Fireside Chat with Scott Brinker & Kerry Cunningham — The Agentic Marketing Stack

1226
03:31:33.460 –> 03:31:34.930
Scott Brinker: Great to have you.

1227
03:31:34.930 –> 03:31:35.360
Julia Nimchinski: here.

1228
03:31:35.360 –> 03:31:38.050
Scott Brinker: Hi, Kerry, great to be here with you.

1229
03:31:38.660 –> 03:31:39.400
Julia Nimchinski: Bye.

1230
03:31:39.980 –> 03:31:40.360
Scott Brinker: Man.

1231
03:31:40.360 –> 03:31:40.700
Julia Nimchinski: Amazing.

1232
03:31:41.150 –> 03:31:45.990
Scott Brinker: It’s like just drinking from a fire hose. This is, like, awesome how you set these things up.

1233
03:31:47.470 –> 03:31:52.029
Julia Nimchinski: Amazing, welcome to the show, the Gentec Marketing Stack. Let’s go!

1234
03:31:53.460 –> 03:31:56.969
Scott Brinker: Right. Terry, I so look forward to having this chat with you.

1235
03:31:57.400 –> 03:32:12.610
Kerry Cunningham: Yeah, this is… this is really fun. This is great for me. And so, first of all, I thought maybe, you could just tell everybody what you’re doing. I… we know you did a career transition, recently that we saw, so, what are you focused on? And then I have a question also that I’m going to ask I to tell you about.

1236
03:32:12.610 –> 03:32:30.100
Scott Brinker: Okay, sounds great. Well, for the past 18 years, I’ve sort of been, you know, this armchair MarTech analyst, writing chief Martech. During the first part of that, I was also building a SaaS company myself on interactive, Interactive Content Platform. I then joined HubSpot in 2017 to help them

1237
03:32:30.100 –> 03:32:37.440
Scott Brinker: create a true open platform ecosystem. Did that for 8 years. Took us from around 50 integrations to over…

1238
03:32:37.460 –> 03:32:52.909
Scott Brinker: 2000? And just a couple weeks ago, decided, okay, this was a good inflection point for me to move on. And so at the moment, I’m just deep diving on all the things I need to catch up on with, like, AI and MarTech that, you know, changed in the past week.

1239
03:32:53.860 –> 03:33:13.489
Kerry Cunningham: Amazing, amazing. So, great to catch up with you. And, you know, when I first saw your announcement, that you were leaving HubSpot, my reaction was, wait a minute, you did that in 8 years? That seemed like decades, and when you see the growth, even of what you’ve done with Chief Martech and the list of

1240
03:33:13.810 –> 03:33:18.570
Kerry Cunningham: Companies, that’s at 15,000-something, right now.

1241
03:33:19.260 –> 03:33:20.640
Kerry Cunningham: Pretty staggering.

1242
03:33:21.270 –> 03:33:29.300
Scott Brinker: By the way, I should just give a shout out, for Sixth Sense, as a partner of HubSpot. That was a… yeah, love it. Still a great partnership, so, anyway.

1243
03:33:29.300 –> 03:33:29.840
Kerry Cunningham: Thanks.

1244
03:33:30.750 –> 03:33:43.949
Kerry Cunningham: One, one… and so this is, this is something I think everybody’s, dealing with, and this is a slightly off the AI and agents topic, but, in your last report, there was a, a passage that read something like,

1245
03:33:44.250 –> 03:33:54.319
Kerry Cunningham: If success is redefined as a profitable business with happy customers and happy employees, we think there will be many successful MarTech ventures ahead.

1246
03:33:54.860 –> 03:33:58.279
Kerry Cunningham: Well, one, I hope so, because that’s kind of where we grew up, right?

1247
03:33:59.850 –> 03:34:00.380
Kerry Cunningham: What?

1248
03:34:00.480 –> 03:34:05.249
Kerry Cunningham: You know, and showing… and showing my age, it’s like, wow, we’re redefining

1249
03:34:05.420 –> 03:34:15.490
Kerry Cunningham: a successful business as one that’s profitable and has happy customers. It seems like maybe not being defined as that has been something of an issue.

1250
03:34:15.490 –> 03:34:30.680
Scott Brinker: Well, I think, alright, so the context of that was, yeah, okay, this crazy, crazy growing MarTech landscape, which actually now, with all the AI capabilities for software development, I mean, the truth is that long tail is now stretching into infinity.

1251
03:34:30.760 –> 03:34:46.260
Scott Brinker: In fact, you could have a discussion, there’s probably, like, almost a bifurcation here. There’s a set of products and platforms that are going to continue to be very large-scale products and platforms that very much serve as the center of gravity within people’s tech stacks.

