Your Next Audience Won't Be Human

Varun Kabra 0:17

Thank you. Thank you so much, Jordan. Excited to be here.

Varun Kabra 0:39

Yeah, no, happy to. So I've spent about 20 years having the opportunity to build growth, go-to-market functions across organizations which are at intersection of consumer trust, data, and more recently at AI. Early in my career, I scaled Android and Google across various products for a decade, and then led global marketing at Revolut during our massive expansion phase around the world.

And then ran growth at Proton, which enabled us to bring privacy products to hundreds of millions and maybe more now, mainstream users around the world. As you mentioned, I currently work as a chief growth officer at Concordium.

We're actually, we're looking at a very interesting problem. We're looking to solve the trust problem of AI because if you think about it, if AI is going to act on our behalf, somebody has to be able to vouch for it. So that's kind of what we are building.

Varun Kabra 2:14

Now, I don't think anybody knows, but if you think about the next wave of AI, it isn't just about people using tools. It's actually AI agents making decisions, whether it's shopping for you, negotiating or paying on your behalf.

And the problem is somebody has to guarantee, even at the first level, A, they say who they are and they're doing what they're allowed to do. And this is what we are building at Concordium, the trust and identity infrastructure underneath this whole agentic economy that allows for agents to carry a trust mark, very similar to how you would think of ISO certifications back in the 90s or even the HTTPS, you remember the padlock, and that travels with them around the internet.

And it's not an easy problem. It's actually exceptionally hard problem to solve, which is also why and what got me here and what keeps me here.

Varun Kabra 3:41

Yeah. Forget the technical language for a second, even the Web3 aspect. I don't think it's a Web3 or a crypto problem. I think if you think about it as a marketer, a lot of consumers are simply going on to ChatGPT and acting on an answer.

They're asking a question, they're getting an answer, and there is no visit to a website, no click through, nothing to retarget. And people like us who grew up on the SEO era, now the reality is there is no page two on ChatGPT.

So if on Google you didn't rank today, you could still sit in a position where you can be found by a really, really dead mind buyer or you can buy your way to page one. Now on ChatGPT, there is no scroll, no second page. You're either there or you simply don't exist.

And I think that's the consumer experience, but what it poses for brands as a challenge is they have now to earn the right to be in that answer. And that starts with something like what an AI system can actually verify.

It's not something which is prevalent today, which is why you see a lot of brands not getting the right picture reflected in ChatGPT or Claude answers, but that is going to become more and more prevalent in terms of how do these systems even verify this is a brand answer. So that's kind of what Concordium solves for, providing a patch to an agent to prove that it is who it is claiming to be.

It is allowed to do these things, and there is a verified human behind it. So that just separates the whole thing because not every agent might have a human, but I think Concordium provides for that connection between a human and an agent.

Varun Kabra 5:57

I mean, unfortunately, most marketing orgs are not even at a stage where we are being completely agentic. I kind of think of marketing organization in three stages of AI maturity. Stage one is you're using AI as chat, do first drafts, tweaking the language.

Stage two is deploying those skills and workflows. And stage three is actually the fully agentic process redesign, but actually the agents are actually transacting and acting with real autonomy. Now, most of the teams, like I said, including honestly parts of my organization, we're still stuck at stage one, but this will very quickly change.

Now, picture a marketing or an agency team in stage three. There's an agent that buys media or negotiates with a partner. And the question is what if it does something wrong?

Who did it? Did the AI do it? That can be the end of the conversation. Somebody has to be accountable, and that's kind of what is really, really important to solve for as more and more organizations go from stage one to stage three.

Varun Kabra 7:50

Yeah, I think the best way to think about is go back to history. If you remember when the internet flourished and Googles and eBays and Amazon's came, the concept of digital marketing came in, and then there was the same question. There are traditional marketers and then there are digital marketers.

I think over a period of time, that divide or that line went away because if your consumers do not differentiate between what is digital and what is not, by the time I go and buy my car at a showroom, I might have already made my decision because you've been guiding my decision making as a brand online.

So I think one has to reflect it in the same way. I think ultimately AI will not be visible. There won't be anything like this is technical AI marketing, this is not, but it'll be part of our workflows.

Now, if one has to really think about how to do this, one has to reimagine the user journeys and your own workflows because the companies which have been successful with this digital marketing, as one would call it, have been the ones who didn't think about, who didn't differentiate.

They said, "We will just follow the user journey and we will serve the user in each of those touch points with the brand message, with the right proposition, which is relevant for them." And I think one has to apply the same thing. So from a workflow, from an organization perspective, one has to reimagine today with the technology, agentic, non-agentic skills, capabilities, all that exist, how would you redesign your workflow?

Where would you put humans? Where would you not put humans? I actually think right now it's such an exciting time for marketers while there is a worry amongst a lot of people that AI was going to take away our jobs.

