Episode 2: Generative AI and Persado’s Role in the Ecosystem

1
00:00:20.100 –> 00:00:21.990
[email protected]: All right, so we are recording.

2
00:00:24.170 –> 00:00:33.880
[email protected]: Welcome back to the motivation. AI matters Podcast. I’m really excited to be joined by this episode’s guest. Paul Ritzer.

3
00:00:34.040 –> 00:00:39.000
[email protected]: Paul is the founder and CEO of the marketing. AI Institute.

4
00:00:39.130 –> 00:00:50.780
[email protected]: Paul Great to have you. It’s great to be here, looking forward to the conversation. We’ve we’ve had plenty of offline conversations over the years, but for a chance to think we get to do it for a podcast, so it should be fun.

5
00:00:50.780 –> 00:01:01.200
[email protected]: I know I’ve been looking forward to this one for a while. We’ve been talking about appearing on a podcast together for a couple of months. We’ve had a lot of really interesting conversations about

6
00:01:01.230 –> 00:01:12.240
[email protected]: generative AI. You’ve been a long time, friend of Persados, and I figured we’d make it official and bring you on the podcast, and talk about your point of view

7
00:01:12.370 –> 00:01:15.210
[email protected]: on the industry and your background.

8
00:01:15.570 –> 00:01:27.670
[email protected]: and I think that’s a great place to get started, so could you share with us a little bit about your background. How you arrived at where you are today

9
00:01:27.820 –> 00:01:41.800
Paul Roetzer: from the college days, I guess, because I came out of journalism school. So I think the first thing people need to know about me is, i’m actually a liberal arts, Major. I’m a writer by trade. My wife is a painting major, and in our history, major, so like

10
00:01:41.860 –> 00:01:56.170
Paul Roetzer: I come at this from an angle of I. I have a deep respect for the arts and the human ability to create. So I am not an a machine learning engineer or data scientist, or, you know, an AI researcher in a traditional sense.

11
00:01:56.300 –> 00:02:03.470
Paul Roetzer: I I started working in a marketing agency, and eventually founded my own in 2,005,

12
00:02:03.530 –> 00:02:16.190
Paul Roetzer: and that that we became Hubspot’s first partner in 2,007, and which really threw me into the marketing technology space and the emergence of Mobile. You know, iphone social media. Everything was sort of bubbling up at that time.

13
00:02:16.280 –> 00:02:28.970
Paul Roetzer: and then in 2,011. I just developed a curiosity about artificial intelligence, and started trying to understand what Ibm Watson was, and how it worked, and how it one on jeopardy. And so I spent a couple of years just researching AI and

14
00:02:29.050 –> 00:02:35.230
Paul Roetzer: wrote about it a little bit my second book in 2,014 to some theories about how it could be applied to marketing.

15
00:02:35.360 –> 00:02:41.130
Paul Roetzer: and then I started just doing all the public speaking about it. People started asking me to come and give talks on that topic, and

16
00:02:41.150 –> 00:02:50.150
Paul Roetzer: and then, in 2,016, we turned that sort of interest and curiosity into the marketing. I Institute and just started publishing a couple of times a week what we were learning and

17
00:02:50.250 –> 00:03:03.400
Paul Roetzer: interviews we were doing, and companies we were following like Prisato. You know the companies that in the early days early days of AI in the mid teens we’re actually like doing some interesting things with machine learning and natural language processing.

18
00:03:03.420 –> 00:03:16.490
Paul Roetzer: And yeah, and so then I sold my agency in 2,021 to focus on marketing. I Institute, and that’s what I do today. We’re a media Event and Education Company run conferences, online courses and publish a bunch of content. And

19
00:03:16.570 –> 00:03:18.280
Paul Roetzer: I have a podcast. So

20
00:03:18.420 –> 00:03:33.080
Paul Roetzer: yeah, just so. I wouldn’t say I fell into it. It’s been like a 12 year. Run to end up where we are today. But chat Gpt certainly accelerated the interest in what we’re doing, and I think just overall awareness about artificial intelligence

21
00:03:33.340 –> 00:03:47.250
[email protected]: absolutely. You know, we at Prisato also have a lot of people who say they fell into artificial intelligence and fast. In fact, our prior guest, Lisa Spira, who oversees our Content Intelligence team.

22
00:03:47.250 –> 00:03:59.200
[email protected]: She is a linguistics expert by trade, and did not have formal education and machine learning. And I think that affords her a very unique perspective, being able to work with our technology.

