Fireside Chat with Amos Bar-Joseph & Meagen Eisenberg — The Rise of the System of Action

465
01:00:16.690 –> 01:00:24.169
Julia Nimchinski: We’ll definitely promote it. And next up, we want to welcome Amos Barr Joseph, CEO and co-founder of Swan AI.

466
01:00:24.390 –> 01:00:32.270
Julia Nimchinski: And Megan Eisenberg, CMO and board member at Sensera, Navin, MongoDB, and G2!

467
01:00:32.650 –> 01:00:34.690
Julia Nimchinski: Welcome to the show, how are you doing?

468
01:00:35.020 –> 01:00:58.260
Meagen Eisenberg: Doing well, thank you for having us. Amos, I think we’re in good shape. The last thing they talked about was, just because you give humans tools doesn’t mean they’ll be necessarily more productive or deliver earlier. They actually need deadlines, but if you give it to an agent, they will. So I think that is right in line with what we’re going to talk about, really a system of action.

469
01:00:58.260 –> 01:01:04.760
Meagen Eisenberg: So I’m excited to, be talking with you today, and have you on, and, and…

470
01:01:04.760 –> 01:01:11.099
Meagen Eisenberg: Really kick off with, you know, you described a new paradigm, the system of action. You know, how do you define it?

471
01:01:11.920 –> 01:01:30.300
Amos Bar Joseph: Amazing. So, we’re starting with the hard questions right off the bat, I like it. We didn’t come here to play games. So, a system of action is basically the intelligence layer that stands between your data and systems.

472
01:01:30.300 –> 01:01:53.349
Amos Bar Joseph: and your execution, your ideas, basically. So it’s where ideas become automated workflows with no technical complexity involved. And we want to break it down here for a second, Megan, before this might sound abstract, and, you know, a lot of people are throwing a lot of big words in the air in the last, you know, few months, let’s call it like that. And so.

473
01:01:53.700 –> 01:02:03.890
Amos Bar Joseph: if you look at this new software paradigm shift that our, you know, AI agents are bringing to the table, one of the things that will radically change is how we’re interacting with our systems.

474
01:02:04.070 –> 01:02:13.940
Amos Bar Joseph: Right? And, AI agents, what they can be really good at is becoming that intermediate between us, what we want, and the systems.

475
01:02:13.960 –> 01:02:34.829
Amos Bar Joseph: And then, you can imagine two different futures here that are already unraveling. One is that every application will develop their own concierge, let’s call it, like we’re trying… maybe seeing a bit here. So, instead of going to Swan, I will talk with Swan agent, and then instead of going to Salesforce, I will go with… I will talk with Salesforce agent.

476
01:02:34.830 –> 01:02:41.470
Amos Bar Joseph: But, another future that we might imagine is that there is one system of action

477
01:02:41.470 –> 01:02:46.549
Amos Bar Joseph: One agent to rule them all, that all of the systems will connect to that

478
01:02:46.670 –> 01:03:03.690
Amos Bar Joseph: agent, and it will be able to orchestrate ideas and turn them into workflows across your systems, and will use your data to act upon it. And that is the future that we’re betting on, actually. We’re already seeing unraveling, basically.

479
01:03:04.820 –> 01:03:10.259
Meagen Eisenberg: Sounds good. So, why do you think SaaS as we know it is being replaced by the agentic layer?

480
01:03:10.700 –> 01:03:14.489
Amos Bar Joseph: Yeah, so, that’s a… that’s a great question. So…

481
01:03:14.630 –> 01:03:19.440
Amos Bar Joseph: If you look at the, you know, the… cloud rev…

482
01:03:20.020 –> 01:03:35.809
Amos Bar Joseph: evolution, in that sense. So, what we’ve witnessed is that instead of going to these apps that I have on-prem, I can now go to the browser, and from the browser, I can access all of these apps, right? And…

483
01:03:35.810 –> 01:03:46.569
Amos Bar Joseph: The new software paradigm shift, it’s also a platform shift. Again, every software paradigm shift brings a new platform shift. We’ve seen it in mobile as well, by the way.