1252
03:34:46.290 –> 03:35:05.669
Scott Brinker: And there will continue to be consolidation forces that happen, you know, in that head of the tail. But the long tail, I think, is just gonna keep getting longer and longer. And for a while, I think a lot of people thought of that as, like, a negative thing in multiple ways. Like, oh, well, wait, isn’t the whole point of building a business, is you’re gonna raise a bunch of money, and you’re gonna become a multi-billion dollar business?

1253
03:35:05.670 –> 03:35:06.140
Kerry Cunningham: Yeah, yeah.

1254
03:35:06.140 –> 03:35:18.760
Scott Brinker: surely that’s the only kind of software business you could ever want. And we’re actually saying, no, no, you actually have a bunch of people who build very focused, you know, software businesses, they have a small target market.

1255
03:35:18.880 –> 03:35:28.809
Scott Brinker: Their customers love them, they love what they’re doing. You could actually argue this whole conversation that’s happening these days around, it’s not software as a service, but service as a software.

1256
03:35:28.810 –> 03:35:45.590
Scott Brinker: You start going down that path, you realize, oh my god, the world is full of tens of thousands, hundreds of thousands, millions of, like, services businesses as they become more and more software-enabled, and, like, where the boundary lines between them are. I almost think it doesn’t really matter.

1257
03:35:45.590 –> 03:35:46.590
Kerry Cunningham: Doesn’t matter.

1258
03:35:46.590 –> 03:35:52.189
Scott Brinker: You have a business that, yeah, you love what you do, and customers love what you do, too.

1259
03:35:52.190 –> 03:36:02.130
Kerry Cunningham: Yeah. You know, I think that’s an interesting, point to think about, because I think as we look at the role of AI and agentic AI in business, and particularly within

1260
03:36:02.220 –> 03:36:13.410
Kerry Cunningham: MarTech, one of the things that seems to me has gotten lost over the years, you know, I’ve spent almost my entire career in and around B2B tech companies as well.

1261
03:36:14.000 –> 03:36:26.760
Kerry Cunningham: It seems one of the things that gets lost is the concept of what we need to accomplish, as opposed to what we can accomplish or can do. And the technology brings in a lot of the what can we do.

1262
03:36:27.140 –> 03:36:44.879
Kerry Cunningham: But it’s the people who run the business who have to be thinking about the what should we be doing? How should we be identifying our buyers? How should we be serving our customers? And I think that that gets really easily lost in

1263
03:36:44.980 –> 03:36:49.380
Kerry Cunningham: All of the mayhem around what can we do, right?

1264
03:36:50.770 –> 03:36:55.300
Scott Brinker: Yeah, no, I think that’s true. I mean, it’s interesting, again now.

1265
03:36:55.300 –> 03:37:13.660
Scott Brinker: this black background, like, really emphasizes the gray hairs, you know, but maybe one of the small advantages of the gray hairs is having seen, like, a certain pattern of technology adoption that’s repeated again and again, and, you know, it’s not that every one is exactly the same, but there are some common patterns, and one of them is…

1266
03:37:13.660 –> 03:37:16.770
Scott Brinker: Boy, most technology implementations

1267
03:37:17.100 –> 03:37:25.220
Scott Brinker: either fail, or they don’t work out really well, and it’s not really the technology. It’s almost always the organizational thing of, like, well, we weren’t really clear about

1268
03:37:25.290 –> 03:37:44.789
Scott Brinker: what we wanted to do with that, or we didn’t think about, like, well, what sort of skills or talents do we need? Or, you know, like, how do we even organize our business so that other things that might have been, like, bottlenecks? And we’re certainly living through this right now with AI, particularly because of the fact that so many of these AI technologies, one of the things they deliver in spades is speed.

1269
03:37:44.790 –> 03:37:50.669
Scott Brinker: You know, like, there’s things that used to take a long time that are increasingly now able to be done in a matter of minutes.

1270
03:37:50.810 –> 03:37:54.699
Scott Brinker: Okay, well, to really harness that in a coherent.

1271
03:37:54.700 –> 03:38:10.930
Scott Brinker: non-chaotic way, you know, requires us to rethink a lot of, like, how people work together, how they coordinate, how’s there some coherence, you know, to what the organization is doing, when all of a sudden you have the ability to, like, you know, run the movie at 10x speed.

1272
03:38:11.250 –> 03:38:30.819
Kerry Cunningham: Yeah, and I think, we’re not here to put together a framework for AI adoption, but I don’t know that there would be a better person to do it, so while we’re talking about it, it does seem to me that as organizations think about, and we’ll talk about, you know, use cases and stuff in a minute, but as organizations think about what they can do.

1273
03:38:30.820 –> 03:38:44.030
Kerry Cunningham: with AI and what they should be doing with agents, the first thought always needs to go back to what do we really need to do for our customers, for shareholders? You know, what are the things that…

1274
03:38:44.030 –> 03:38:59.149
Kerry Cunningham: that, we shouldn’t just do faster because we can do them faster, we should stop doing them at all. You know, would be… there’s a lot of things that we should just stop doing altogether. And so I think I… I hope, anyway, that our audience can think about that.

