Agentic Commerce: Buying Without Buyers

00:00:00: I know that a lot of people said when the computer came up, yeah.

00:00:03: This is a trend.

00:00:05: we don't need this.

00:00:06: It's not optional.

00:00:07: There there is.

00:00:08: no i'm gonna participate.

00:00:10: you will be out When You Don't Participate.

00:00:12: But now it's Not About Being Not Ranked As Good as Possible.

00:00:16: Regarding SAO Its More about being The One Answer.

00:00:20: In Terms We Are Really Changing From SAO To Generative Engine Optimization Or ANSA Engine

00:00:27: Optimisation.

00:00:28: Just get started.

00:00:29: Don't wait, try learn.

00:00:31: repeat that endlessly.

00:00:32: keep an open mind.

00:00:34: don't close the back door of your store.

00:00:38: Roll out the red carpet for all of the urgent

00:00:41: fire.

00:00:42: The boardroom memo powered by net economy.

00:00:50: Welcome to the Bodromimo.netconomy's podcast about the future of commerce, customer relationships and digital business.

00:00:57: In this episode we explore at Chantic Commerce And how intelligent agents could reshape their way companies buy sell and build Customer loyalty.

00:01:06: Todays guests are Manuela Fritzl Head of Consulting at Netconomy and Daniel Reicher customer director at Metcolony.

00:01:13: Let's get into it!

00:01:14: So last week I had a very interesting conversation with somebody who said, actually we decided that we don't really participate in agentic commerce and that got me thinking... And i tried to find the good metaphor for it.. It is like storing your best products you have on physical warehouse and locking door and turning all of lights off.

00:01:39: That's the best example I could think of, because customers will eventually use it and that change is coming.

00:01:49: And its already there!

00:01:51: Yeah... It's a statement to say i'm not gonna participate in this.

00:01:56: You can take back all those big changes over time.

00:02:01: I do remember there were a lot of people, or...I don't remember because that doesn't fit age-wise.

00:02:06: But i know from reading and stuff That A LOT OF PEOPLE SAID WHEN THE COMPUTER CAME UP, YEAH!

00:02:11: This is a trend.

00:02:13: We dont need this.

00:02:14: Nobody will ever have.

00:02:16: No household Will Have A Computer.

00:02:17: It Will Be Just An Office Thing.

00:02:20: Now You Basically Have A computer In Your Pocket Sometimes Like On Your Finger Or Something And it Will be The Same With Agentech.

00:02:28: Its Not Optional There is no, I'm not going to participate.

00:02:33: You will be out when you don't

00:02:34: participate.".

00:02:36: Totally agree with that.

00:02:38: I mean, agentic commerce.

00:02:39: in my opinion it's really the biggest shift there was to commerce as such.

00:02:45: so e-commerce started what?

00:02:47: Twenty five years ago And since then, we optimized everything for human eyes.

00:02:55: The two of us... We have been in so many conversations arguing if the buy button should be green or light-green or dark-green whatever fits better In order to increase conversion by very little percentage.

00:03:07: But what we see now is a total shift.

00:03:10: So Human psychology What we optimize towards doesn't apply anymore.

00:03:17: And I mean, it's going to be interesting because It is not just human psychology.

00:03:24: Before when you buy something for the first time and also do that... You fall in love a little bit with your brand or product based on research.

00:03:36: what we see?

00:03:46: Sixty euros shipping.

00:03:48: now you're already hooked.

00:03:50: You really love your wanted and use it okay?

00:03:51: I'll pay the sixty years.

00:03:53: The agent doesn't care, they agent is just like ok year out of the selection process And i think a lot of companies are not ready for that and also not ready to show more of themselves.

00:04:08: that's on their own sides.

00:04:10: I think this is also a very important point because the agent will not just go on your own side and basically how you present yourself, it's like a Tinder profile where you say I'm the best!

00:04:20: I'm awesome.

00:04:20: please swipe right?

00:04:22: But an agent will look everywhere.

00:04:23: they will also check TrustPilot.

00:04:26: They would check Reddit forums Where A LOT OF CONVERSATIONS HAPPEN AND THEY CAN DO THAT IN SECONDS And That might completely shift The agents perception Of Your Product.

00:04:36: I think it will completely shift.

00:04:38: So the agent doesn't really care about images or very expensive video productions, Or at best Instagram clip where a lot of influencers highlighted that product.

00:04:51: It really cares about data and hard facts.

00:04:53: And as you said before there is only a millisecond Of time when an agent has to look on your data Either have good connector data or you don't and in that millisecond it's going to decide if this option is considered compared to others, or just simply ignored.