484
01:03:46.570 –> 01:04:02.770
Amos Bar Joseph: And mobile brought, as well, a new platform shift. And what we’re seeing here is that this new browser is the system of action, basically. And we’re already starting, if you’ve been, like, a savvy user of ChatGPT or Gemini, you notice that you can connect your email.

485
01:04:02.770 –> 01:04:19.559
Amos Bar Joseph: for, all of a sudden, to ChatGPT. And you might be able to connect your Drive and Notion, and then HubSpot just announced that you can connect it to Claude and ChatGPT, and all of a sudden, these just consumer apps are becoming

486
01:04:19.630 –> 01:04:35.319
Amos Bar Joseph: this browser to the world of AI and to the world of your systems and data, and that’s where, basically, the fight is going towards. And if you look at Salesforce, for example, with AgentForce and HubSpot, with all that agents.

487
01:04:35.430 –> 01:04:55.099
Amos Bar Joseph: What they’re realizing, this is not an offense move, this is a defensive move. They’re understanding that the strategic control of the stack moves from the system of record to the system of action. And the one that will control that layer, basically, will control the entire stack around it.

488
01:04:56.660 –> 01:05:20.250
Meagen Eisenberg: Yeah, that was one of the questions I was going to ask, what sort of parallels or differences in these past paradigm shifts with SaaS, cloud, mobile. I think you talked to that a little bit. Also today, I saw an announcement where there’s now a MCP registry, so I think that’ll be interesting as we start to develop into this new paradigm and world. So looking at what does that mean.

489
01:05:20.250 –> 01:05:39.299
Meagen Eisenberg: as a company that’s built our own, and the marketing org that’s firing off a lot of different things and having agents run. So I’ll be curious on that. But let’s talk a little bit about the vision of the system of action. Walk us through Swan’s vision. What does a go-to-market team’s day look like once they have a system of action?

490
01:05:39.530 –> 01:05:41.860
Amos Bar Joseph: Yeah, definitely. So,

491
01:05:42.360 –> 01:05:48.440
Amos Bar Joseph: Swan, the… our customers, they actually… they look at it and call it lovable for GTM.

492
01:05:48.770 –> 01:06:04.590
Amos Bar Joseph: Okay. It’s an AI go-to-market engineer. It’s kind of like a developer, a technical resource, that works together with sales and marketing and RevOps to turn any go-to-market idea into an agentic workflow in seconds.

493
01:06:04.590 –> 01:06:18.359
Amos Bar Joseph: You know, from prompt to pipeline, so you could really scale demand and revenue with intelligence, not head count, and iterate on your go-to-market at the speed of thought. And what we were…

494
01:06:18.410 –> 01:06:22.350
Amos Bar Joseph: accustomed to is that, you know, we have an idea.

495
01:06:22.730 –> 01:06:42.129
Amos Bar Joseph: And we want to say, you know what, maybe our go-to-market should look a little bit different, we want to change our ICP, we want to change a specific buyer journey, we want to change our messaging in some way, or we want to change how, you know, inbound leads are captured and processed and turned into meetings, or whatever. And we have, like, this small idea, and the distance

496
01:06:42.330 –> 01:06:52.530
Amos Bar Joseph: Between that idea and its execution is so long, it takes so many hours of implementation, but what if that distance became zero?

497
01:06:52.680 –> 01:06:58.359
Amos Bar Joseph: became, like… a second. I could just tell Swan, Swan, this is what I want.

498
01:06:58.670 –> 01:07:05.819
Amos Bar Joseph: to do, basically. This is the new idea, and Swan will be able to translate it into an agentic workflow in seconds.

499
01:07:05.900 –> 01:07:18.810
Amos Bar Joseph: And so, it’s not like Swan and the system of action is about replacing sellers. It’s about removing all the technical complexity that stands between go-to-market teams

500
01:07:18.810 –> 01:07:26.380
Amos Bar Joseph: And their ideas, and their ability to differentiate and to iterate on what makes them more efficient and better than their competitors.