1275
03:38:59.270 –> 03:39:02.090
Kerry Cunningham: As we transition to talk about, okay.

1276
03:39:02.810 –> 03:39:21.200
Kerry Cunningham: what can they do, what do they do? And so, there have been about 15 different definitions of AI agents today so far. I would like to hear yours, because yours is one that I think, I think our audience probably ought to pay attention to.

1277
03:39:21.200 –> 03:39:34.130
Scott Brinker: Very kind there. I’m actually not super religious about the definition, because I agree with you, there are a lot of definitions, and I think for me, the clearest way is to think of it on a spectrum, where, you know, one end of the spectrum, where we’ve been

1278
03:39:34.230 –> 03:39:45.729
Scott Brinker: grown quite comfortable is this idea of, you know, very rules-based, deterministic automation. Right. And that still has a big role to play in most businesses. You know, there’s this…

1279
03:39:45.730 –> 03:40:04.449
Scott Brinker: other extreme, you know, that is largely still science fiction of, like, oh, well, I just tell the agent, hey, I would like an amazing, you know, marketing campaign, go off, create it, do it, and if you just want to put deposit the money from that, here’s my account, have a nice day. You know, we’re a long way away from

1280
03:40:04.450 –> 03:40:08.610
Scott Brinker: That, thank goodness, most marketers are like, alright, gotta…

1281
03:40:08.610 –> 03:40:13.520
Scott Brinker: But, like, in between that, we’re talking about probably the degree to which

1282
03:40:13.540 –> 03:40:17.500
Scott Brinker: The amount of autonomy that something has, and also…

1283
03:40:17.510 –> 03:40:23.210
Scott Brinker: the scope of what it does. You know, so for instance, like, some of the examples of

1284
03:40:23.210 –> 03:40:25.759
Scott Brinker: Agents, that are being used.

1285
03:40:25.760 –> 03:40:45.139
Scott Brinker: Is, like, if I give an agent, like, hey, this is a customer that I’m interested in, can you go and research this customer on the web? And let it actually make some choices about, like, okay, which websites does it visit? As it starts to crawl through the customer’s website, you know, like, which pages are it paying attention to, how’s it synthesizing that?

1286
03:40:45.440 –> 03:40:49.160
Scott Brinker: That’s actually a non-trivial amount of autonomy.

1287
03:40:49.160 –> 03:40:49.480
Kerry Cunningham: Yeah.

1288
03:40:49.480 –> 03:40:57.229
Scott Brinker: Because it could go a million different ways, and what you get back could have high variance in, like, then what you decide to do based on that information.

1289
03:40:57.370 –> 03:40:59.340
Scott Brinker: But that is, nonetheless.

1290
03:40:59.460 –> 03:41:07.759
Scott Brinker: a relatively constrained box in which I think we’re gonna increase the capital, saying, like, yeah, that’s a great place to let, you know, an AI agent do some autonomy.

1291
03:41:07.780 –> 03:41:24.090
Scott Brinker: But it gets triggered in a more deterministic workflow that figures, this is the moment when we do that, you know, and on the other side of that, when it comes back, we maybe have a, you know, still relatively structured process of, okay, what are we going to do with that information? So I don’t know, to answer your question.

1292
03:41:24.090 –> 03:41:24.579
Kerry Cunningham: Yeah, yeah.

1293
03:41:24.580 –> 03:41:28.449
Scott Brinker: It’s just something that has some autonomy, but it doesn’t have to have a lot.

1294
03:41:29.100 –> 03:41:39.259
Kerry Cunningham: Yeah, so good. And I actually like that, and that’s kind of what I thought you might say, because I think, one, everybody’s got to know we have to have a grounding in what we’re talking about at the same time.

1295
03:41:39.260 –> 03:41:51.710
Kerry Cunningham: I don’t think anybody should care very much, right? It’s, you know, there are things that we can automate in, as you’re talking about, a deterministic way, if this, then that, make that happen.

1296
03:41:51.710 –> 03:42:04.830
Kerry Cunningham: And then, now, there’s the ability to have the machine that’s doing the automation supply more of the logic, supply more of the steps, you don’t have to determine each one of those in advance. So that’s great.

1297
03:42:04.830 –> 03:42:15.329
Kerry Cunningham: Now, my experience in B2B is that B2B leaders are not great at probabilistic, that, you know, we really like. Deterministic.

1298
03:42:15.350 –> 03:42:24.390
Kerry Cunningham: And the outcomes from the machine, if they… if there’s variability in the outcome of the machine that we don’t anticipate, I think we’re gonna…

1299
03:42:24.840 –> 03:42:35.489
Kerry Cunningham: struggle with that, a little bit. And I think that may not be built into the expectations that a lot of people have about what agents are going to do for them, is this idea that.

