When Bots Come Knocking, Human Answers: POV with Jay Benach

Speaker A: Hey, it's IB Moss. And this is another episode of Insider Interviews where you get the insider scoop on the business of media marketing and advertising. This is a short form series and it's called POV Possible because we're here at Possible, the fourth year of a terrific conference. One of the themes here at Possible is connecting humans and, uh, everything that's happening in tech. So. So the scoop today is coming from Human executive Jay Benik.

Speaker B: Great to see you.

Speaker A: Thank you. Great to see you as well.

Speaker B: Thank you.

Speaker A: As part of this series, I have certainly met some people with some very interesting names. We've spoken to the cool company and now there's Human. So I kind of love the fact that you're pretty blatant about what you call the company. Tell me about why it's called Human and a little bit about how that came about.

Speaker B: Sure, sure. The derivation of Human Security is an evolution of some of our earlier company names. We started out as Bot or Not llc.

Speaker A: Bot or not.

Speaker B: Correct. The company was started as a. Think of it as a science experiment. Right. If we could identify bots with precision, uh, we then evolved to ydops, which is very much a. Think of it as a proactive defense, uh, concept.

Speaker A: Okay.

Speaker B: And as a result of further evolution, we said, look, what we're ultimately looking to do is to be able to identify and be able to help our customers understand the real and right human that's impacting their business.

Speaker A: So that's the whole crux of it really, is the real or the right human?

Speaker B: That's correct.

Speaker A: Okay, so you know me and my puns. What labor pain did you set out to solve for. How would you describe that?

Speaker B: The beginning of the company? The, the, the founding insight is that we were able to distinguish with very high precision when a browser session was being operated by a bonafide authentic real person versus it being automated or scripted in some way.

Speaker A: Uh-huh.

Speaker B: That is the foundational insight. And from that we built a lot of capabilities and a lot of solutions to meet the needs of various Personas in the market. But the fundamental concept is being able to distinguish without friction.

Speaker A: Okay, distinguish without friction, Bot or not.

Speaker B: Correct.

Speaker A: Okay, great. Um, Jay, you had a bit of an interesting origin story yourself too. You, uh, did you work in gaming before this?

Speaker B: So prior to this, I was doing a turnaround of an ad network focused on video games, where the concept was to run advertising inside a variety of different types of video games. And, um, from there, what we found is that when we would engage with certain gaming publishers, we said Gosh, how are they able to deliver so much of an audience so quickly? Uh-huh. And I was always perplexed by this. I couldn't figure it out. There were medium sized companies out there that one would have never heard of, not household names that were able to deliver phenomenal amounts of audience. And it always didn't sit right with me. And then a couple of years later, through a friend of mine, I was introduced to the founders of what was Bottlenut llc. And they described the problem they were trying to solve and they thought, hey, maybe it might work for digital advertising. And I said, gosh, I think you're onto something with that. I've seen stuff that doesn't make any sense. And from there, uh, what happened was, is the people in advertising started to say, hey, can we try this? Can we take a look? And it was very hush hush. Uh, and at the time it was something that was very uncomfortable to even talk about. A definitive way to identify in black and white clarity what's real and what's not in advertising. And that is what we brought forth into the market. And it was the type of thing where we made it available. And the thought leaders in the industry, they are the ones who said, we want this, we want this, we want this. And then there was a wave of followers after that.

Speaker A: Wow. And that was a dozen or so years ago.

Speaker B: Yeah, we're going back to the 2013, 14ish area.

Speaker A: Yes, I remember those years.

Speaker B: Mhm.

Speaker A: Okay. Uh, so the problem it seems has only escalated, um, the proliferation with AI and everything. And um, non humans are much smarter machines and so it's really kind of ubiquitous. It seems that we don't know how to decipher between humans or not. Tell me now about AI and what's happening.

Speaker B: So, so from our perspective, and me personally, this has been quite a crazy experience when, when we started this, uh, basically people didn't initially believe us that there were bots on the Internet at scale that were focused at consuming pages, content and ads. And initially people doubted us. And it wasn't until we showed them the technology and they were able to evaluate it themselves that they had their own personal aha, uh, moment. Now what's been so surreal for both the company and myself 12, 13 years later, is now everyone is talking about the agentic web automation bots and we no longer have to be like, so there's this thing you can't see. Okay. And we're the only ones that can show you what this microphone. So now the conversation is very much advanced. Where it's not just, oh, is this a bot or not? Is this a human or not? It's okay, there's humans, there's bots. What bot is this? What is it doing here? Why is it come to visit me? Who is operating it? Who's claiming to operating it? How, how often does it visit us? All these types of things. So now the demand for precision and insights.

Speaker A: Yeah.

Speaker B: Has skyrocketed.