501
01:07:27.660 –> 01:07:51.960
Meagen Eisenberg: Yeah, I guess that’s what you mean by one agent will interface for all workflows. You go to one browser, one place, it’s all… instead of having to log into several different systems, connect them, you know, and work maybe with different teams that own different parts of what you need to put out there, certainly as a CMO, with multiple teams building content, and then bringing it all together, and then trying to create that experience around the buyer.

502
01:07:51.960 –> 01:07:56.219
Meagen Eisenberg: How do you orchestrate that from one place? And it sounds like this.

503
01:07:56.220 –> 01:07:59.120
Meagen Eisenberg: This could be the answer to that. I’m curious…

504
01:07:59.120 –> 01:08:16.890
Amos Bar Joseph: Yeah, Megan, I want to double down on that, and also refer to something you mentioned earlier, because you talked about, you know, the MCP registry. So, for those of you who don’t know what MCP is, so it’s basically a simple way for, you know, the developer community to agree on how to connect systems

505
01:08:16.890 –> 01:08:19.419
Amos Bar Joseph: To agents, like, in a very simple way.

506
01:08:19.460 –> 01:08:34.379
Amos Bar Joseph: And, it’s… if we all follow the MCP, so it would be very easy for a developer to build an AI agent that could talk with HubSpot, but then also with Salesforce, but then also with Slack, but then also with whatever, and etc.

507
01:08:34.729 –> 01:08:42.019
Amos Bar Joseph: That signal shows us that the world wants to connect these systems to agents, right?

508
01:08:42.050 –> 01:08:51.599
Amos Bar Joseph: And so, the buyer always prefers that one source of truth, that one place to control everything from.

509
01:08:51.600 –> 01:09:10.960
Amos Bar Joseph: And that was a promise that, you know, buyers have been sold to for many, many, many years, that, you know, the one system that will do everything, especially in go-to-market. And, buyers got, you know, tired of that promise, but they really want it. And for the first time in software history, we got the right technology.

510
01:09:10.960 –> 01:09:11.990
Amos Bar Joseph: To power that.

511
01:09:12.819 –> 01:09:21.449
Meagen Eisenberg: Yeah, it makes sense, because my other question would be, okay, I’ve got a CRM, I have Marketo for marketing automation, I’ve got data warehouse.

512
01:09:21.449 –> 01:09:33.999
Meagen Eisenberg: You know, these all seem to be the record or the one source of truth, although they, they bi-directionally sync information, and are, are, you know, they have different people logging in and using them. So, I guess…

513
01:09:33.999 –> 01:09:43.339
Meagen Eisenberg: it doesn’t really matter, because as long as you know where the data resides, and you have a way to connect into it with the NCP layer, the agents can go find it.

514
01:09:44.160 –> 01:09:46.389
Amos Bar Joseph: Yeah, that’s true, and

515
01:09:46.800 –> 01:10:05.749
Amos Bar Joseph: the mindset that changes here is that, in a world where you have that system of action, that intermediate, basically, that helps you either build out these workflows or initiate these one-offs, like, you know, find me that, and perform that action, etc. You have that intermediate.

516
01:10:05.910 –> 01:10:10.819
Amos Bar Joseph: What happens is that all your systems become service providers to that agent.

517
01:10:11.020 –> 01:10:11.760
Amos Bar Joseph: Basically.

518
01:10:11.910 –> 01:10:36.850
Amos Bar Joseph: So they stop serving users, and they start serving agents, in that sense. And we’re already seeing that in Swan. We see that the best teams, what they’re trying to do is they look at, you know, their data, and they look at their systems, and they try to understand how, you know, agents could use them, basically, in a way, right? And we’re at that very inconvenient stage where

519
01:10:36.850 –> 01:10:38.729
Amos Bar Joseph: We’re trying to change

520
01:10:38.730 –> 01:10:45.700
Amos Bar Joseph: that stack. That is what’s happening right now. We’re trying to change what we already have so that agents

521
01:10:45.700 –> 01:10:59.310
Amos Bar Joseph: could use them. And so what happens to Marketo and Salesforce, etc, they just become the back end of that intermediate system. They are just service provider, and so they should focus on what they know to do best.