23
00:03:59.220 –> 00:04:11.390
[email protected]: And I would say the same thing about you. So, taking your your background in in writing, and even your wife’s passion for art, this must be a very

24
00:04:11.440 –> 00:04:14.940
[email protected]: interesting time for you to observe

25
00:04:15.040 –> 00:04:34.640
[email protected]: this wave of generative AI technology being able to author its own content, using a series of prompts in your own words what is generative? AI, and what are some of the most useful applications that you’ve seen of Generative AI: so far?

26
00:04:34.890 –> 00:04:40.480
Paul Roetzer: Yeah. We think of Jenner AI’s the ability for the machine to create something from a prompt, so

27
00:04:40.650 –> 00:05:05.430
Paul Roetzer: language is an obvious one, the ability for it to write things. But you also have image generation, video generation, audio generation, you know, music or the AI Drake thing, you know. A few weeks back or month or 2 ago. It’s able to generate audio, and then code is the other one. But those are the kind of the 5 major things now that it’s going to keep expanding, and there’s other things like synthetic data and like. But in terms of you know, most business and marketing people.

28
00:05:05.430 –> 00:05:25.300
Paul Roetzer: those are the categories that you’re gonna be really be thinking about. And so the used cases are basically anywhere where those things exist anywhere where you’re creating content. So if you’re in sales and you you know, sales, emails and proposals. If you’re in marketing, you know, marketing emails, landing pages, website copy ads, social copy.

29
00:05:25.300 –> 00:05:39.430
Paul Roetzer: These are all things that AI can assist. And again, like Prasado’s, been doing stuff like this for years. Where your custom, training these things, these capabilities, based on, you know, brands and data. So it’s not new, but Chat Gbt sort of

30
00:05:39.780 –> 00:05:50.100
Paul Roetzer: the generative AI term for it emerged in 2,022, and then chat Gpt Just brought it to everyone like, I think. What was the 100 millionusers in January or something?

31
00:05:50.100 –> 00:06:01.860
Paul Roetzer: So that was really what changed, as there were some breakthroughs in the capabilities. But it was really just the mass access and adoption to the technology that brought us to where we are.

32
00:06:02.160 –> 00:06:07.280
[email protected]: Yeah, I agree there. There have been some precursor technologies to what we now

33
00:06:07.330 –> 00:06:17.640
[email protected]: commonly referred to as generative AI, we’ll get to Prisado in a bit, but we we do pride ourselves on having a decade, long track record of

34
00:06:18.110 –> 00:06:19.370
[email protected]: humans

35
00:06:19.390 –> 00:06:25.700
[email protected]: inserting a certain prompt about the audience. They’re trying to reach the emotion that they’re trying to convey.

36
00:06:25.820 –> 00:06:44.700
[email protected]: And then Prisado, using statistical analysis and machine learning. Now we refer to as motivation AI. In order to optimize language that an audience will see based on the audience’s historical Internet behavior, their interaction with the brand over time.

37
00:06:45.390 –> 00:06:58.630
[email protected]: What other? And this can be outside of marketing? But what other enterprise use cases and applications? Have you seen in the past few months that are really interesting or even slightly controversial.

38
00:06:58.830 –> 00:07:17.520
Paul Roetzer: Yeah, I mean, I I do think that most organizations we’re talking to are trying to solve for the the language model side, because it does touch every function of business. So anywhere where words are created marketing sales, Service, Ops. Hr. Finance legal. It’s going to affect

39
00:07:17.560 –> 00:07:37.320
Paul Roetzer: all of it all knowledge work, basically so I would say, a lot of organizations are just trying to get a grasp of what exactly is this technology, and what are they going to do now? We’ve seen really interesting examples like actually, I just saw one yesterday what Wendy’s is putting like

40
00:07:37.320 –> 00:07:50.220
Paul Roetzer: a chat interface basically at the drive through that’s powered by language model, or, you know, like it’s it’s trained on like the acronyms people use for the food. And so I think you’re gonna start seeing a lot of these higher profile examples

41
00:07:50.370 –> 00:07:53.790
Paul Roetzer: where people are going to start interacting with these AI agents.

42
00:07:53.870 –> 00:08:11.280
Paul Roetzer: and it’s going to become second nature, you know, having to think about it. But for right now I think a lot of corporations are really racing to figure out what to do on the language side. And do they need to train their own models? And where’s the data come from? And so that’s what a lot of our conversations are centering around right now.

43
00:08:12.050 –> 00:08:14.380
[email protected]: I agree. It’s a very interesting

44
00:08:14.500 –> 00:08:22.320
[email protected]: time to be working in the Enterprise space. We are having conversations daily with our clients around.

45
00:08:22.660 –> 00:08:27.240
[email protected]: how to integrate versato with other generative AI capabilities.