Speaker A: So this sounds like something out of the wizard of Oz. When Glinda says to Dorothy, are, ah, you a good bot or a bad bot? Are there good bots? Get it? Good witches, bad witches.

Speaker B: So think of it this way. Uh, there are bots that are operated. And remember, the bots don't necessarily operate on their own. They are controlled. You can think of it as like an invisible line, like a marionette is controlling these bots. Right. And is the controller of the bot, are they trying to do something that's only good for them? Oh, are they doing it on behalf of others?

Speaker A: Yeah.

Speaker B: Okay. And the thing that they're doing on behalf of others, hey, is this for good? Is it constructive? Or is there actual evil intent, criminal taking, theft involved? So that's, that's the lens that one has to use. So, you know, where do you go in, in this? Well, couple different ways to think about it. Um, there is, in the criminal underworld, there is specialization, just like all forms of industry throughout history. Okay. The people that write the exploits that are able to get onto someone's computer are different than the people that distribute the exploits. Okay. That's the way the criminal element operates. And if you're looking to do bot activity and you either want to try and not get caught.

Speaker A: Mhm.

Speaker B: Or you want to do it, uh, to accomplish a specific type of evil outcome, that's the path you'll take. Other people will say, oh, no, we're operating a bot, we're publicizing it. And the most basic form of it is everyone kind of understands the Google bot, where their website gets indexed, visited by Google, to be included in Google search results. I think everyone understands that as a very basic form of what would they consider a good bot. Google has convinced everyone to say, listen, let our bot come to your website, then don't block us and you will appear in search results.

Speaker A: Oh, oh, oh, that's fascinating.

Speaker B: That's by definition what many would consider a good bot. Now fast forward to today where the stakes have gone up.

Speaker A: Uh, yep.

Speaker B: Right now you have a scenario where as a result of large language models, Right. Um, people think of it as the chatgpt or the GROK interface that so many people are familiar with. In order for those experiences to be current, fresh and accurate. Where do they get all their information from? Yeah, they have to get it from the Internet, they have to get it from the web. They have to either operate their own bots or subcontract to other companies to operate bots to get all that knowledge. So that when you ask ChatGPT a question, it gives you an awesome answer. This phenomenon of LLMs, call it going back two, three years, has caused an explosion of VOD activity.

Speaker A: Okay, uh, this is kind of a sidebar question, but do you have an opinion about LLMs having advertising?

Speaker B: Oh, um, LLMs having advertising is an awesome evolution of the brilliant insights from ooverture and Google AdWords.

Speaker A: Uh-huh.

Speaker B: Which is imagine paying only when a person expresses interest or intent.

Speaker A: Exactly.

Speaker B: In a very specific area. And I can target my money to spend and I can spend as much as I want to get the highest level of intent and precision on that. With LLMs, it's brilliant. Imagine being able to surface, uh, your brand. Mhm. Surface your product.

Speaker A: Hm.

Speaker B: In the context of an interactive conversation with an AI model. It's incredibly impactful. Perhaps even more so than a simple search from Google. Why? Because now it's not just being mentioned, there's interactive dialogue about it. Wait, I don't understand. Does it come, uh, is it in canvas or is it in leather? Right. Where is the leather sourced from? Now if I'm in the sneaker business, that is a click. That is an ad opportunity that is really worth it to me.

Speaker A: I see. And it's less sort of skippable, if you will. We're so used to sponsored search results being at the top of the Google results that we just know to skip down anymore. Now it's sort of embedded into the LLM type of answers.

Speaker B: Uh, think of it this way. Some of the brightest minds in advertising are now working to figure out how to make a high impact, rewarding, ad oriented experience within the LLM environment. Right.

Speaker A: Yep, yep.

Speaker B: That's the idea. And I promise you by November it will be improved and by next November will be further.

Speaker A: Absolutely. It's already improved from last week, Correct. Right. So I want to get back to the value prop of human. I don't know if you're game for this, but let's try it. I started a new segment, um, in insider interviews and it's called Pitch me, Pinch me. So you get to pitch me your best like 10 floor elevator pitch um, as let's say a layperson or a quasi educated marketer on human. And I get to see if I want to pinch myself. It's so good. And I have to buy it. Game.

Speaker B: Sure. It fundamentally comes down to, do you want to know if you're being robbed?

Speaker A: Yes, I do.

Speaker B: If the answer is yes, there is a solution to that. If the answer is, well, I'm not sure, that would be uncomfortable. Okay, that's maybe a rock I don't want to turn over. Well then no. You don't want the world's most powerful differentiator of Internet at advertising traffic. You follow me?

Speaker A: I do, yeah. Okay, I'm sold.