522
01:10:59.310 –> 01:11:16.280
Amos Bar Joseph: It’s not like the system of action will perform everything on its own, it will just be intermediate. So if you’re really good at A-B testing, you know, performance marketing, for example, you still… you will still do that, but the one that will control that will be an agent on behalf of the user.

523
01:11:17.580 –> 01:11:19.209
Meagen Eisenberg: It makes a lot of sense.

524
01:11:19.210 –> 01:11:39.100
Meagen Eisenberg: So you describe Swan as lovable for go-to-market, and I have to admit, I love, I love that tool. I’ve, played with it to build some fun apps, so I’m curious, if I’m an AI go-to-market engineer that can turn these ideas into workflow, can you share just a concrete example of a customer turning a raw idea into a pipeline?

525
01:11:39.420 –> 01:11:56.149
Amos Bar Joseph: Yeah, so, let’s talk about what I believe is one of the most impactful motions in go-to-market right now, but also the one that requires the most engineering to perform, actually, and that is first-party intent, basically. So.

526
01:11:56.160 –> 01:12:19.780
Amos Bar Joseph: If you look at, you know, generating pipeline, right? Now, it’s not… you know, I’ve never talked to a CMO or CRO that says, yeah, we have enough pipeline, we don’t need more, we’re totally okay with that, no need to help. And if you look at what generates pipeline in the most efficient way is actually, you know, acting upon first-party intent. People that are engaging with your assets.

527
01:12:19.980 –> 01:12:35.300
Amos Bar Joseph: And you know exactly who they are, and they’re showing interest in what you’re selling, or the content that you’re putting out there, and you want to turn that engagement into qualified pipeline, basically.

528
01:12:35.890 –> 01:12:43.119
Amos Bar Joseph: But it’s extremely complex, because what happens if it’s already, you know, closed lost?

529
01:12:43.330 –> 01:12:59.910
Amos Bar Joseph: So there’s a lot of context there that we want to understand what to do with it. What happens if it’s in the pipeline, even? Hmm, that’s super interesting. The AEs could actually leverage that and understand, okay, they are coming to our website. What does that mean to the exact deal stage that we’re at? But what happens if it’s net new?

530
01:12:59.910 –> 01:13:11.079
Amos Bar Joseph: Is it a relevant company? Is it within our ICP? It’s not within our ICP. Is the level of intent enough for us to double down here with an SDR, or should we just put it in a nurturing sequence?

531
01:13:11.180 –> 01:13:17.379
Amos Bar Joseph: So, just for a small motion here, that is first-party intent, the engineering efforts

532
01:13:17.380 –> 01:13:24.849
Amos Bar Joseph: To start experimenting on understanding what is the right buyer journey for every scenario here is so vast

533
01:13:24.850 –> 01:13:49.829
Amos Bar Joseph: That 99% of go-to-market organizations just result to spray and pray, okay, someone visits our website, let’s put them in a 20-step sequence and hope that they will convert, and they’re saying, wow, it’s working much better than cold outbound, okay, but what we’re missing out here is a huge potential of much more efficient conversion that all across the journey and up until the pipeline late stages.

534
01:13:49.830 –> 01:13:57.580
Amos Bar Joseph: Basically. And so Swan can turn all of these technical… all of that technical complexity into something that you can just

535
01:13:57.580 –> 01:14:16.170
Amos Bar Joseph: iterate on it in natural language, and speak to it, and just say, yeah, you know what, let’s try like that, but let’s try like that. Yeah, we’re seeing great results, but I’m actually… I want you to notify the reps, don’t send an email, just update the hub, just update the CRM, etc. Whatever you want it to basically work, you can just tell someone to do it, and it will build it for you in seconds.

536
01:14:17.070 –> 01:14:24.319
Meagen Eisenberg: And how does that work? How does one connect to our existing systems of record, or third-party data, or even other agents?