1300
03:42:35.890 –> 03:42:54.309
Scott Brinker: Put your finger on exactly the weirdest thing that everyone is struggling with in the mental model, is we are so used to computers in particular, being just this incredibly deterministic, like, if-then-the-else machine. And so this whole idea of thinking probabilistically,

1301
03:42:54.310 –> 03:43:00.440
Kerry Cunningham: Yeah, it’s just very foreign, and it feels very uncomfortable for all the operations and infrastructure we’ve put in place.

1302
03:43:00.700 –> 03:43:03.720
Scott Brinker: Being said, we step back, and we look at

1303
03:43:03.730 –> 03:43:19.029
Scott Brinker: the real world, which actually does exist beyond the domain of, you know, this digital microcosm that we felt we’ve so tightly controlled. In the real world, yeah, the relationships between customers and how they make decisions and how we interact have probabilistic components like just…

1304
03:43:19.030 –> 03:43:28.019
Scott Brinker: strewn throughout it. So, I don’t know, it’s weird trying to, like, adjust to this, but I don’t actually think it’s as fatal of a flaw as

1305
03:43:28.050 –> 03:43:32.849
Scott Brinker: Perhaps we might have that initial reaction of, like, wait, the computer shouldn’t act, you know.

1306
03:43:32.850 –> 03:43:43.699
Kerry Cunningham: Yeah, but I think that’s it, right? It’s, we expect the computer to come back with the right answer and exactly the right thing every time, and a lot of people find it super off-putting if it doesn’t.

1307
03:43:43.850 –> 03:43:54.160
Kerry Cunningham: And I think, just as a culture, I guess we’re going to have to adapt our expectations, so that we don’t start throwing out

1308
03:43:54.370 –> 03:43:59.089
Kerry Cunningham: What would be better results than we’re gonna get otherwise, because they’re not perfect.

1309
03:43:59.740 –> 03:44:10.339
Scott Brinker: Well, and that’s why I think one of the easiest first steps for companies, like, adapting some of these, you know, AI capabilities is giving the AI tasks

1310
03:44:10.440 –> 03:44:19.740
Scott Brinker: that actually we weren’t really able to do or do well before. Like, the perfect example, you know, that research thing was one. Another is, like, all this unstructured data, like, you know, hey.

1311
03:44:20.270 –> 03:44:32.389
Scott Brinker: have, like, a call transcript, I’m gonna have a string of emails. Historically, we didn’t really have the mechanisms in the computer to say, like, you know, analyze this, you know, summarize this thing, make a decision based on that.

1312
03:44:32.670 –> 03:44:50.700
Scott Brinker: technically, it took a lot of effort, very few… it was not accessible to most folks. And now, it’s, like, pretty much accessible to anyone. The fact that it’s kind of a new thing, and it’s dealing with a kind of interaction that almost we don’t expect to be as deterministic… I don’t know, it feels like…

1313
03:44:50.700 –> 03:45:01.890
Kerry Cunningham: A nice first step to be like, oh, okay, well, this is a useful new thing to add into the toolbox before we take things that used to be deterministic and start to make them more probabilistic.

1314
03:45:02.190 –> 03:45:07.889
Kerry Cunningham: Yep, absolutely. And I think, you know, there are occasions, like, for instance, right now, in our research work.

1315
03:45:07.890 –> 03:45:22.880
Kerry Cunningham: we abhor open text fields, right? You can’t really get good data out of it, and most of it’s… but that’s not true anymore, and we’ve actually been a little bit slow to respond to that. Like, of course we can get good data out of it now. We don’t have to comb through it ourselves anymore.

1316
03:45:22.880 –> 03:45:26.000
Kerry Cunningham: So there are those places where you have to ask yourself.

1317
03:45:26.000 –> 03:45:41.720
Kerry Cunningham: you know, what are the things that we do really need to do that we either haven’t done, or we’ve done badly, or that take too long, we’ve ignored? And I think we have to continually revisit, a lot of the assumptions that we’ve made about our businesses, and ask.

1318
03:45:41.930 –> 03:45:52.650
Kerry Cunningham: can we do that now? But starting with the things that we want to be able to do, right? That we know we should be able to do, rather than the thing that we can do, because the machine can do it.

1319
03:45:52.960 –> 03:46:07.310
Kerry Cunningham: So, on the topic of what can and what should it do, maybe talk about what you think you see as the most important applications or use cases for AI agents in marketing.

1320
03:46:08.260 –> 03:46:10.090
Scott Brinker: Yeah, so,

1321
03:46:10.200 –> 03:46:29.929
Scott Brinker: And boy, it keeps changing, you know, every week. What was it before you got it? I think for me, because I like to try and, like, organize things into, you know, buckets, hence that crazy Martech landscape. You know, I sort of think of there being, like, 3 large buckets of AI agent technology that’s having an impact on marketing.