Speaker B: Okay, thank you. That's the basic premise of it. And like all things, um, you know, there's regular lenses, there's magnified lenses, there's high quality glass. Um, people who want to get down to a level of precision and not understand even further, not just if they're being robbed, but if they want to get into who is trying to steal from me, that's the next level why people want to run this sort of advanced technology.

Speaker A: Uh, so thank you, I buy it. I'm pinching myself. Uh, I hope nobody's robbing me. I want to understand something that I think you've said before that, um, viewable doesn't necessarily mean human. Oh yeah, it's a little bit unsettling.

Speaker B: So how did we get here so many years ago? Going back to the early 2000s, if you think of what a basic computer browser experience is, a lot of people in digital advertising realized that there were a number of ads that would be served, but they would be served. The terminology was below the fold, getting off the screen. And so the ad server would say, hey, I served your ads.

Speaker A: Right.

Speaker B: But to the actual practical marketer would say, I know, but no one ever saw them. So the concept of viewability became a fundamental precept, uh, of advertising. Hey, I gave you some ads to run and you showed them, but were they actually viewable? Right. And this was the most basic tenant, and it's very reasonable. Okay. Who would argue with that?

Speaker A: Mhm.

Speaker B: The reason why, like all things, it gets more complicated. Okay.

Speaker A: Of course.

Speaker B: Okay. Is because as technology evolves, what is viewable becomes a more complex question. Okay. Because in a world where people are scrolling, well, the ad was viewable, but for how long? If it's viewable for half a second, does that really count as viewable? And again, one has to have empathy with marketers and advertisers to be like, look don't make me play police. Right? We're spending a lot of money here, okay? You got to guarantee me that if I'm paying you to show my ads and you show my ads, they're like someone, if you say they saw it, they actually saw it, right? And so standards developed about what the definition of viewable is, okay? Then people said, oh, well, great, now that it's viewable on page for a certain amount of time. But wait a second, what amount of the page is my ad taking up? Is it like a little skinny thing in the corner, okay. Or is it sort of a, uh, predominant or reasonably sponsored piece? So further standards were developed, right? Then where things start to get really complicated is. And like all things, the criminal mind is generally superior. The bot operators who were involved in designing bots to inflate audiences for many websites figured out that in order to make their audience desirable, they had to make sure that the ads were perceived as viewable. So the bots were designed in a certain way to scroll the page, move the mouse in a certain way to make the ads come through as viewable by traditional measurements. And so that's why you have a large number of people say, oh, but my ad was viewable. Yeah, it was viewed by a bot. It's completely fraudulent. That's the basics of viewability and why people come to us, okay? As a result of that, people said to us, human, can you at least tell if it's viewable? So they twisted our arm and we did the thing to make it possible to say, listen, according to, uh, the standard practices, the way the Media Ratings Council specifies it, now we'll tell you what's viewable or not, but fundamentally, a viewable ad that was viewed by a robot that you didn't want is not what you want.

Speaker A: Yes. And you don't have to pay for it.

Speaker B: Yes.

Speaker A: Okay, does that help so much? And terrifies me at the same time because it's just the tip of the iceberg. Still, the iceberg continues.

Speaker B: The good news is that viewability is a fairly well understood concern and many companies are able to help advertisers and their marketers and agencies work through some

Speaker A: of these complexities, thanks in part to human technology. When brands are now able to kind of filter out this non human activity or low quality traffic and they see maybe their results drop, does that mean that they're anxious? Is it more clear to them? We've been facing things when, uh, rating terms change and now we're losing what might have been perceived as, oh, look, at all the impressions it delivered, et cetera. How do you calm an advertiser? How do you explain that?

Speaker B: So there's the front end, which is the ingestion acquisition of customers that marketers are facing where they are running advertising to acquire customers. Once you eliminate all the stuff that you don't want.

Speaker A: Yeah.

Speaker B: The non viewable, the fraudulent activity, the things that are deemed not okay for

Speaker A: this brand, the bad witches or whatever

Speaker B: you want to call it, there's going to be less audience available.

Speaker A: Yeah.

Speaker B: Supply and demand just means the price per ad that one pays will go up. But that is not necessarily a bad thing because what's perceived as a low price might include a whole bunch of junk that you don't want. So on the front end, the net impact of this filtration. Mhm. And the tech that it's used to filter this out is just a higher per unit, per impression, per click, per whatever dollar cost for the advertiser. But that's not a bad thing because they know that what they're getting is real.

Speaker A: Yeah.

Speaker B: Right on the front end. Now what's changed in 2025, 2026 and is coming is now brands, advertisers, e commerce companies now have bots and agents coming to their website, uh, more frequently and to accomplish more goals. Where it used to be just a Google bot indexing your site to get some basic content for inclusion, or it might have just been a tool that you wanted to run to measure your site's uptime. Now you have what are known as sniping bots where competitors will come in and put products in shopping carts which will skew your metrics to make you think that you should be either retargeting this bot or retargeting this product set on the web and depleting your campaigns. You'll have competitors who hire companies to basically click on your ads.