46
00:08:27.320 –> 00:08:35.530
[email protected]: even more long-term conversations around job displacement, new career creation

47
00:08:35.890 –> 00:08:42.370
[email protected]: and outside the scope of the work that we do it forsato. But a really interesting observation that i’m seeing

48
00:08:42.500 –> 00:08:52.620
[email protected]: in the news and in market, is what are the implications for education and job readiness in the next decade. So I I really do leave, and it sounds like

49
00:08:52.710 –> 00:08:56.210
[email protected]: in your experience. You’re You’re seeing something very similar

50
00:08:56.280 –> 00:09:04.340
[email protected]: that this wave of generative AI technology has very disruptive capabilities to accelerate time to market.

51
00:09:04.620 –> 00:09:13.520
[email protected]: amortize R&D. Expenses, but also produce very fresh new content for marketers and for other lines of business.

52
00:09:13.790 –> 00:09:27.100
Yeah, we’re definitely seeing it at the University level. I’ve had quite a number of conversations with different universities, and everyone is screaming. I even just hired like I’ve talked with high schools. I’ve even talked with primary school like principals and heads of

53
00:09:27.260 –> 00:09:41.500
Paul Roetzer: districts that just don’t know what to do like. They don’t understand the tech, and they have teachers. They, they may know, are using it, or maybe they don’t. They have students that they’re not sure if they’re using, or if they should encourage them to use it.

54
00:09:41.500 –> 00:10:00.300
Paul Roetzer: So it is very disruptive, and it is the challenges we still have to teach critical thinking like we we can’t have students coming through any little school. My kids are in fifth and fourth grade, and i’m very cautious and how to show them how to use the technology. And I’m very much teaching it more like a calculator at this point that I am.

55
00:10:00.300 –> 00:10:11.270
Paul Roetzer: They’re exposed to the technology, and they know it exists. I’m sure. Their friends, you know, are aware of it. But we’re using it as a tool to like check our work and and try creative alternatives to what they’re doing.

56
00:10:11.410 –> 00:10:26.190
Paul Roetzer: And then i’ll actually want them to analyze what the machine outputs. And then in a perfect world. You’re actually ta having them critically analyze how that output maybe, is different. And what did it do? Maybe better than we did? And what can we learn from it.

57
00:10:26.190 –> 00:10:33.010
Paul Roetzer: So I do think there is a very important conversation to be had about education. It’s moving slowly.

58
00:10:33.110 –> 00:10:41.290
Paul Roetzer: i’m. I’m guessing a lot of leaders of schools are going to spend the summer trying to figure this out. It kind of got thrown at them like in December, and

59
00:10:41.470 –> 00:10:52.850
Paul Roetzer: in the middle of the school year. You got all the sudden try and figure out how to transform education. That’s a big ask for these people. So I think you’re gonna have a a a summer where there’s gonna be a lot of work probably done around this space going into next school year.

60
00:10:52.960 –> 00:11:00.370
[email protected]: you know. I completely agree, and the the implications for how students interact with AI

61
00:11:01.260 –> 00:11:05.650
[email protected]: has a lot of crossover with how employees and workers will interact with AI.

62
00:11:06.240 –> 00:11:15.760
[email protected]: So I want to pivot back to the work that you’re doing. I know that you recently launched a series titled Piloting AI for marketers

63
00:11:15.890 –> 00:11:26.910
[email protected]: now without giving away the series in its entirety. What would you say? Are 3 key points, or takeaways that professional marketers should know

64
00:11:27.110 –> 00:11:32.590
[email protected]: to understand and optimize how they’re using artificial intelligence?

65
00:11:32.810 –> 00:12:00.340
Paul Roetzer: Yeah, I mean, it starts with a base, level understanding of what it is and how it works. So to me all of this is anything you do has to be premised on that. You have an understanding of what it is, and you can explain it at a you know, confidently explain it to your peers or to your your leaders in the organization. So you have to invest what’s needed to do it. Now you can get that pretty quickly like we do an intro day, I for marketers class. The last one we just did had over 1,300 people registered for it.

66
00:12:00.340 –> 00:12:08.560
Paul Roetzer: and that is like 30 min of me explaining the fundamentals and then doing Q. A. For 30 min. That could be enough like that can give you enough of what you need.

67
00:12:08.650 –> 00:12:17.120
Paul Roetzer: Then from there you can start getting into okay, how do I actually identify a used case and prioritize those use cases to pick them?

68
00:12:17.150 –> 00:12:35.020
Paul Roetzer: And then, once I have my used cases. Maybe it’s an eye writing tool, or a social media tool, or an analytics tool, or an advertising tool. Then how do I actually go find the right vendors to do this. How do I assess those vendors? So understanding of the technology, identification and prioritization of use cases, and then

69
00:12:35.230 –> 00:12:39.340
Paul Roetzer: knowing how to like really evaluate these vendors is is critical

70
00:12:40.280 –> 00:12:41.810
[email protected]: that makes sense, and

71
00:12:41.820 –> 00:12:48.730
[email protected]: really resonates with me, and a lot of conversations that I have with our clients on a pretty regular basis.