I actually think for me, it's taken the most boring part away of my job completely, and it's kind of forcing me to do the interesting parts faster. And I think that's honestly very, very exciting because I didn't want to do those parts anyways.

And thankfully I have an agent now who does that and it gives me time, mind space, but also pushes me forward onto that.

So one has to reimagine the workflow, rethink this is where agent makes sense, this is where a human makes sense, and then that's the best process or best design and stay nimble because what we are seeing of AI today is just like 10%. I think this is going to completely change in maybe next month.

Varun Kabra 11:45

Yeah, so I think two things. One, I don't think the principles are very different. Ultimately, if as a brand you keep the idea that you need to super serve your user, that principle still applies, whether it's SEO, whether it's AE or GE or whatever we want to call it, I don't think that changes.

You're absolutely right. The role of earned media is becoming more and more important. I mean, there's a reason why it was called earned media, not paid media, because it was credible, it was important, and I think it's great actually that we're seeing that come up again.

I think from a brand perspective, I actually think it's a very, very interesting time for two reasons. One, one has to not assume that your customer is a human anymore.

Your audience couldn't be people more and more. It could be an AI agent reading your content, comparing the prices, making recommendation to the human on that analysis. Now, a lot of brands are still writing and designing purely for human scrolling, and that's going to create a huge gap.

And brands and content built only for human creation in a world where actually growing share of decisions are going to be made by agents or relied on information which is going to be given by the agents, that's going to create an opportunity.

The second thing I think is very, very important is the role of human judgment becomes even more important. I'll give you an example. So I think there was a stat that in the UK last year we had 8,000 iterations of a beer and pizza ad, which all look the same.

I take the public transport in London when I go to work and you can now, as a human, spot what is AI written and what's not AI written fairly easily.

I think humans are smart enough to now judge the pattern of this is AI written. And the challenge with AI is everybody is training them on the same model. Now we know as marketers and comms people, the best brands have been built on rationality and a little bit of irrationality on top, a belief system.

Expecting a model which is built to optimize for the median of the internet will never build great brands or great creatives. So I think the role of judgment peace makers becomes really, really important.

And I again see a huge opportunity over there to stand out from the crowd. So I know there's a rat race of let's just make everything AI. I think the brands which will figure out what not to AI will probably be more successful in the next phase in my view.

Varun Kabra 15:38

Yeah, so I think let's take those two separate parts. One, the buyer or the consumer is an agent, and in one case, the consumer is human. So we'll take the human example because that's probably more relevant right now.

I think anything which is things like research, first drafts, rule, basic stuff, which is more an efficiency sort of play, I think AI is very good at it. Even in my team and many organizations I know, there is a rational argument that why should the human be involved there?

But anything which is judgment, like in our team, for example, we do have research, which is agentic. So there are a few agents which do that.

We still have the approval, the judgment, the changes, which are owned and evaluated and approved by the human because agents will never be able to, at least for a foreseeable future in my view, be able to understand humans and what sparks us, what excites us. So I think that's where I would differentiate in terms of where a human plays a role and where it does not.

The second opportunity is who maintains not just a repository of insights, but the ability to evaluate what insight is important from a consumer perspective that a brand needs to know and evaluate and learn and do something about it.

I think that's where humans possibly today as of now still have a much better ability to evaluate that, and that should not be given to agents in my opinion, because you will get pretty much the median outcome.

What you're looking for is something which is guided by data, but has that creative element, has that irrationality, has that unknown captured, which has not been captured in the conversations. So I think those are two separate lanes I would think would be very, very useful to keep as for any marketer or for any actually business leader.

Varun Kabra 18:53

Yeah, so knowing your agent basically means three things. One, you know who's actually responsible for it in the background. Second, you know exactly what's allowed and what's not allowed. And the third is a way to check that on spot, not just a claim on the website, which is really important because the internet operates as a web, not as individual sort of websites.

And I think the third one is the most important step that people skip. So right now, every platform, an agent shows up, just giving an example, has to verify it from scratch, like who it is, what does it do, et cetera, et cetera.

There's no passport it carries with it. That's like you going and showing a passport or providing your credentials fresh at every place you go now. We know that is not practical, that is not scalable. I think if I think and take that same thought for a marketing team, it cuts two ways.

One, when your brand's own agent goes out to buy media or negotiate with a partner, you want it to be recognized as a legitimate everywhere it shows up, not just on your own site.

And then when an agent shows up claiming to act for a customer, buying on their behalf, even redeeming an offer, you as a brand on your end or your agent need a fast way to tell a real authorized agent from a scraper or a fraud bot or the same way we are now really good at spotting fake traffic today.

So I think that's kind of where I see the know your agent concept being applied to content marketing areas.