1322
03:46:29.930 –> 03:46:38.780
Scott Brinker: So one, which is probably the largest and the one we have the most discussions around, are these internal AI agents. They’re agents that are operating

1323
03:46:38.780 –> 03:46:56.399
Scott Brinker: behind the scenes with the marketer to help everything from, like, brainstorming ideas to, you know, doing analysis, implementing the production, getting things live. Lots of great stuff. Keeps expanding. Like, the scope of, like, what people are able to do with these agents, awesome. But it’s behind the scenes.

1324
03:46:56.560 –> 03:47:05.040
Scott Brinker: There’s a second category of agents, which are still controlled by the company, by marketers, but they are customer-facing, the most obvious being

1325
03:47:05.220 –> 03:47:08.510
Scott Brinker: The chatbots that are ubiquitous on websites, and…

1326
03:47:08.510 –> 03:47:09.030
Kerry Cunningham: Yep.

1327
03:47:09.030 –> 03:47:09.970
Scott Brinker: Let’s be honest.

1328
03:47:10.310 –> 03:47:23.420
Scott Brinker: those have largely sucked for many, many years. But, you know, in this past year, they’ve actually started to get really good, because leveraging these LLMs, you know, they’re very good now at that sort of conversational interaction pattern

1329
03:47:23.420 –> 03:47:31.480
Scott Brinker: But also, companies are getting better and better at connecting those to the right back-end data in, like, you know, our knowledge bases, ticket histories, things like this.

1330
03:47:31.480 –> 03:47:38.030
Scott Brinker: that can actually provide useful answers, and so this is a great win, I think, for both most customers and companies.

1331
03:47:38.190 –> 03:47:53.870
Scott Brinker: There’s other kinds of agents we’re playing with that are customer-facing. Probably one of the more controversial ones would be the whole AI SDR, you know, thing. But, anyways, this is a whole category of, like, okay, these are agents, but they’re designed to interact

1332
03:47:54.080 –> 03:48:07.900
Scott Brinker: With other generally human customers at the moment. And then there’s this third bucket, which to me, in many ways, is the most fascinating and potentially the most disruptive, which is AI agent technology

1333
03:48:08.150 –> 03:48:32.870
Scott Brinker: that actually the marketers don’t control, that is running in… on behalf of the buyers, you know? And we’re just starting to see that. Like, an obvious example in, you know, like, B2B would be, like, okay, if I now turn to ChatGPT deep research, and I’m like, listen, I want you to go out and analyze, you know, what is the state of, you know, AI capabilities and CRMs, and who’s doing interesting things, and how should I weigh and evaluate

1334
03:48:32.870 –> 03:48:33.490
Scott Brinker: that.

1335
03:48:33.800 –> 03:48:38.930
Scott Brinker: Oh my goodness, that thing goes off, and it’s pulling stuff from all the websites, it’s going to third.

1336
03:48:38.930 –> 03:48:39.300
Kerry Cunningham: Yeah.

1337
03:48:39.300 –> 03:48:47.060
Scott Brinker: Of course, you know, you start to see also, you know, things like, coming up, like, perplexity shopping and all these things. Like, we’re getting to a…

1338
03:48:47.250 –> 03:48:59.190
Scott Brinker: We’re on the cusp here of buyers being able to turn to more and more AI to, play a more intermediary role in the, you know, buyer-seller relationship.

1339
03:48:59.660 –> 03:49:14.969
Scott Brinker: that’s gonna be really interesting for us as sellers. In fact, actually, that whole AEO, GEO, take your pick, the, you know, optimizing content, you know, for these AI engines, in many ways, right, this is part of us, like, starting to adapt of, like, okay, it’s not just Google.

1340
03:49:14.970 –> 03:49:22.920
Scott Brinker: We now have multiple of these AI intermediaries between us and our customer, and we have to think about how we serve them

1341
03:49:22.920 –> 03:49:24.390
Scott Brinker: As much as the humans.

1342
03:49:24.750 –> 03:49:37.840
Kerry Cunningham: Yeah, it’ll be interesting to see how long that takes, because, you know, one of the things when you… when you study buyers in B2B, when you were making, kind of, enterprise, kind of, important purchases for their businesses.

1343
03:49:37.840 –> 03:49:46.630
Kerry Cunningham: You know, two of the most important things, that buyers are looking for are, defensibility. So…

1344
03:49:46.780 –> 03:49:55.560
Kerry Cunningham: I want to make sure that if I say, yes, we want to spend this money, that it’s not going to come back on me, or not on me alone, right?

1345
03:49:55.560 –> 03:50:09.890
Kerry Cunningham: And so, maybe I would just leave it with that… with that one thing. That suggests very strongly what we’re seeing in our data, that backs that up, is that buyers are not letting go of their primary research their own when they’re in market.