72
00:12:48.960 –> 00:13:00.080
[email protected]: One of the most common topics that we’ve come across in the last decade, and certainly more common now is the idea of building technology versus buying it.

73
00:13:00.430 –> 00:13:11.910
[email protected]: And I think the recent surge of interest in Generative AI, specifically chat Gpt has put some wind in the sales of the build versus by argument

74
00:13:12.700 –> 00:13:13.700
[email protected]: and

75
00:13:13.910 –> 00:13:19.910
[email protected]: the accessibility to creating code pseudo code

76
00:13:20.250 –> 00:13:33.370
[email protected]: functionality a lot more quickly than has been available to R. And D. Teams and marketers. I think, is really placed a lot more emphasis on homegrown technology versus to your point.

77
00:13:33.470 –> 00:13:36.250
Paul Roetzer: Selecting the right vendor, you know.

78
00:13:37.070 –> 00:13:46.750
[email protected]: taking a step back from that just in in general. How do you think generative AI is going to disrupt

79
00:13:46.820 –> 00:13:50.160
[email protected]: business models within the within the enterprise?

80
00:13:52.120 –> 00:13:54.850
Paul Roetzer: So I think what’s gonna happen is

81
00:13:55.440 –> 00:13:56.790
Paul Roetzer: we’re seeing

82
00:13:58.350 –> 00:14:08.180
in the past you had to go seek these tools out. You had to go find third-party applications that were smarter versions of what you were using or new kinds of technologies that enable you to do things.

83
00:14:08.310 –> 00:14:26.990
Paul Roetzer: What’s about to happen in 2,023? Is it just going to be infused into all your core tech platforms. So if you use box for knowledge management storage, they have box. AI. Now, where you can just have a conversation with your knowledge. Base, like whatever you’re looking for, and you know, query it you have. Slack is going to have Einstein Gpt

84
00:14:26.990 –> 00:14:37.180
Paul Roetzer: baked into it. Hubspot has chat spot salesforce as Einstein. Google Workspace is going to have language models, baked in microsoft 365 co-pilot all this year.

85
00:14:37.530 –> 00:14:40.970
Paul Roetzer: So if you think about any industry.

86
00:14:40.980 –> 00:14:50.580
Paul Roetzer: and the fact that all of your workers, including administrative staff, like everyone, is going to have access to generative tools baked right into the platforms they’re already using.

87
00:14:50.920 –> 00:15:01.060
Paul Roetzer: and they have no idea what those technologies are, nor how to use them, or how to change their processes based on them. And it’s just going to be like flipping a switch. And all of a sudden everybody’s got these capabilities

88
00:15:01.240 –> 00:15:04.540
Paul Roetzer: that that on its own is insanely disruptive.

89
00:15:04.630 –> 00:15:11.060
Paul Roetzer: Then you take it and say, okay, well smart entrepreneurs could look at this technology and say.

90
00:15:11.120 –> 00:15:19.060
Paul Roetzer: this whole industry is backwards and slow and inefficient. What if I just go find a couple of people. We just build a smarter version of this company.

91
00:15:19.480 –> 00:15:26.170
Paul Roetzer: I think you’re gonna see a lot of that in different industries, too like. There are some industries who are just ripe for disruption. They are

92
00:15:26.210 –> 00:15:37.940
Paul Roetzer: slow and inefficient by design, because that’s how they make money, and not to get like too specific. But law firms like they, They profit from inefficiency.

93
00:15:38.030 –> 00:15:44.360
Paul Roetzer: and that is just not going to cut it in the age of AI. So I. I just feel like there are some industries that

94
00:15:45.490 –> 00:15:59.220
Paul Roetzer: it’s very obvious. What’s about to happen? They don’t want it to happen, and they’re going to resist it and pretend like it’s not going to happen, but it’s coming, and it’s because it’s just going to be infused into everything we do, whether we seek it out or not. I think that’s the major change. Here

95
00:16:00.220 –> 00:16:04.290
[email protected]: I agree, and the the 2 axes that we

96
00:16:04.470 –> 00:16:14.390
[email protected]: talk about when we’re discussing the economics of generative AI with executives, that our clients are efficiency and efficacy.

97
00:16:14.490 –> 00:16:17.300
Paul Roetzer: So efficacy. You know, how do you

98
00:16:17.460 –> 00:16:23.230
[email protected]: improve the performance of an existing process or system, and then efficiency. How do you

99
00:16:23.640 –> 00:16:26.040
Paul Roetzer: produce something more quickly, you know.

