Demo • N.Rich — Your Agentic ABM Orchestration Team

415
01:15:24.680 –> 01:15:29.530
Julia Nimchinski: And we are going to talk about the Gentec ABM orchestration. Welcome, Marcus!

416
01:15:30.130 –> 01:15:32.010
Julia Nimchinski: If you’re actually curious, Julius?

417
01:15:32.010 –> 01:15:34.620
Markus Stahlberg: All good, all good, amazing content so far.

418
01:15:35.260 –> 01:15:37.489
Julia Nimchinski: Yeah, we are excited to see the demo!

419
01:15:40.400 –> 01:15:49.659
Markus Stahlberg: Right, so let’s get started. So, so my name is Marcus Stahlberg, I’m the CEO and co-founder of Enrich.

420
01:15:49.740 –> 01:16:04.200
Markus Stahlberg: And, today I’m going to talk about, talk about the, HNTIC… HNTIC approach to, ABM, essentially, like, how to,

421
01:16:04.280 –> 01:16:14.369
Markus Stahlberg: how to automate, how to, enrich, how to, like, make, make your ABM more efficient. So let’s…

422
01:16:14.710 –> 01:16:22.499
Markus Stahlberg: Let me share… We’ll make this… Can you see my screens?

423
01:16:33.150 –> 01:16:34.050
Markus Stahlberg: Right.

424
01:16:35.840 –> 01:16:47.950
Markus Stahlberg: So, let’s get started. So, here, in the next 15 minutes, I’ll go through, what is Enbridge, what is the state of AI and agents in GTM,

425
01:16:47.950 –> 01:17:01.250
Markus Stahlberg: I’ll do two parts of the demo, integrating AI into ABM workflows, then do a bit of a recap of, like, the results that I’m showing you, or, like, what kind of results this approach produces.

426
01:17:01.250 –> 01:17:12.339
Markus Stahlberg: then, showing the, introduction to the agentic approach, within ABM, and then, like, finally, finally, finishing with the demo.

427
01:17:12.590 –> 01:17:14.269
Markus Stahlberg: The second part of the demo.

428
01:17:14.450 –> 01:17:38.959
Markus Stahlberg: So, getting started, just briefly about Enrich, like, I’m sure everybody’s not aware of, like, the category or the company. So, we are, 10 years in the business, we are recognized in Gartner Magic Quadrant 2 years in a row now, and, we’ll be there this year again. And, like, we get, we have gotten the highest ratings on, like, all categories, where customers’ feedback actually matters on

429
01:17:38.960 –> 01:17:40.309
Markus Stahlberg: On G2.

430
01:17:41.250 –> 01:18:06.049
Markus Stahlberg: And, just to, emphasize this, like, we have worked a lot on, like, capabilities and also kind of support functions that make it easier for customers to use and run ABM, ease of use, meeting requirements, high-quality support, NPS, ease of doing business with, ease of admin, ease of setup, and overall, direction. In all of those, like, we are, we are basically beating our

431
01:18:06.050 –> 01:18:14.570
Markus Stahlberg: our competition. So, very proud of that, and this is a good leeway then to start talking about the AI.

432
01:18:14.820 –> 01:18:27.659
Markus Stahlberg: So in go-to-market, like, the starting point, unfortunately very often is that, like, AI, and the agents are, like, eating, soup with a fork.

433
01:18:27.990 –> 01:18:36.050
Markus Stahlberg: It works, and it delivers, but, like, is it the most efficient way to deliver those results that can be questionable?

434
01:18:36.490 –> 01:18:38.260
Markus Stahlberg: So,

435
01:18:38.770 –> 01:18:58.539
Markus Stahlberg: agents do… agents do deliver, like, many companies get really good results with the, with AI SDR, like, with, with automating… automating the approach, but the question really is, like, how does it work? And this is always something that is, is good to think about. So, there are two

436
01:18:58.540 –> 01:19:16.200
Markus Stahlberg: components, there’s conversion, and still, based on, like, our data, humans can still deliver better conversion than AI, like, on average, if you think about, like, individual outreach. But what AI can do is that it can do volume.

437
01:19:16.390 –> 01:19:34.160
Markus Stahlberg: And this is how the results actually, like, are generated. Humans can’t do volume, AI can. But this is not really the right way or the long-term approach that we would recommend to utilize AI just for the volume, but actually you need to come up with ways

438
01:19:34.220 –> 01:19:38.160
Markus Stahlberg: Ways how to make it work, also in terms of conversion.