1346
03:50:10.570 –> 03:50:11.700
Kerry Cunningham: When they’re not in…

1347
03:50:11.700 –> 03:50:23.469
Scott Brinker: Right. But again, it’s almost like it doesn’t have to be one or the other. It’s like, particularly the larger the purchase is, the longer the, you know, research, discovery, evaluation cycle is.

1348
03:50:23.470 –> 03:50:23.910
Kerry Cunningham: Yeah.

1349
03:50:23.910 –> 03:50:37.169
Scott Brinker: I think it just, you know, we are seeing more and more of people leveraging AI to even get, like, a deeper, perspective of things earlier on in the journey than before. How would they have… yeah, you know, anyway.

1350
03:50:37.170 –> 03:50:48.920
Kerry Cunningham: I think when buyers are not in market, you know, if I’m an accounting manager, and you know, my company buys accounting software, I use it, I’ve been using it forever.

1351
03:50:48.920 –> 03:51:01.040
Kerry Cunningham: I’m probably not going to your website anymore when I’m not in market. Now, when we are in market and we’re actively looking at things, I probably am, but if I’m in between, I’m probably going to ChatGPT or Claude or something like that, and…

1352
03:51:01.050 –> 03:51:05.289
Kerry Cunningham: Anything you’ve done with your brand and your colors and all of that good stuff you’d like me to see?

1353
03:51:06.020 –> 03:51:10.310
Kerry Cunningham: Is it gonna come through on ChatGPT? No. Yeah, maybe, maybe not.

1354
03:51:11.150 –> 03:51:19.339
Scott Brinker: Well, I mean, in all fairness, what we have not yet really seen take hold outside of, you know, the things, you know, Google is doing is,

1355
03:51:19.390 –> 03:51:20.530
Scott Brinker: Advertising.

1356
03:51:20.570 –> 03:51:41.720
Scott Brinker: In this world, and I feel pretty safe in the bet that, like, oh, that’s coming, and I don’t actually think that’s a bad thing, either for buyers or sellers. You know, I think there are going to ultimately be vehicles of, like, okay, when I am in the context of trying to get, you know, the answers to something here, yeah, the fact that maybe I’m seeing things that…

1357
03:51:41.850 –> 03:51:50.960
Scott Brinker: is a chance for the brand to, like, actually represent some particular idea, opportunity, offer, something like that. I think that’s gonna be a really cool error when we get.

1358
03:51:50.960 –> 03:51:57.710
Kerry Cunningham: Yeah, yeah, and I… so, you know, in a lot of ways, maybe it doesn’t look very different than it does today, or in the recent past when you’ve been Googling, so…

1359
03:51:57.710 –> 03:51:58.900
Scott Brinker: We’ll find out!

1360
03:51:58.900 –> 03:52:08.569
Kerry Cunningham: Yeah, so I wanted to ask you, I think, what everybody would want to hear from you is, so what do you think the Martech stack looks like, kind of big picture.

1361
03:52:08.570 –> 03:52:24.439
Kerry Cunningham: over the next few years. You’ve got, you mentioned earlier, you know, there are the big platforms, there are the decent-sized applications, and then you’ve got this term that you talk about, the hypertail, which I think is great. So maybe just explain what that is, and where it fits, and…

1362
03:52:24.440 –> 03:52:25.979
Kerry Cunningham: How should we be thinking about that?

1363
03:52:26.120 –> 03:52:38.420
Scott Brinker: Yeah, yeah, no, it’s funny, the session that was right before us, John Brunswick was, answering one of these questions, and he’s like, listen, you know, even folks who are building, like, the winning race car.

1364
03:52:38.420 –> 03:52:58.300
Scott Brinker: they’re not making their own brakes, they’re not making their own tires. I think the… what’s the phrase for this? Like, the rumor of SaaS’s demise is greatly exaggerated, you know? I think there’s still, like, right, I mean, a tremendous opportunity for there to be these, like, core platforms and core services that nobody wants to reinvent.

1365
03:52:58.300 –> 03:53:01.730
Scott Brinker: You know, we want them to be that, you know, center of gravity.

1366
03:53:01.970 –> 03:53:03.630
Scott Brinker: That being said.

1367
03:53:03.790 –> 03:53:18.759
Scott Brinker: I think we also recognize that, okay, on top of those things, we have the ability to now start to customize the interaction experiences, you know, our internal employees have with these things, the experiences we’re gonna, like, deliver to customers.

1368
03:53:18.760 –> 03:53:24.749
Scott Brinker: you know, the way in which, like, workflow gets, you know, configured and, like, adapted.

1369
03:53:24.750 –> 03:53:31.089
Scott Brinker: And I think that’s where we see a lot of AI is happening on this layer above the core MarTech stack.

1370
03:53:31.310 –> 03:53:32.150
Scott Brinker: Actually.

1371
03:53:32.190 –> 03:53:41.600
Scott Brinker: don’t expect the core MarTech stack to change a lot here over these next few years. I think we have a data layer. In some cases, people think of this as, you know, like.