100
00:16:26.090 –> 00:16:37.700
[email protected]: with less of an impact on your bottom line. And I think specific to service industries, Law firms are a great example where the unit of measure is the billable hour.

101
00:16:37.940 –> 00:16:45.580
[email protected]: The more they work, the more they get paid a solution like Chat gpt, or something to be released.

102
00:16:45.940 –> 00:16:56.610
[email protected]: can really threaten the business model like that where not only are the documents being prepared sharper than what may exist today, but they’re being produced

103
00:16:57.190 –> 00:16:59.250
[email protected]: in fewer available hours.

104
00:16:59.880 –> 00:17:15.349
[email protected]: Just one example, I am sure, in the coming years you’ll see this type of technology pro proliferate into even very advanced use cases like a pharmaceutical laboratory to be able to run synthetic experiments, accelerate

105
00:17:15.390 –> 00:17:18.200
[email protected]: R. And D. Cycles for medication

106
00:17:18.420 –> 00:17:25.690
[email protected]: which on one hand, can replace the humans who are working in those labs, on the other hand, has tremendous upside

107
00:17:25.790 –> 00:17:27.520
[email protected]: to be able to deliver.

108
00:17:27.550 –> 00:17:31.370
[email protected]: You know, very highly needed medicine out into the market.

109
00:17:32.080 –> 00:17:34.330
Paul Roetzer: There will be trade offs. No doubt

110
00:17:34.810 –> 00:17:38.680
[email protected]: there definitely will be trade offs. So let’s let’s talk about that a little bit.

111
00:17:38.760 –> 00:17:48.340
[email protected]: What what do you believe? The barriers to entry are for companies to invest in this type of technology.

112
00:17:48.600 –> 00:17:50.880
[email protected]: and then quick, follow on after that.

113
00:17:51.260 –> 00:17:56.050
[email protected]: If companies don’t invest in this technology, do they risk being left behind?

114
00:17:56.950 –> 00:18:11.070
Paul Roetzer: Yeah, I I mean to answer the the second question. First, My thesis is the future of all businesses. AI are obsolete. So either AI native built smarter from the ground up AI emergent you evolve to infuse AI across the organization, or you just become relevant

115
00:18:11.310 –> 00:18:22.840
Paul Roetzer: and depending on the industry that a relevancy could take years. If you’re a Sas company could be months or days like it. How quickly it happens is going to be based on what kind of industry you’re in, and how quick adoption is happening

116
00:18:23.030 –> 00:18:28.820
Paul Roetzer: Now the barriers, the first one that comes to mind to me is this is a complex buying environment.

117
00:18:28.990 –> 00:18:36.120
Paul Roetzer: There, there’s literally thousands of generative AI tools. And you just go. If you just focus on language or image, generation or vision like

118
00:18:36.150 –> 00:18:38.110
Paul Roetzer: it. It’s really hard

119
00:18:38.210 –> 00:18:52.180
Paul Roetzer: to to know who the players are. And so you’re now, especially in larger corporations. You’re likely starting with the known trusted entities. You’re going to go work with organizations who have a proven track record of working with larger enterprises

120
00:18:52.280 –> 00:18:57.550
Paul Roetzer: because they understand the dynamics of a business well beyond just your need for smarter technology.

121
00:18:57.630 –> 00:18:58.480
Paul Roetzer: So

122
00:18:58.620 –> 00:19:10.760
Paul Roetzer: I think that for a lot of organizations can be very hard to figure this out, because the tech is evolving so quickly, and the second piece that I I think a lot about is the lack of talent that understands this stuff.

123
00:19:10.970 –> 00:19:20.890
Paul Roetzer: So if you look around your your group in your company, or you know, even at you know, higher level. Go find somebody who can explain to you like what a language model is, or

124
00:19:20.910 –> 00:19:36.940
Paul Roetzer: how generally I works like they don’t exist like they they, they’re very hard to find. So you’re talking about like upscaling a workforce potentially hiring. But what do you hire like? I I don’t know, like they’re not coming out of universities trained on this stuff yet. So

125
00:19:36.960 –> 00:19:42.590
Paul Roetzer: it’s all new, and I think that’s the big. The Talent Gap is probably in it gonna end up being

126
00:19:42.630 –> 00:19:59.540
Paul Roetzer: the biggest barrier the other one that comes to mind. I know you all work with highly regulated industries. There are some industries, those aren’t going to be allowed to do this stuff like in the near term, like you, you’re gonna have governance, or regulations or laws that prevent you from benefiting from all the things that’s capable of

127
00:19:59.540 –> 00:20:03.880
Paul Roetzer: for different reasons within your industry. And so that’s a reality, too.