Speaker A: Uh-huh.

Speaker B: Click on your search results to deplete your budget, to skew your metrics so you target the wrong thing. Now this sounds, it's this, this gets to the point where it sounds very uncomfortable. All generally speaking bonafide top tier marketers, the people that are here for possible to expand and grow their business. They are generally not hiring a criminal enterprise to do these things. They're hiring a marketing consulting firm, they're hiring a growth marketing agency.

Speaker A: Mhm.

Speaker B: Okay.

Speaker A: Like a salient MG for example.

Speaker B: All types of companies that are designed to help marketers maximize impact along the way, to improve KPIs along the way to improve actions to Improve performance of campaigns. Mhm. Performance Marketers, ad operation experts will subcontract, will use certain tooling that brings about and attracts this type of behavior, the unsettling behavior. And that's why when human, today we talk about trust. If you think about it from the perspective of an E commerce brand advertiser, it's really hard. How do you grapple with what your CDN is telling you?

Speaker A: Mhm.

Speaker B: What your credit card merchant uh, provider is telling you? How do you deal with what your customer service returns department is telling you? How do you deal with what your marketing agency is telling you about what's working and what's not working at the front end? And so therefore it becomes very difficult to say what is ground truth here. What can I trust to know what to go attack first? What we have tried to do is take our solutions that were meant for advertising, which is the acquisition side of things, to make them very usable on site for brand marketers, for brand advertisers, for E commerce providers, so that they can understand what's coming to their store, what's coming to their site, so they don't optimize for the wrong thing.

Speaker A: Right?

Speaker B: Uh, yep. And from a very, very like primitive basis, all the marketers understand, which is, is this my customer? Is this someone I want to be my customer? Or is this someone here for a different purpose?

Speaker A: And a someone is, uh, Right. Yeah.

Speaker B: And if, if there's one thing we all need to have, it is empathy.

Speaker A: Mhm.

Speaker B: For the, for the marketer, for the advertiser. Today, when a bot or automated piece of traffic shows up at their website, the question is, is this thing here to take from me and do not so good. Or is this bot acting on behalf of a customer who's doing a price comparison, who's shopping for products and wants to genuinely understand what I have to offer? How do I make sure I give this bot and the bot operator, in this case the bonafide human shopper, what they need so that they hopefully pick me and buy from me or come back from me?

Speaker A: Yes.

Speaker B: And so now advertisers have to contend with, well, I'm running ads that talk to humans. How do I serve the good and the right bots? How do you do that? Well, no, no, hold on. Human can help. You know, if the bot that's visiting you is who it says it is, um, or if it's being hired by a nefarious actor. But once we, once we tell you that this is a bot. Yes, once we tell you it's not coming from one of these dark pool things. Okay. It's actually coming from a shopping comparison engine that is user instantiated. Now it's up to the marketer, the advertiser to say, look, what content? How do we make it easy? How do we deliver our pitch to this bot in a way that the bot can synthesize the content and position us back to their master, that is the shopper, in a way that is compelling for the shopper. And that's an area in which one has to have tremendous empathy. And this is the vertical learning curve for state of the art marketers right now.

Speaker A: Jay, I feel like I just got a PhD in 20 minutes.

Speaker B: Welcome to Miami.

Speaker A: Since I think a lot of people are going to need some human help, how do we find human, um, humansecurity.com.

Speaker B: right?

Speaker A: That's easy.

Speaker B: Our marketing team has put a tremendous amount of effort into making us easy and accessible.

Speaker A: Okay.

Speaker B: Um, the best experience that we have when customers is when they come in. They're able to be genuinely open, kimono and transparent with the concerns they have. And believe it or not, every advertiser is a special unique snowflake. Everyone has their near term goals, they have their corporate principles. There's aspects of corporate social responsibility here. Okay. And many marketers are now playing, playing on a global, on a global basis. It's no longer just one market. So now marketers have to be sensitive and have to genuinely have a global palette. Right? Because when you're delivering a message like, yeah, you think it's coming from the United States, but what if it's not? Right? So these are the types of areas where people, um, when they come in and they're able to be transparent with what they're looking to accomplish, I think we're able to help them that.

Speaker A: Jay Benik, thank you so much for joining GM Media Security. Human and a human.

Speaker B: Thank you so much.

Speaker A: Well, I hope you got some good insider scoop from this episode of Insider interviews with me, E.B. moss and media and marketing pros. If you did give it a like, give it a share. And if you'd like to learn more about how you can get your own podcast, reach out to me@podcastsossappeal.com or follow Insider Interviews anywhere. Thanks again for listening.