537
01:14:24.320 –> 01:14:26.699
Amos Bar Joseph: Yeah, so,

538
01:14:27.130 –> 01:14:37.179
Amos Bar Joseph: you mentioned MCP at the beginning, I’m going back to it. That was a great data point to bring up, Megan. So what we did, actually, we built kind of, like, our own version of that.

539
01:14:37.220 –> 01:14:41.869
Amos Bar Joseph: We built a, you know, a connector that Swan

540
01:14:41.870 –> 01:14:59.000
Amos Bar Joseph: can either connect to a system, and so that we have, like, a system connector. That system could be either, you know, Salesforce, or HubSpot, or Outreach, or SalesLoft, or whatever, basically, you have in yours, or Marketo, in that sense. And so it can, you know, integrate into a system.

541
01:14:59.100 –> 01:15:13.520
Amos Bar Joseph: read and write into that system based on, you know, the data that… first-party data that you have there, and the actions that you perform. Then, you have another connector, which is the data connector. So, we actually offer third-party data

542
01:15:13.520 –> 01:15:24.789
Amos Bar Joseph: connections out of the back with Swan, so if you want to find contacts, enrich contacts, enrich accounts, use intent signals like first-party intent, you can use

543
01:15:24.850 –> 01:15:33.990
Amos Bar Joseph: from Swan directly, and Swan can integrate with these data providers, like third-party data providers. Off the bat, you can either work with yours, or with ours.

544
01:15:34.240 –> 01:15:53.059
Amos Bar Joseph: Then, finally, you have that agent connector. So, for example, right now, if you want Swan to research online an account that just submitted a demo request, so Swan will use Perplexity, which is an AI search engine to do it. It’s an AI agent that goes online, so we didn’t have to invent

545
01:15:53.060 –> 01:16:04.900
Amos Bar Joseph: you know, researching online on our own. We just connected Swan to the Perplexity agent, and it can now just tell Perplexity Agent, hey, Megan wants to research that lead, just do it for her. And it does.

546
01:16:06.270 –> 01:16:25.880
Meagen Eisenberg: Yeah, you know, it’s interesting. My team’s been doing similar things, where they’re doing the deep research using perplexity to create account. Like, I’m preparing a sales rep to go into a sales meeting and giving them all the background on the account and being able to pull that in. It’s pretty cool how far we’ve come on that. Something that would take a few hours is

547
01:16:25.880 –> 01:16:40.440
Meagen Eisenberg: minutes, so it’s great to see that you’re sort of orchestrating that across many different systems and use cases. So where are your customers finding the biggest ROI? Is it speed, scale, cost reduction, creativity?

548
01:16:41.080 –> 01:16:43.769
Amos Bar Joseph: Yeah. So,

549
01:16:44.130 –> 01:16:57.919
Amos Bar Joseph: all of the above, I would say, but I’ll explain. The currents… the main pillar here is speed of iteration, we call it, okay? So, we believe that right now.

550
01:16:58.110 –> 01:17:16.870
Amos Bar Joseph: the biggest barrier in go-to-market is not better data, and it’s not better AI, it’s just to remove all the technical complexity that stands between teams and their tech stack. That is the biggest barrier. The moment that you remove that barrier, what happens is that your speed of iteration

551
01:17:16.870 –> 01:17:27.080
Amos Bar Joseph: And your ability to execute on what you think is the best thing to execute on just increases by 2 or 3 scales, basically.

552
01:17:27.150 –> 01:17:52.029
Amos Bar Joseph: And that is the main differentiator that we believe that would always be in go-to-market, your ability to understand what makes your organization unique and your organization better than your competitors, and iterate on it faster and double down on it. So, Swan is really a tool that is designed to amplify your knowledge and your team in the way that you

553
01:17:52.030 –> 01:17:52.730
Amos Bar Joseph: think.

554
01:17:52.750 –> 01:18:02.350
Amos Bar Joseph: We’ll do it in the best way, and helps you to experiment on it in a super-fast way until you get to that gold mine, so you can double down on it and scale it.