128
00:20:04.590 –> 00:20:08.160
[email protected]: I think this opens up a lot of

129
00:20:08.540 –> 00:20:29.670
[email protected]: philosophical questions around how humans and AI interact. One of the things we haven’t touched on, which I think is fascinating is, who does the intellectual property belong to you? Brought up a great example about the the drake and the weekend song which has been bouncing around, spotify and soundcloud a couple of other streaming platforms.

130
00:20:29.730 –> 00:20:36.060
[email protected]: just using that as a hypothetical. Who owns that IP? If someone, If someone

131
00:20:36.260 –> 00:20:42.460
[email protected]: generated that using one of these, you know, thousands of of AI software. I think

132
00:20:42.560 –> 00:20:54.840
[email protected]: the questions that are going to be asked in industry, around regulation and copyright and intellectual property ownership are going to be new, and I think that it’s a totally green field

133
00:20:54.980 –> 00:21:00.490
[email protected]: opportunity, and I agree with you. I think it’s AI or obsolescence.

134
00:21:01.340 –> 00:21:02.080
Paul Roetzer: Yep.

135
00:21:02.500 –> 00:21:06.640
Paul Roetzer: So on the ownership front that I mean the latest guidance from the copyright offices.

136
00:21:06.760 –> 00:21:19.390
Paul Roetzer: In the case of Drake no one owns it. So there the Copyright office on March sixteenth, 2,023 issued updated guidance. That, said a prompt is not considered human authorship, and only a human in on a copyright.

137
00:21:19.690 –> 00:21:43.350
Paul Roetzer: So if it’s images, videos, text logos, anything you create. If AI created it, you cannot copyright it, and I can steal it and put it on hats and sell it, and like that, that there is no protection. Basically, I always at this moment say, talk to your IP attorney like I am not an IP attorney. I have spent lots of money with IP attorneys, and learned a few things along the way. But this is not my ex Don’t. You know domain of expertise.

138
00:21:43.350 –> 00:21:55.840
Paul Roetzer: but that is the guidance. I’m I’m. Passing along, I would say, from the copyright offices, do not assume you on it, and if you hire agencies to do work for you. This is a really important one, and you have a work for higher agreement where the copyright passes to you.

139
00:21:55.840 –> 00:22:10.080
Paul Roetzer: If that agency unbeknownst to you is using AI to generate these things, then you technically Don’t actually on the copyright that you think you own. So it’s an important issue for people to to wrap their heads around. And I say, talk to IP. Attorneys like

140
00:22:10.810 –> 00:22:18.030
[email protected]: It’s none of us really like to have to go talk to attorneys, but this one that’s it’s good good advice, especially because

141
00:22:18.100 –> 00:22:25.640
[email protected]: this technology is being used sometimes without the for lack of a better term. Corporate chain of command being aware.

142
00:22:25.990 –> 00:22:27.960
[email protected]: and we have seen

143
00:22:28.160 –> 00:22:37.090
[email protected]: in especially highly regulated industries, that we work with bands on Chat Gpt. For exactly the reason that you’re bringing up, because the

144
00:22:37.470 –> 00:22:50.410
[email protected]: the understanding of intellectual property and regulations is so new and so opaque that executives are putting a full stop to using this type of technology for fear of falling into one of these traps.

145
00:22:50.780 –> 00:22:51.450
Paul Roetzer: Yup.

146
00:22:51.670 –> 00:22:56.390
[email protected]: So, pivoting more to a a topic closer to home. For myself.

147
00:22:57.320 –> 00:23:12.570
[email protected]: I know that You’ve You’ve worked with Prisado for a variety of years. You’ve been a great partner of ours. We, as I mentioned earlier, pride ourselves on having a decade of experience in what is now referred to as the generative AI industry.

148
00:23:13.000 –> 00:23:15.560
[email protected]: We’ve been taking human prompts.

149
00:23:16.220 –> 00:23:21.620
[email protected]: personalizing emotions and text based on the intended recipient

150
00:23:22.170 –> 00:23:29.590
[email protected]: in as a sub sector of generative AI, that we refer to now as motivation. AI.

151
00:23:30.520 –> 00:23:37.890
[email protected]: So that’s how we position ourselves differently in your experience. Working with Prasad over the past couple of years.

152
00:23:38.080 –> 00:23:42.440
[email protected]: What do you perceive to be our unique differentiator in the marketplace?

153
00:23:42.840 –> 00:23:53.190
Paul Roetzer: Yeah. So there’s 2 things that seem to differentiate companies right now and be defensible in this. The software world and those are data and distribution meaning.