439
01:19:38.410 –> 01:19:39.820
Markus Stahlberg: And,

440
01:19:40.240 –> 01:19:49.599
Markus Stahlberg: MIT had this, very, influential report about state of AI in business, and, like, done this, recently, this year.

441
01:19:49.600 –> 01:20:02.769
Markus Stahlberg: And the lessons that matter is that AI needs to learn, it needs to integrate, and it needs to target bottlenecks. And, like, so it is more about the quality than it is about the quantity.

442
01:20:02.770 –> 01:20:20.900
Markus Stahlberg: And here, this is the approach that we have taken to AI, so we’ve been really kind of working hard on understanding, like, how, how people, how marketers generate, create these, like, ABM campaigns, and where are those bottlenecks, how to integrate AI into the workflows.

443
01:20:21.060 –> 01:20:44.889
Markus Stahlberg: Now, as a crash course into, into ABM, so, you start ABM by defining your target accounts, your ideal customer profile, and once you have that, you want to segment them, these accounts, into different stages based on, the signals that are available. So, basically, from awareness to, to consideration and, like, all the way to, to purchase.

444
01:20:44.900 –> 01:21:05.209
Markus Stahlberg: And this is then what marketing needs to do, and this is really what we will be talking about today, about how to move accounts from cold to hot, and how AI is going to help with this process, and how is it making this process smoother and, like, less of friction.

445
01:21:05.210 –> 01:21:06.540
Markus Stahlberg: With less friction.

446
01:21:06.710 –> 01:21:21.659
Markus Stahlberg: So, then, going into campaigns, so we have, here we have awareness conspiration Decision campaigns, and I will just show you, like, just to get the context, like, how, how these campaigns, and ads work, so you get an idea.

447
01:21:21.840 –> 01:21:38.160
Markus Stahlberg: about it. So, we have designed the ad formats. Here is an example of an ad format that, is used within the platform. We have designed it in a way that it’s optimally AI-friendly. So, it’s basically image plus

448
01:21:38.220 –> 01:21:51.890
Markus Stahlberg: text. So there are headlines, descriptions, and images. And this ad, this is one ad, basically, that has 350 different variations, so you can just, like, basically, the system automatically, in the background.

449
01:21:52.140 –> 01:22:05.660
Markus Stahlberg: rotates the different ads and tries to find, like, what is the ideal combination for each of the… each of the individuals, like, each of the people who are, who are basically seeing the ads. And here, this is kind of the first

450
01:22:05.660 –> 01:22:21.079
Markus Stahlberg: let’s say, generation of our AI, like, so I just cleared those, those, headlines and descriptions. So basically, AI can, within the platform, can, generate content that is,

451
01:22:21.140 –> 01:22:34.179
Markus Stahlberg: on brand that is, like, related to the kind of, value propositions, kind of messaging that, that you want. So this is trained in the, in the background, and this works really well.

452
01:22:34.430 –> 01:22:42.590
Markus Stahlberg: So AI is super good in this kind of task where you have, like, limited space, and you need to be, like, communicating

453
01:22:42.740 –> 01:22:52.490
Markus Stahlberg: pretty complicated things. So, our AI is called Aurora, and so this way now, we have the different, different variations here.

454
01:22:52.700 –> 01:23:06.830
Markus Stahlberg: And, like, it would be basically ready to go, like, and, and, and run then. There’s another ad format which also is, has been designed to be AI-friendly, so this is…

455
01:23:07.520 –> 01:23:20.870
Markus Stahlberg: an article format, so it is basically, again, text and images. And when you go to media, like, you can see the ad, like, it’s a scrollable article, where the system is tracking the scrolling.

456
01:23:20.960 –> 01:23:27.980
Markus Stahlberg: And, here you can use AI to generate the images and generate, generate the content itself.

457
01:23:28.210 –> 01:23:40.129
Markus Stahlberg: Which makes this very adaptable and personalized then, like, to the, to the broader audience. So, in our back end, we have a demand site platform, which means that we can target people

458
01:23:40.380 –> 01:23:53.050
Markus Stahlberg: within millions of websites. So, when you target certain accounts, different people will be seeing different ads, different variations, and the system is just automatically, like, optimizing that.