555
01:18:03.440 –> 01:18:27.930
Meagen Eisenberg: Yeah, I keep thinking about Tim’s chart that he was showing earlier around, it’s all about deadlines, and so do you program that into your agents and give them deadlines? Anyway, as a leader, I’m going to take that back. We bought all these AI tools, and we’re seeing all this interesting productivity and things coming out of it, but at the end of the day, what’s the deadline is about where we’re going to deliver. So I find, you know, I think that’s interesting.

556
01:18:27.930 –> 01:18:35.540
Meagen Eisenberg: Well, you’ve onboarded over 300 companies already. What are some of the patterns that are emerging from successful adoption?

557
01:18:35.950 –> 01:18:39.420
Amos Bar Joseph: Yes, so, wow, that’s a question I love, because I feel like there’s…

558
01:18:39.520 –> 01:18:58.450
Amos Bar Joseph: Most AI implementations fail, and it’s the hard truth, and I think that there’s, like, one main reason for it, basically, that if we’ll focus on that, then we would probably, you know, make most of them successful, actually.

559
01:18:58.680 –> 01:19:09.169
Amos Bar Joseph: And the reason is that the entire mindset around implementing AI is trying to optimize it for perfection, not for adaptation.

560
01:19:09.750 –> 01:19:15.369
Amos Bar Joseph: What does that mean is that we foolishly think that we know

561
01:19:15.760 –> 01:19:18.670
Amos Bar Joseph: How should AI be implemented within the company?

562
01:19:19.340 –> 01:19:34.689
Amos Bar Joseph: And we invest a lot of efforts in building that version together with the vendor, maybe internally, whatever, and then it breaks, it doesn’t work, and we say, hmm, the project just… we need to think about something else.

563
01:19:34.940 –> 01:19:39.730
Amos Bar Joseph: But the truth is that we don’t know. This is a new technology.

564
01:19:39.900 –> 01:19:56.459
Amos Bar Joseph: No one knows. No one on planet Earth knows what will work, Megan, for your specific go-to-market organization. No one knows. Not even your organization, but the only ones who are able to understand it is your organization in an iterative process.

565
01:19:56.760 –> 01:19:59.300
Amos Bar Joseph: Okay, and if you…

566
01:19:59.440 –> 01:20:06.520
Amos Bar Joseph: focus on an AI implementation that will allow you to discover, iterate, and change, and adapt.

567
01:20:06.550 –> 01:20:23.030
Amos Bar Joseph: then eventually, in a very fast, actually paced environment, you’ll be able to get to an AI that really moves the needle in a consistent and efficient way. And Swan is designed, is a product that is designed from the ground up.

568
01:20:23.220 –> 01:20:48.189
Amos Bar Joseph: for feedback loops, and for adaptation. And so, it’s not about onboarding for, like, 3 months, and then seeing if it could just, you know, run for an entire year on its own. It’s actually a very lightweight onboarding, and then every touchpoint with Swan just, you know, makes it more adaptable to what you want. It learns from your feedback. You talk to it, you can tell it, I want you actually to be like that. This is what I want.

569
01:20:48.190 –> 01:21:00.769
Amos Bar Joseph: learned that what we’re doing right now isn’t working. We need to change, we need to go to a different direction, and so one can adapt to it. And so, we should start optimizing for adaptation, not for perfection.

570
01:21:01.270 –> 01:21:03.889
Meagen Eisenberg: And here I thought it was just going to be about prompts.

571
01:21:03.890 –> 01:21:21.750
Meagen Eisenberg: If I could just write the right prompt, it’ll give me what I need. You know, it’s interesting, we’re deploying it across 12 different functions here at Samsara, so my events team is using it, of course content, of course marketing ops, we’re looking at data analysis.

572
01:21:21.750 –> 01:21:43.839
Meagen Eisenberg: We’re creating ABM pages and web pages, and we’re doing a ton in, on the creative side with video production and all of that. So I kind of have a sense of the adoption across marketing, but when you look at, across your customers and the go-to-market functions, who’s adopting it the fastest? Is it the SDRs? Is it workflows? Is it marketing ops? CS?