154
00:23:53.610 –> 00:23:58.990
Paul Roetzer: Yeah, let’s say you an AI writing tool like they want to use it the right blog post whatever

155
00:23:59.030 –> 00:24:17.790
Paul Roetzer: I could use chat chpt as of November of 2,022. I could use, you know, any number of other players. But then grammarly showed up, and grammarly is 30 million customers. That’s distribution like you can show up late to the party, and you got massive audience, and you can introduce a tool, and you can be a winner.

156
00:24:17.860 –> 00:24:24.530
Paul Roetzer: What Persado has always had is the data, and I mean distribution as well, like a proven customer base. But

157
00:24:24.710 –> 00:24:42.080
Paul Roetzer: what the way these tools work is, They make predictions about words based on general knowledge. They go learn the Internet what they don’t have is performance data. So when it’s generating things for you, it doesn’t know if what it’s writing actually will work. It just knows that it looks like what it learned from.

158
00:24:42.170 –> 00:24:51.240
Paul Roetzer: And those can sound like really good emails. They can look like really big landing pages or blog posts or social shares or ads, or whatever you’re creating

159
00:24:51.290 –> 00:24:59.210
Paul Roetzer: the language models, and these, like app companies that are building interface on top of them, can look really good and sound. Really good.

160
00:24:59.410 –> 00:25:03.200
Paul Roetzer: Are they actually going to perform? They have no idea

161
00:25:03.290 –> 00:25:10.690
Paul Roetzer: they have no performance data to tell them it’s going to perform, and that’s where Prisado shines is that’s what it was built on was

162
00:25:11.080 –> 00:25:25.170
Paul Roetzer: performance data like learned performance data, and what words work and what words motivate action, and that is probably like the the the most valuable asset for Prisado is that that’s what you have is data that these other models just Don’t have.

163
00:25:25.810 –> 00:25:29.570
Paul Roetzer: So I would see it, at least from like an outside or objective perspective.

164
00:25:29.690 –> 00:25:33.640
[email protected]: I think I think it’s a very important positioning statement. We

165
00:25:33.690 –> 00:25:39.120
[email protected]: We do continually train the model on hundreds of millions of digital interactions

166
00:25:39.710 –> 00:25:50.330
[email protected]: distribution wise. Our Our core use case was a couple of years ago. In email we expanded into social SMS web.

167
00:25:50.640 –> 00:25:57.580
[email protected]: We’re moving even further down funnel into optimizing shopping cart based on past abandonment behavior.

168
00:25:57.650 –> 00:26:05.700
[email protected]: So I think our distribution is, you know, to your point, also a very strong competitive position for us as well.

169
00:26:05.930 –> 00:26:10.420
[email protected]: and you know we’re we’re very excited about where the company is is today

170
00:26:10.640 –> 00:26:13.400
[email protected]: about where it’s heading. You know we

171
00:26:13.780 –> 00:26:20.590
[email protected]: not to give anything away. But you know we are also really looking at the idea of

172
00:26:20.920 –> 00:26:32.490
[email protected]: higher level. More abstract prompts that creative teams can use to brainstorm campaigns. So it’s not necessarily just an input output. It is more of a dialogue with the machine.

173
00:26:32.670 –> 00:26:35.550
[email protected]: So I can interact with the software. And say.

174
00:26:35.740 –> 00:26:38.080
Paul Roetzer: yeah, like the reduce

175
00:26:39.090 –> 00:26:42.360
Paul Roetzer: where you started saying that I think my Internet might have dropped for a second.

176
00:26:42.410 –> 00:26:47.800
[email protected]: Okay, I just because you broke up as you’re starting to say that last part.

177
00:26:48.110 –> 00:27:03.430
Paul Roetzer: So if we go back to like after I finish the differentiation, let me just run a check real quick on my Internet. Okay, yeah, it seemed all clear on my side. But I all I can start over. Yeah, cause you’ve you’ve frozen. So I wasn’t sure if it was you or me? I didn’t actually hear your answer to that.

178
00:27:03.690 –> 00:27:06.240
Paul Roetzer: No worries. Let me just check it real quick.

179
00:27:17.270 –> 00:27:27.400
Paul Roetzer: Yeah, i’m, I’m assuming okay, though, again, yeah, I mean. And you’re You’re You’re loud and clear. Can you? Can you hear and see me fine? Yeah, you just froze for a second

180
00:27:27.570 –> 00:27:30.390
[email protected]: ready to get going in about 3 s.

181
00:27:30.850 –> 00:27:33.470
Paul Roetzer: Yeah, go for it seems okay, perfect.

