vipul Vyas: …or air traffic control. To say, you know, here’s the kind of language that resonates with this person at this time in this context. And so, we should use across channels to essentially at the end of the day. You’re trying to always drive a couple of things. you’re trying to influence behavior…

215
00:30:56.280 –> 00:31:01.939
Vipul Vyas: …and drive loyalty. And so you’re not gonna be able to do that with…

216
00:31:02.160 –> 00:31:05.539
…disparate messages that aren’t very well…

217
00:31:05.620 –> 00:31:07.600
Vipul Vyas: …coordinated across platforms.

218
00:31:08.430 –> 00:31:20.449
Vipul Vyas: And that’s one thematic issue that I think people in marketing should specifically worry about, which is that now we have 15 Generative AI solutions. They all speak to our customers in different ways.

219
00:31:21.870 –> 00:31:25.410
Vipul Vyas: And then also our customer here is 15 different ways of us talking to them.

220
00:31:26.790 –> 00:31:37.989
Alex Olesen: So, on the topic of air traffic control. I know you you recently attended, I believe it was a congressional hearing on AI…

221
00:31:38.810 –> 00:31:49.549
Vipul Vyas: …any main takeaways you want to share from that with the listeners?

222
00:31:50.100 –> 00:32:02.969
Vipul Vyas: They didn’t head off the social implications of social media in time; they didn’t understand or grasp the unintended consequences, or the natural…

223
00:32:03.570 –> 00:32:10.549
Vipul Vyas: …progression of what would follow the implicate natural implications of something like social media because they didn’t bother to step back and understand them…

224
00:32:11.240 –> 00:32:15.670
Vipul Vyas: …and so they don’t want to make the same mistake twice…

225
00:32:16.030 –> 00:32:28.470
Vipul Vyas: As a result, I think AI is going to get a lot more scrutiny from a political perspective because it’s clear. I mean, you don’t have to imagine you can…

226
00:32:28.760 –> 00:32:34.429
Vipul Vyas: …feel it, that you can see and sense it, that it’s clear potential impact on society…

227
00:32:35.120 –> 00:32:39.289
Vipul Vyas: …and so, the big takeaway to summarize that is that…

228
00:32:39.590 –> 00:32:44.980
Vipul Vyas: …the political world will be active and much more engaged with Generative AI…

229
00:32:45.280 –> 00:32:52.700
Vipul Vyas: …then it has been with technologies in the recent past. I think Italy potentially banned a lot of them…

230
00:32:52.780 –> 00:32:56.819
Vipul Vyas: …you know, Chat GPT usage already as a country…

231
00:32:56.920 –> 00:33:00.810
Vipul Vyas: …enterprises have also done the same thing…

232
00:33:01.170 –> 00:33:03.279
Vipul Vyas: …so…

233
00:33:03.340 –> 00:33:09.030
Vipul Vyas: …I know that’s not governmental, but I think that’s going to be the theme is that you’re gonna have a very active…

234
00:33:09.270 –> 00:33:13.300
Vipul Vyas: …political response to anything that’s happening in the space…

235
00:33:16.220 –> 00:33:25.619
Alex Olesen: …that that makes sense. Well, I know we’ve covered a lot of ground today, and thank you. You’ve been a tremendous guest. I have one final question for you…

236
00:33:26.290 –> 00:33:33.000
Alex Olesen: …of all of your observations in the market, everything that you’re working on with Persado…

237
00:33:33.110 –> 00:33:39.469
Alex Olesen: …what would you say you’re looking forward to the most in this space in the next six to 12 months?

238
00:33:41.230 –> 00:33:52.529
Vipul Vyas: I know we spoke a little bit about potential, you know, social doom and gloom here and there, but I think that the other part the other side of the coin. Is that…

239
00:33:53.070 –> 00:33:58.300
Vipul Vyas: …all the creativity that is trapped in many, many people’s heads…

240
00:33:58.540 –> 00:34:00.580
Vipul Vyas: …for want of a way to express it…

241
00:34:00.630 –> 00:34:10.409
Vipul Vyas: …because they can’t get it out, because, you know, they can’t draw. They can’t write, maybe, as well as they think they should be able to write all the things, all the creativity…

242
00:34:10.449 –> 00:34:13.499
Vipul Vyas: …that is locked away in humanity…

243
00:34:13.560 –> 00:34:18.190
Vipul Vyas: …now has a mechanism by which to express itself.

244
00:34:18.320 –> 00:34:27.630
Vipul Vyas: You know, if you want to create an application, you need to learn how to code. There was an inherent barrier for you…

245
00:34:28.880 –> 00:34:35.290
Vipul Vyas: …expressing and bringing to reality what you wanted to bring to reality. Now…

246
00:34:35.370 –> 00:34:44.319
Vipul Vyas: …I mean code is an abstraction. Right? Basically, ultimately, code is just computer programming code. is just an abstraction layer above…

247
00:34:45.400 –> 00:34:48.939
Vipul Vyas: …turning on and off a bunch of transistors on a chip…

248
00:34:49.210 –> 00:34:50.280
Vipul Vyas: Ultimately…

249
00:34:50.320 –> 00:34:54.389
Vipul Vyas: …it’s just different layers of getting that now…

250
00:34:54.570 –> 00:34:57.989
Vipul Vyas: …simple spoken words, normal language…

251
00:34:59.290 –> 00:35:01.610
Vipul Vyas: …can effectuate that…

252
00:35:01.680 –> 00:35:05.609
Vipul Vyas: …you can actually describe. I want a program that does Xyz…

253
00:35:05.980 –> 00:35:09.549
Vipul Vyas: …as an example. And generally, I can create such a thing…

254
00:35:10.570 –> 00:35:15.169
Vipul Vyas: …right? And so, I look at that to say, there’s a lot of trapped…

255
00:35:15.370 –> 00:35:26.610
Vipul Vyas: …creativity that for simple sources of friction that have historically existed, couldn’t get out. Now, it can. Same thing on the marketing front. We, wanna…

256
00:35:26.740 –> 00:35:33.120
Vipul Vyas: …you know, as you mentioned, there’s plenty of things that never happened because riders block so many things that didn’t happen. Well…

257
00:35:33.400 –> 00:35:38.160
Vipul Vyas: …because I just didn’t know certain things that are now knowable.