573
01:21:43.840 –> 01:21:45.520
Meagen Eisenberg: Sale, who do you think?

574
01:21:45.800 –> 01:21:57.579
Amos Bar Joseph: Yeah, so… Right now, AI is better in… Automating, low-level tasks.

575
01:21:57.750 –> 01:21:58.550
Amos Bar Joseph: Okay.

576
01:21:58.720 –> 01:22:04.240
Amos Bar Joseph: And… we should think about AI more as, like, an intern, rather than an expert.

577
01:22:04.680 –> 01:22:12.379
Amos Bar Joseph: And the areas that are, you know, you could easily move the needle are the areas where you have the most intern-level tasks.

578
01:22:12.510 –> 01:22:29.980
Amos Bar Joseph: Okay, and so, of course, if you look at the top of the funnel area, where, you know, there’s a lot, like, SDR work, basically, and there’s a lot of pipeline generation efforts in, I would say, relatively low ACB accounts.

579
01:22:30.230 –> 01:22:45.939
Amos Bar Joseph: Okay? That area, you might find a lot of value in, you know, just using AI to automate a lot of the work. But if you look at, you know, higher ACV deals, when still you need a lot of human there in the loop, then you cannot just, you know.

580
01:22:45.940 –> 01:22:58.850
Amos Bar Joseph: throw AI to just automate everything, what you start needing to think, and that’s an important thing, Megan, because I don’t really believe there’s a specific role, by large, that could benefit more from AI. I think the…

581
01:22:58.850 –> 01:23:06.189
Amos Bar Joseph: Questions should always be contextual to who are we talking about, specifically within our company, specifically.

582
01:23:06.190 –> 01:23:12.219
Amos Bar Joseph: Our company. We’re looking at our SDR team, we look at our marketing team, we look at our AEs.

583
01:23:12.430 –> 01:23:27.430
Amos Bar Joseph: what can… how can we amplify them? How can we turn them into the 100x version of themselves by taking AI, automating all the intern-level works that they’re doing, and amplifying what is outside of that?

584
01:23:28.690 –> 01:23:49.449
Meagen Eisenberg: Okay, well, I did read something today that said, Gemini wins the gold at this global coding competition, so I don’t know if it’s just the intern. Sounds like it beat… it got 12 out of 12, or 11 out of 12, or something, for the first time, so it sounds like it’s getting pretty smart and can do some pretty sophisticated things.

585
01:23:49.450 –> 01:23:54.740
Meagen Eisenberg: So I’m excited to see where it’s going. A question… one more.

586
01:23:54.740 –> 01:24:09.149
Amos Bar Joseph: Megan, I’ll actually… I would actually, you know, reply that, because it’s super interesting, and, you know, we’re living in this world where AI gets smarter and smarter every day, and, you know, from one side of things, it’s scary, but…

587
01:24:09.480 –> 01:24:19.889
Amos Bar Joseph: What we’ve seen, and by working with all the best models out there, and really building with it, is that the importance

588
01:24:19.980 –> 01:24:39.689
Amos Bar Joseph: of the human in the process doesn’t get eliminated, it gets amplified in that sense. And because of these, for example, coding agents, so let’s take the frontier of AI. You’re right to say that in coding, AI is the best at the moment, but without a human steering the wheel.

589
01:24:39.690 –> 01:24:55.809
Amos Bar Joseph: It can go crazy and get rogue and get lost very fast, no matter how smart it is. And what we’re seeing is that what that AI developers are doing, they’re opening a window into much more productive environment

590
01:24:55.810 –> 01:25:14.210
Amos Bar Joseph: But the humans that are controlling them in the right way are able to output 10x more output at a given time. And so, what we’re seeing is a world where humans can just create much more by working with these agents, rather than these agents replacing them and doing their work.