1372
03:53:41.630 –> 03:53:44.570
Scott Brinker: systems of the record, like CRMs or CDPs.

1373
03:53:44.600 –> 03:53:58.050
Scott Brinker: I think at larger businesses, you start to see more and more they’re using the Cloud Data Warehouse or Lake House as kind of, you know, that, you know, more flexible repository. Then on top of that, we’ve got these systems that we use to, like, orchestrate and manage,

1374
03:53:58.050 –> 03:54:21.999
Scott Brinker: you know, workflows. So that might be a marketing, like a marketing automation or customer engagement platform. You see people using, the… I love these iPaaS providers, Integration Platform as a Service folks, like Zapier and Workato and Make and then Aiden. They keep evolving, and some of them, like, man, they are just, like, have become these really phenomenal platforms for orchestrating, you know, Agenic workflows and things like that. And so.

1375
03:54:22.000 –> 03:54:33.229
Scott Brinker: But they’re a platform, they’re a core app, and so you, like, have those systems, and then they give you the foundation upon which you do more and more custom stuff on the fly.

1376
03:54:33.250 –> 03:54:38.980
Scott Brinker: And that’s, you know, what we coined the hyper tail of… the long tail was this long tail of all these, like, you know.

1377
03:54:38.980 –> 03:54:59.690
Scott Brinker: micro-commercial apps out there. The hypertail is like, hey, I am, you know, Scott working here on a particular event project. Oh, I want to just spin up a custom workflow and app for this, you know, thing. I do it, it’s mine. When the event is over, I toss it away. Tungas recently called this, like, ephemeral software.

1378
03:54:59.690 –> 03:55:05.580
Scott Brinker: And I don’t think it takes away from SaaS. I think in many ways it’s complementary and augmentive to it.

1379
03:55:06.710 –> 03:55:23.229
Kerry Cunningham: You know, I think one of the ways that I’ve seen this over time is that, like, it’s… software in B2B, and especially in MarTech, is like a fractal phenomenon. Like, every time one pops up, you’re opening up a space where another one needs to go on either side of it, you know, it seems like.

1380
03:55:23.230 –> 03:55:35.680
Kerry Cunningham: And so they’re just this continuing… I mean, in some ways, that might explain why there’s 15,000-something, going on right now, is that, for everyone that does come up, there’s a couple of spots on either side of it now.

1381
03:55:35.680 –> 03:55:42.149
Kerry Cunningham: So, it looks like there might be a question from the audience.

1382
03:55:43.530 –> 03:55:54.910
Kerry Cunningham: Let me see… let’s see to that one. As agentic marketing stacks start automating execution? You mentioned AI, SDR tablets, etc. How do you see them changing the way companies map and understand the… oh, understand the customer journey?

1383
03:55:56.060 –> 03:55:57.680
Kerry Cunningham: First one to do it wins.

1384
03:55:57.980 –> 03:56:11.299
Scott Brinker: So I think understanding the customer journey, we’re actually starting to see that a lot more now, because what is it? It’s basically this ability to do analysis, you know, on different data sets, and to be honest.

1385
03:56:11.400 –> 03:56:13.299
Scott Brinker: For a lot of marketers.

1386
03:56:13.410 –> 03:56:30.979
Scott Brinker: beyond very simple data reporting, they often had a bottleneck of, like, okay, now I need to reach out to, you know, an analyst who’s gonna, like, dig into our, like, you know, snowflake, or our amplitude, and, you know, sort these things out. And marketers, if they have one superpower, it’s like.

1387
03:56:30.980 –> 03:56:47.529
Scott Brinker: they are just these perfect question-generating machines, you know? Like, I’ve never met a marketer who doesn’t have, like, a billion questions in their head. It’s that curiosity that drives them, but a lot of them, they’ve, like, even self-censored these questions for ages, because it’s like, okay, well, the amount of work it’s gonna get to answer that.

1388
03:56:47.530 –> 03:56:48.250
Kerry Cunningham: Can answer that, yeah.

1389
03:56:48.520 –> 03:56:59.200
Scott Brinker: You know, and this is one of the things that some of these AI tools that we’re using to be, you know, again, they’re not fully replacing, you know, expert data scientists, you know, expert analysts.

1390
03:56:59.200 –> 03:57:09.870
Scott Brinker: For a lot of, like, lower-level junior analyst-type work, yeah, these AI tools, you know, they have the ability to take these English language questions, natural language questions from the marketer.

1391
03:57:09.870 –> 03:57:21.450
Scott Brinker: translate them into, you know, queries, be able to pull back data. You know, it doesn’t have to be a black box, too. They can, like, make it very clear, okay, this is the data I pulled, these axes, this is what, is this what you wanted? Do you want to tweak it?

1392
03:57:21.450 –> 03:57:31.440
Scott Brinker: And I think those tools are very exciting because they… they’re helping marketers now be able to get better and better at analyzing these interactions that they have.