182
00:27:34.900 –> 00:27:56.950
[email protected]: So, Paul, I think that’s great competitive positioning for Prisado to build on your answer. Distribution, I think, has been one of the the strengths that we focused on building in the last couple of years. If you were to look at Prisado 5 years ago our primary use case was an email. We’ve since expanded to social to web.

183
00:27:57.130 –> 00:28:02.660
[email protected]: And then we’ve moved even further down funnel to focus on the shopping cart page and specific

184
00:28:03.210 –> 00:28:15.010
[email protected]: addressing past abandonment behavior based on demographics based on shopping trends. And we’re not to give anything away but really focused on

185
00:28:15.410 –> 00:28:25.860
[email protected]: How do you abstract the idea of an input and an output? So creative teams can effectively brainstorm the types of campaigns and emotions that they want to use

186
00:28:25.920 –> 00:28:28.280
[email protected]: when interacting with their audience. So

187
00:28:28.300 –> 00:28:31.620
[email protected]: that’s more to more to come, for sure. But I think.

188
00:28:31.700 –> 00:28:37.370
[email protected]: leading the charge and and adopting similar trends to the rest of the industry.

189
00:28:37.520 –> 00:28:46.340
[email protected]: understanding how marketing and data teams can work well together and human and machine can work alongside each other effectively as well.

190
00:28:47.030 –> 00:28:50.260
Paul Roetzer: Yeah, I would think that the the data alone is

191
00:28:50.970 –> 00:28:56.460
Paul Roetzer: grounds for excitement. I mean, I I do think you have what a lot of these companies are going to be looking for.

192
00:28:56.830 –> 00:28:57.930
[email protected]: Absolutely

193
00:28:58.110 –> 00:29:05.140
[email protected]: So Speaking of the excitement, Paul, and thank you again for for coming on the podcast. This has been a great conversation.

194
00:29:05.420 –> 00:29:15.540
[email protected]: What does the future of the Marketing AI Institute have in store. And is there anything else that you’re working on, that you want to talk about with the audience?

195
00:29:15.900 –> 00:29:29.190
Paul Roetzer: Yeah. And our big event is the marketing AI Conference every year. So it’s July and Cleveland, and this this one’s shaping up to be the certainly the biggest one yet. It’s already the biggest one yet, I think, in terms of the sales.

196
00:29:29.260 –> 00:29:42.280
Paul Roetzer: So we’re seeing a whole new level of interest around that so really excited about that. And then I think just events overall. We did an a. For writers summit in March. We thought we’d get a 1,000 people. We have 4,200,

197
00:29:42.330 –> 00:29:46.720
Paul Roetzer: so I just so much like of a ground swell of interest.

198
00:29:46.780 –> 00:29:58.450
Paul Roetzer: and we have that we have the AI for agency summit in the fall, and we’re looking at some other like kind of expanding that strategy because it’s just the quick way to get this information to people and help them.

199
00:29:58.540 –> 00:30:06.770
Paul Roetzer: And then online education. Our piloting AI for Marketer series you mentioned is sort of the the the core offering at the moment.

200
00:30:06.910 –> 00:30:19.640
Paul Roetzer: But we have some pretty cool ideas around some other original series that we’re working on. So I think there’s just going to be a lot of focus on creating as much value as possible through the events and online education and continuing to expand those areas

201
00:30:20.310 –> 00:30:30.260
[email protected]: fantastic. And for those in the audience, i’m sure there will be many who would benefit from working with Maii. What’s what’s the best way to engage you? What’s the best way they should reach out?

202
00:30:30.260 –> 00:30:47.720
Paul Roetzer: Yeah, marketing AI institute.com is where everything’s housed, the book, the courses, the newsletter, the podcast, the blueprints about different industries. You can find it all there and then personally, i’m very active on Linkedin and Twitter. So either of those linkedin’s, great though, and let me know you heard, you know, on the podcast and

203
00:30:47.900 –> 00:30:52.320
Paul Roetzer: definitely connect. So those are the 2 main areas you will kind of connect with me

204
00:30:52.780 –> 00:31:05.180
[email protected]: fantastic. Well, Paul, really appreciate you coming on again, for the audiences with Paul Rateser, founder and CEO of the Marketing AI Institute and 3 time Author.

205
00:31:05.440 –> 00:31:23.060
Paul Roetzer: No more off. I don’t think we’re gonna more of those coming any time. I was. I was gonna ask if there’s a fourth book coming. It’s like we can. We can check in later. Maybe some digital books. I don’t know a lot of print book anytime, maybe author. But by Chat Gpt: Who knows? Yeah, there we go.

206
00:31:23.370 –> 00:31:26.650
Paul Roetzer: Thanks a lot. Paul is a pleasure.