459
01:23:53.050 –> 01:24:04.490
Markus Stahlberg: And all of this is basically AI-generated, but it’s still, at the moment, in a way that you need to generate it, using, like,

460
01:24:04.490 –> 01:24:11.890
Markus Stahlberg: kind of manually, one ad at a time. So now, going into the… into the results.

461
01:24:11.890 –> 01:24:13.360
Markus Stahlberg: So, this…

462
01:24:13.360 –> 01:24:34.339
Markus Stahlberg: particular, like, approach has worked out really well. So these are click-through rates of, of human versus AI-made cross-channel ads within our platform. So you can see here, these are individual variants that the purple ones here are the AI ones, and the red ones are the, human-made ones. And

463
01:24:34.340 –> 01:24:39.459
Markus Stahlberg: I was telling before that it’s not about the volume, it is… it is about the…

464
01:24:39.720 –> 01:24:51.309
Markus Stahlberg: Conversion. And how does it work here, with AI, the… and with ads especially, it is about, it is about, like, having a lot of different variations.

465
01:24:51.310 –> 01:25:10.889
Markus Stahlberg: And AI is super good at generating those variations. So, what ends up happening is that people are not creating enough of the variations, and it’s also super hard to come up with a lot of variations that are, like, about the same topic, but, like, just worded in a bit different way. So that’s… that’s why this works.

466
01:25:11.330 –> 01:25:26.320
Markus Stahlberg: Then, let’s move forward, like, to the, to the next topic, which is about why are agents needed. So, already I showed you that this is a bit, bit manual, manual process, and, if I look at now,

467
01:25:26.770 –> 01:25:35.659
Markus Stahlberg: an average… team, working with ABM. You need about 4 full-time employees to, to do that.

468
01:25:35.830 –> 01:25:49.890
Markus Stahlberg: to create those, sustain, like, a solid ABM approach, you need a… you need to do a lot of things, like there is campaign planning, ICB definition, data wrangling, content personalization, and so on and so forth. So this is a lot of effort, and normally.

469
01:25:50.020 –> 01:26:00.320
Markus Stahlberg: companies don’t have four people to do that. Normally, there’s one… one person, two people who are working on it. So it’s super hard, and this is where, like, help is definitely needed.

470
01:26:00.560 –> 01:26:05.120
Markus Stahlberg: So our approach is essentially that, like, instead of saying that, okay.

471
01:26:05.780 –> 01:26:20.709
Markus Stahlberg: just optimize this to your resources and do less. We say that… actually do more, and, like, replace certain functions with, with agents. So we have campaign planning agents, content design, campaign optimization, campaign analyst, agents,

472
01:26:20.710 –> 01:26:28.680
Markus Stahlberg: That, basically, are managed by the Demand Generation Manager, or, or, like, whoever, like, manages the ABM, ABM campaigns.

473
01:26:29.660 –> 01:26:53.679
Markus Stahlberg: And the way this works is that it’s a workflow where we start from, like, the agent starts from sensing, interacting with the user, doing reasoning, planning, blueprint of the campaign, collaborating with the user, then acting, interacting with other agents, and then eventually delivering, like, the learnings back to the agent.

474
01:26:53.680 –> 01:26:56.989
Markus Stahlberg: Like, back to this, this system.

475
01:26:57.670 –> 01:27:03.410
Markus Stahlberg: And now we can go to the second part of the demo.

476
01:27:03.910 –> 01:27:11.130
Markus Stahlberg: So we start here, actually, just a second, I’ll just refresh this. Sorry about that.

477
01:27:12.220 –> 01:27:13.290
Markus Stahlberg: So…

478
01:27:13.650 –> 01:27:28.870
Markus Stahlberg: Here, so we go to the agent’s home, and we have the campaign creator agents and content creator agents that I will demo here. So there are several other agents that, that are, like, under process.

479
01:27:28.870 –> 01:27:37.269
Markus Stahlberg: But I’ll start with the campaign creator agents. So, the first thing here is, like, let’s start by creating or planning the campaign.

480
01:27:37.610 –> 01:27:48.099
Markus Stahlberg: So I will now just copy-paste it here a bit of briefing for the agent, and then I will just add our

481
01:27:48.510 –> 01:27:50.430
Markus Stahlberg: website here.

482
01:27:50.590 –> 01:27:54.600
Markus Stahlberg: as the source, and then clicking confirm.