258
00:35:38.980 –> 00:35:48.160
Vipul Vyas: And I think, as I alluded to before, when you have that ultimate nexus of efficiency and efficacy, that wonderful…

259
00:35:48.320 –> 00:35:52.390
Vipul Vyas: …confluence of the two you remove constraints. Now…

260
00:35:52.720 –> 00:36:03.629
Vipul Vyas: …things you can’t even imagine were possible before, such as speaking individually to Alex in the way Alex wants to be spoken to in a way that’s compelling to him…

261
00:36:04.580 –> 00:36:12.560
Vipul Vyas: …is possible because I can create copy on the fly images on the fly, video, on the fly. All contextually…

262
00:36:13.610 –> 00:36:14.770
Vipul Vyas: …modified…

263
00:36:15.790 –> 00:36:21.609
Vipul Vyas: …that before you know, just a concrete example, if I want to shoot a video, I’ve got to…

264
00:36:21.830 –> 00:36:38.539
Vipul Vyas: …get studio time. You gotta get actors. Gotta get, you know, a script written. Get all that stuff done. Get a, basically, shoot this thing, and do a lot of post-production editing. It’s a whole thing. So, you better make sure that the video generically is resonating with as many people as possible.

266
00:36:39.810 –> 00:36:44.430
Vipul Vyas: The constraints that force that are gone. So actually, I can create…

267
00:36:44.500 –> 00:36:56.170
Vipul Vyas: …interactive and engaging material and content for a person. In this case, you know yourself, Alex. dynamically, on the fly. for essentially close to nothing…

268
00:36:56.180 –> 00:36:57.450
Vipul Vyas: …in terms of cost…

269
00:36:58.800 –> 00:37:03.260
Vipul Vyas: …and the that settled to it better, faster, cheaper, and that…

270
00:37:03.460 –> 00:37:15.059
Vipul Vyas: …a fundamentally different way of engaging people. You know, a much more personalized way that’s going to resonate with them is just going to be different. I mean, it’s gonna open up an entire new…

271
00:37:15.180 –> 00:37:20.290
Vipul Vyas: …way of people interacting with enterprises, people interacting in general…

272
00:37:20.540 –> 00:37:25.749
Vipul Vyas: So, I think those are the kind of interesting, positive things I think, that are on the near-term horizon.

273
00:37:25.860 –> 00:37:28.499
Alex Olesen: It’s fascinating. It is such a…

274
00:37:28.860 –> 00:37:36.109
Alex Olesen: …watershed moment. In my opinion, I know you. You mentioned some others. I’ll summarize them for the listeners. Now…

275
00:37:36.560 –> 00:37:39.249
Alex Olesen: …I do think, sequentially…

276
00:37:40.430 –> 00:37:46.309
Alex Olesen:… manufacturing the computer being invented, the internet proliferating…

277
00:37:46.500 –> 00:37:55.799
Alex Olesen: …I do think that we are looking at another moment with the advancement of Generative AI that will have the ability to transform the way we work…

278
00:37:55.890 –> 00:37:58.319
Alex Olesen: …transform the way we live, the way we learn.

279
00:37:58.590 –> 00:38:05.899
Alex Olesen: And I appreciate all of the examples that you’ve given. You know you’ve had a front-row seat…

280
00:38:05.990 –> 00:38:09.969
Alex Olesen: …to this industry for the better part of the last 20 years…

281
00:38:10.210 –> 00:38:28.770
Alex Olesen: …and finally, the wealth of knowledge you’ve been sharing with Persado’s clients. Now, our listeners will be able to benefit from Vipul Vyas. Thank you very much. for joining the podcast again. This was Vipul Vyas. He is the Senior Vice President of Go-To-Market Strategy at Persado.

283
00:38:30.180 –> 00:38:40.929
Vipul Vyas: thanks, Alex. I’ll I’ll not let you stop me there. I’ll say two more quick things in terms of one of the things that buffered manufacturing decline and job creation was actually the rise of the It sector…

284
00:38:41.580 –> 00:38:56.689
Vipul Vyas: …which actually created jobs that people who no longer go to manufacturing can now assume to some degree. So, that was a safety net of some degree. So, we’ll need something like that. And the second is, you know, another big technology trend…

285
00:38:56.750 –> 00:39:06.540
Vipul Vyas: …that didn’t quite take office people into it was blockchain, as an example. And I think actually, it will have new relevance because you can imagine scenarios where…

286
00:39:06.600 –> 00:39:12.430
Vipul Vyas:… a company’s enterprise system is a negotiating with another company’s enterprise system or…

287
00:39:12.530 –> 00:39:22.190
Vipul Vyas: …AI bots are actually interacting with each other, and the resolution of the interaction because it’s now no one entities…

288
00:39:22.870 –> 00:39:28.310
Vipul Vyas: …ownership of that. The final adjudication can leverage things like blockchain to track…

289
00:39:28.550 –> 00:39:39.349
Vipul Vyas: …the outcome of such interactions. So, that may actually be a technology that sort of came and lost a little bit of luster recently. We actually find new life…

290
00:39:39.550 –> 00:39:51.009
Vipul Vyas: …I’ll leave it at that. Thank you for having me. I definitely appreciate it.