Varun Kabra 21:27

Yeah, no, exactly. So I'll give you an example of what we are trying to do with an ice hockey team. So we are a sponsor of the Danish ice hockey team this year. And even as a marketer, the badge on the jersey is the boring part for me.

What we are doing is the ability for a fan to create an AI agent which can handle everything from, I need tickets to this game, I want this merchandise, I want this experience. And you don't need to worry about those days.

I'm sure you and I have both woken up four in the morning to buy this ticket or ticket which is cheaper, better, and we want to jump the queue or we've queued up, get the right seats. Now you don't need to anymore, right? How do you get your agent to work on your behalf 24/7 under a defined framework?

I am willing to, this is not what I'm willing to. And on the other side, for the national hockey team, there is another agent which can then negotiate with your agent to unlock value from both sides. So that's kind of where I see a lot of these things going and we are very, very excited actually. We will be rolling out at the Women's Ice Hockey World Cup later this year, this technology.

Varun Kabra 23:26

Yeah, because at the background, the agent is verified and you as a human are also verified. So the club or the national team absolutely knows that you are this fan, this is your history with the team, this is what your club level, and you can actually prove that you're a genuine ticket holder or a genuine fan without sharing any of your personal data. And that in my view unlocks a fan experience which is never possible before.

Varun Kabra 24:12

No, I mean we are a platform, so how it works is we might not be the front of the user or we might be in front of the user. It totally depends. So for example, in this case, one has to just download the Concordium ID app, upload their passport one time, and then use that to verify yourself anywhere on the internet.

And then the same way, your agent can inherit the same verification with the constraints you define, and then they can verify anywhere on the internet.

So it's like going to a bar and rather than showing your license, which shows your name, where you live, which honestly, the doorman does not need to even have a look, and you don't want to show. And the idea is you go over there and you give a blue ticket. The blue ticket proves you're allowed to come in and have a drink.

If you don't have a blue ticket, maybe you're underage, maybe you have certain other restrictions. And that's great for the business because they can prove compliance. It's great for the user.

I don't have to show my license to every bar doorman around the world, and the process works amazingly well. So that's kind of how we are designing the whole experience to keep it very simple and easy, and you can port this for yourself as well as for your agent.

Varun Kabra 25:56

I'm really bad at prediction, so I'll give it a shot. All right. But I think there are two things which are pretty high on my mind. One, I think in 2027, can a brand's AI agent be independently verified? That will show up directly in consumer trust.

The same way I think we spoke at the start of the conversation how the HTTPS padlock or the blue check mark became a trust symbol that people didn't consciously think about initially, but then absolutely relied upon. If you remember the whole e-commerce era, everybody was scared of putting their credit card details on the internet until you saw that lock sign.

Now the lock sign has gone away because we're living in a completely different era, but it was kind of like a way to define, can I trust this side or should I not trust the site? I think the same thing needs to apply with brands and their agents.

And I also expect in 2027, the agent-to-agent transactions to overtake the human-to-human transactions for a meaningful slice of commerce industries. So what that means, I guess for brands and marketers is brands which are treating verified AI as a visible trust signal now are going to take disproportionate share of attention when the shift lands.

We think it's going to be 2027. It could be later this year. It might take five years. Like I said, I'm really bad at predictions, but I think it's a matter of when rather than if.

Varun Kabra 28:14

Yeah, I don't think it's going to be a binary shift. I think more and more agents will start getting involved in the buying process. I already know there are a few brands in Europe where actually the buyer and seller agents are negotiating on multi-year contracts.

So I think there are early signs. It's honestly a question of not necessarily technological readiness. I mean, from my years of experience, I never thought, and I have learned, technology has never been a bottleneck.

I think trust is. And the moment you unlock trust, whether it's B2B, B2C, you will see these things because they're just better for the businesses, for the users. I think it's a trust gap that needs to be solved for.

Varun Kabra 29:18

No, I think to me, like I said earlier, I think one thing we should all be aware of, and it's going to come faster than we think, your audience isn't going to be people anymore.

And the more and more I think about it, it becomes quite obvious just because there are so many advantages for delegating decisions, research, recommendations, the same way we have done in the past. If you remember going back years and years when we used to book our travel, we used to go and call up a travel agent and the travel agent would actually either come up with the recommendations or if you really trusted it, will also make a decision.

I think the same thing applies today. So it's just that those agents are not going to be humans, they are going to be technical agents which will operate on humans' behalf.

So I think it's going to be really, really important for brands to remember this and make sure in marketing, communications, design, they keep this in mind as they're designing their workflows.

Varun Kabra 30:42

I think the best place probably is LinkedIn. I think I do share my opinions there about marketing, growth, AI, considering this is exciting, but also a little bit nerve-wracking. So that's the best place for anybody to reach me.

Varun Kabra 31:03

Thank you, Jordan.