1393
03:57:31.890 –> 03:57:45.129
Scott Brinker: Now, I think that’s a… there’s a significant distance to getting to a world where, like, oh, given that data, AI agent, can you go ahead and map out a perfect end-to-end customer journey?

1394
03:57:46.000 –> 03:58:01.760
Scott Brinker: I don’t think the AI is… like, again, these things where they become longer and longer, you know, you know, and more complex, like multi-stage things, like the AI right now, it’s not very good at that. A lot of this comes down to context windows and memory and all that sort of stuff.

1395
03:58:01.770 –> 03:58:03.309
Scott Brinker: But, they are…

1396
03:58:03.310 –> 03:58:22.840
Scott Brinker: pretty good at helping us analyze to figure out that customer journey, and as we start to implement these next-generation customer journeys, we’re finding these places, as we were talking about at the beginning of this session, where you can inject pieces of, you know, generative AI to, like, make the experience even more efficient, or make it more delightful for the customer, so…

1397
03:58:23.430 –> 03:58:34.459
Kerry Cunningham: And I think we’re still gonna have to have marketing leaders, or even, you know, frontline people who realize, wait a minute, let me back up from this for just a moment, because

1398
03:58:34.860 –> 03:58:42.290
Kerry Cunningham: A lot of people today would still go look at, okay, I want to map the buyer’s journey, so from the time they become a lead until the time they… no.

1399
03:58:42.290 –> 03:58:58.289
Kerry Cunningham: Your buying… the buying journey for that organization probably started a year and a half ago. It started and stopped 3 or 4 times. How are we going to capture and get to all of the data that is going to help us understand what it actually looks like when a buyer is buying?

1400
03:58:58.340 –> 03:59:08.899
Kerry Cunningham: And when you do that, I think what you’re going to realize is you’re not going to control much, if any, of that, and then you need a system that’s really going to be a signal response system, not a…

1401
03:59:08.900 –> 03:59:19.670
Kerry Cunningham: buyer management system. It’s how do we respond effectively, capture all of the signal, and respond effectively in real time with the right message, right person, etc.

1402
03:59:20.060 –> 03:59:23.100
Kerry Cunningham: And give up the idea of controlling that.

1403
03:59:23.480 –> 03:59:35.710
Scott Brinker: You know, it’s funny, like, when we started some years back on this whole, like, digital journey thing, I think one of the things people kept saying at the time was, like, oh, wow, the customer is in control, the customer is driving these things.

1404
03:59:35.710 –> 03:59:47.009
Scott Brinker: But it’s almost because we got so good at instrumenting all the data of all these journey things, we got so good in these automations, I think actually a lot of us, like, lost track of them. We’re almost like, no, no, no.

1405
03:59:47.170 –> 03:59:51.160
Scott Brinker: we control this journey. We’ve got it fully instrumented, and it’s all that.

1406
03:59:51.160 –> 04:00:11.219
Scott Brinker: And that was the illusion. We still didn’t, you know, and so I think, yeah, this sort of realization that, you know, with these AI agent capabilities on the buyer side said, like, no, really? Yeah, they’re the ones who are going to be driving this and in charge, and to your point, what we want to do is we want to make sure whatever they need to, like, you know, get the right help, get the right decision, get the right service.

1407
04:00:11.390 –> 04:00:12.809
Scott Brinker: We’re there to give it to them.

1408
04:00:13.020 –> 04:00:21.990
Kerry Cunningham: Yeah, perfect. So I know we’re… we’re a minute over, Julia. So, Scott, thank you so much. This was awesome, great to hear from you, here. I think this perfect subject.

1409
04:00:22.510 –> 04:00:27.649
Julia Nimchinski: What a phenomenal session. Thank you so much, Carrie. Thank you so much, Scott. What’s the best way to support you?

1410
04:00:29.980 –> 04:00:41.090
Scott Brinker: Like, I don’t know. I sent over a link for that report. Grab a copy of that report, let us know what you think of it.

1411
04:00:41.090 –> 04:00:45.370
Kerry Cunningham: And, Scott, you’ve got a survey that you’re collecting on now, too, I think everybody should…

1412
04:00:45.370 –> 04:00:53.859
Scott Brinker: You know, thank you so much for raising the very thing that I forgot, and did I have the link, handy? Oh, man, I feel like such a marketing…

1413
04:00:54.110 –> 04:00:56.260
Scott Brinker: Noob here,

1414
04:00:57.130 –> 04:01:00.130
Scott Brinker: I’ll post it in the chat to you after this, I won’t hold up the next thing.

1415
04:01:00.530 –> 04:01:01.080
Kerry Cunningham: Right.

1416
04:01:01.500 –> 04:01:08.409
Kerry Cunningham: You can find me on LinkedIn. Hard to miss me if you look at B2B marketing topics, so… thanks, everybody.