591
01:25:14.940 –> 01:25:30.529
Meagen Eisenberg: I see. Well, one of the questions that came in now was, is how do you… how does SWAN look at context? It came in through, our chat, you know, about… not about just about speed, but context. Can SWAN bring in client knowledge, and if so, how much?

592
01:25:30.890 –> 01:25:37.980
Amos Bar Joseph: Yeah. So… What… the way that we look at context is actually through memory.

593
01:25:37.980 –> 01:26:02.700
Amos Bar Joseph: Okay, and that’s a… I hope that we’ll have enough time to talk, because that’s a huge favorite topic of mine, but in a world where agents are interacting with our systems and with our data, etc, things stop becoming data, they’re becoming memories. What does that mean? Is that… and so on, so on can look at all of the account interactions, for example, that it had with you.

594
01:26:02.700 –> 01:26:18.809
Amos Bar Joseph: And it has memory of that account interactions, and can understand, you know, the timeline, what happens, you know, before and after, and can reason about these data points from a memory perspective, because it’s kind of quote-unquote, remember that. It can also remember that your… what’s your ICP?

595
01:26:18.810 –> 01:26:42.549
Amos Bar Joseph: And if you ask it a question or to perform a task, it would kind of, like, remember your ICP, bring it into the prompt, what you would call, and then perform the action with this ICP in mind. But it’s more than that, actually. If you interact with Swan, Swan would recommend you to send an email to Megan, and to personalize it in a way, and you would tell him, yeah, you know what?

596
01:26:42.700 –> 01:26:58.859
Amos Bar Joseph: But let’s change it. I didn’t really like the way that you wrote it. I want it to be written in a more casual way. Swan will remember that when we’re reaching out to CMOs at, you know, Fortune 1000 companies.

597
01:26:59.000 –> 01:27:15.719
Amos Bar Joseph: we actually want to have a casual tone, and that’s how we want to engage, so we have that memory as well. So it’s not only about the customer knowledge, it’s actually about memory… memorizing the interactions with the customer and the data that you have in your systems of record.

598
01:27:16.580 –> 01:27:23.279
Meagen Eisenberg: I see, so it’s sort of like a best friend. It’s gonna remember I like dark chocolate, I like oat milk lattes, all those things, right?

599
01:27:23.280 –> 01:27:23.950
Amos Bar Joseph: Exactly.

600
01:27:23.950 –> 01:27:29.389
Meagen Eisenberg: Well, we’re up at the top of our time. Where can people see Swan in action?

601
01:27:30.140 –> 01:27:49.299
Amos Bar Joseph: Yes, so, first of all, you can go to GetSwan.com, and there’s a video there of how it works. And you can also follow me on LinkedIn. I love to talk about, you know, what’s working today with AI agents, what’s not working, all of… I, like, love to call BS on a lot of stuff.

602
01:27:49.300 –> 01:27:55.709
Amos Bar Joseph: So, you can follow me on LinkedIn, and I also have a newsletter, that is called The Big Shift.

603
01:27:55.850 –> 01:28:08.360
Amos Bar Joseph: So, you can search it online, and if you want to get a front seat to how we’re building Swan as an AI-native business from day one, and you want to learn from our lessons, so you can just follow along.

604
01:28:09.460 –> 01:28:15.189
Meagen Eisenberg: Awesome. Well, I enjoyed talking with you today. Julia, thank you for having us, and I’ll hand it back over.

605
01:28:15.650 –> 01:28:23.819
Julia Nimchinski: Amazing, super insightful session. Thank you so much, Amos and Megan. How about yourself, Megan? How can we support you best, and where should our people go?

606
01:28:23.960 –> 01:28:37.600
Meagen Eisenberg: Well, certainly, if you’re looking for any fleet solutions, dash cameras or telematics, samsara.com, but I’m on LinkedIn, and always happy to have fun conversations around where marketing and technology are going.

607
01:28:38.030 –> 01:28:49.609
Julia Nimchinski: Our pleasure. Thank you so much. And we are transitioning to our next session. Welcome to the show, Mora Rivera, CMO at Qualified, and Brandi Sender, CMO at Epromore.