483
01:27:55.240 –> 01:28:02.330
Markus Stahlberg: defining how long the campaign, these kind of things still are. You still need human inputs, for this.

484
01:28:03.170 –> 01:28:12.719
Markus Stahlberg: And then, in the background, the system puts all the, data together and utilizes your, your, like, kind of the trained.

485
01:28:13.080 –> 01:28:33.290
Markus Stahlberg: pre-training that is available, and the information that I have given, and then provides suggestions of the messaging pillars, essentially. So this starts from very, very high level. So, from cold to hot accounts, turning anonymous engagement into sales-ready opportunities, and then another one here, proven impact on pipeline. So.

486
01:28:33.460 –> 01:28:35.350
Markus Stahlberg: Let’s move forward.

487
01:28:35.730 –> 01:28:37.520
Markus Stahlberg: here, so…

488
01:28:38.350 –> 01:28:48.790
Markus Stahlberg: Now, the next step is to, create the campaign, and this was a progression campaign, so basically creating these different campaigns for the different stages. So…

489
01:28:48.930 –> 01:28:55.379
Markus Stahlberg: We have cold, in-market, engaged, and hot. So, four, four campaigns, so we create this.

490
01:28:55.900 –> 01:28:57.170
Markus Stahlberg: these campaigns.

491
01:28:57.730 –> 01:29:02.699
Markus Stahlberg: Based on the, account specifications that I, I defined.

492
01:29:03.510 –> 01:29:08.499
Markus Stahlberg: And then here, from here, you go ahead to, to creating the content.

493
01:29:09.290 –> 01:29:26.329
Markus Stahlberg: And this, again, works with a very similar approach, like, I showed in this manual, like, AI approach, but this is just automated. So now, what happens is that instead of using the

494
01:29:26.850 –> 01:29:31.199
Markus Stahlberg: kind of individual, one ad-at-a-time approach, this basic

495
01:29:31.200 –> 01:29:51.659
Markus Stahlberg: creates the entire matrix of, of ads. So typically, you would have, you have these different stages, and then you have, for instance, different buyer personas for which you want to create the ads. And this is all, again, working in this, like, text plus images, approach. So you can create

496
01:29:52.540 –> 01:29:56.230
Markus Stahlberg: different image, image IDS,

497
01:29:56.550 –> 01:30:07.189
Markus Stahlberg: And this is giving, basically, like, just kind of the types of images that, that you would, you would like to use here. So let’s take those, and then move forward.

498
01:30:07.250 –> 01:30:19.009
Markus Stahlberg: And, and this way, you get a lot of, lot of different, different, like, again, images, because the entire purpose of this is to, is to actually, like, get.

499
01:30:19.610 –> 01:30:31.149
Markus Stahlberg: get a lot of variation. That’s… that’s what works, because it’s so many different people and so much volume in the background of advertising still. Okay, so here we go. So now,

500
01:30:31.310 –> 01:30:41.300
Markus Stahlberg: The, next one is, like, what kind of, content we generate, and this way, we end up

501
01:30:41.590 –> 01:30:51.149
Markus Stahlberg: actually generating the, the ads and going all, all the way to the, to the, kind of, end of the process. So that we have, in the end, we have,

502
01:30:51.190 –> 01:31:05.280
Markus Stahlberg: four campaigns, or five campaigns that have been created, like, very quickly and very efficiently. So this way, you don’t need… you don’t need a lot of… a lot of resources to do that, and, that’s, that’s what we have seen works really well.

503
01:31:05.500 –> 01:31:06.730
Markus Stahlberg: So, I think…

504
01:31:07.320 –> 01:31:15.310
Markus Stahlberg: That’s it from, my part, so I’d be happy to answer if there are any, any questions.

505
01:31:15.310 –> 01:31:28.749
Julia Nimchinski: Thank you so much, Marcus. Amazing session. We are, unfortunately, though, have to transition to our next session, our CXO Roundtable, Leveraging Agents to Align Your GTM Teams with Your Buyer. But before we do that.

506
01:31:28.910 –> 01:31:31.300
Julia Nimchinski: Where should our community go?

507
01:31:31.490 –> 01:31:35.559
Julia Nimchinski: What’s the best next step to just learn more about Enrich?

508
01:31:36.150 –> 01:31:42.109
Markus Stahlberg: You can find me on LinkedIn and our website, enrich.io.

509
01:31:42.940 –> 01:31:43.990
Julia Nimchinski: Amazing.