00:05:14: And I think one thing a lot of companies forget... It will be more about speed and availability of data because the agent has no priority on serving as customer or user.

00:05:30: You have an intent buy me that product, I want to get it delivered by Saturday.

00:05:36: I need for this and that

00:05:37: purpose.".

00:05:38: It really cares about speed in data structure.

00:05:42: if your website does not give or you're storefront doesn't gives those answers provide with the data is simply skipped.

00:05:53: very much like compare all of conversations we were in last Decade about search engine optimization.

00:06:01: How can you rank one or two positions higher in Google?

00:06:05: Yeah, how can you get away from page to to page one?

00:06:09: that were Majority of the discussions we were in but now it's not about being Not ranked as good.

00:06:16: this is possible on sale.

00:06:18: regarding sale It's more about Being The One Answer.

00:06:23: So In terms, we are really changing from CO to Generative Engine Optimization or ANSA engine

00:06:31: optimization.

00:06:32: And I mean a lot of projects.

00:06:35: customer wise you see a lot as well.

00:06:38: honestly i haven't seen any customers so far who is ready for that.

00:06:42: there's like bits and pieces happening but a lot still very inaccessible And also a lot of information is outside the control of the customer, which will be interesting because everybody's always talking AI jobs.

00:07:01: Will change?

00:07:02: Jobs are gone in the future but I think there'll still be much more new jobs Because as i said already DAI doesn't just look at your store and what data you provide.

00:07:13: it goes very into.

00:07:15: how do we see on this big black hole that is the internet?

00:07:20: And I think there will be a lot of opportunities for companies also to say, we want to proactively manage that.

00:07:26: We wanna go out there and analyze it in real ways.

00:07:32: so not just with bought content on Instagram but really manage that.

00:07:37: personally i'm hoping might reflect the quality some products because its gonna become more important.

00:07:43: But yeah honestly nobody is ready for this.

00:07:46: Nobody has invested into it.

00:07:50: So I totally agree with that and i would even increase that statement in saying the cards are completely newly dealt now, with agentic commerce.

00:07:59: There will be companies who Are experimenting In The Right Direction And there are Companies right Now That They Really Missed The Trend And they really don't care about it.

00:08:11: And that's the part I don't understand, because there are so many studies out there and so many figures already available.

00:08:18: Just one drop-one figure McKinsey stated by twenty thirty three to five trillion dollars.

00:08:26: That is a market volume we're speaking about right?

00:08:31: It's comparable to the GDP of Germany

00:08:35: If you compare it with more money than Elon Musk has

00:08:37: Yeah

00:08:40: with some of money,

00:08:41: which is.

00:08:41: and then just saying you know we deal with that as a small site project.

00:08:46: And creating a couple of POCs here and experimenting on it not merely enough if you ask me right now invested a lot of money in optimizing everything to care about human emotions.

00:09:04: I mean, just think about the production or picture production or video production how much money that costs and how much is invested.

00:09:11: but compare that with small budgets available for agentic commerce prototypes or any development.

00:09:20: so there's not a lot.

00:09:22: focus right now on it And

00:09:25: I think this is also where i see a lot of companies running into the wrong direction because they treat it purely as an, yeah we do in a POC that's driven by the IT department.

00:09:37: Let's see what happens and that's something that I very strongly recommend against doing Because you always need to check the process first.

00:09:46: You want to optimize the processes best possible?

00:09:49: You don't just wanna have a playground Where maybe some nice comes out.

00:09:55: Housing, especially agentic prototypes purely in IT can just fail my opinion.

00:10:04: You definitely got a point there.

00:10:05: and the thing is we are speaking here about the core of every business And especially every retail business.

00:10:12: you want to position your product?

00:10:14: He wants to sell your products in the end and for that you need Describing data and you need a lot attributes product.

00:10:24: And the last twenty five years, we optimized everything in a way that user can see it on the page.

00:10:29: they can read through it.

00:10:31: there are words that really apply to your specific emotions.

00:10:36: but an agent or within Agente commerce that becomes irrelevant and that's the big shift we're right now.

00:10:44: I mean i also think what become more irrelevant is you only buy with one company, because if you like really nerd out on things like expensive coffee machines they have a lot of parts and accessories.

00:11:02: And usually you become sort of loyal to one brand then you purchase also the periphery from that brand.

00:11:10: but the agent doesn't necessarily have to do that.

00:11:13: They can basically create a multi-basket From A LOT OF DIFFERENT COMPANIES To match exactly what you need.

00:11:20: So also that upselling potential, cross-selling completely changes.

00:11:24: I think it's a very nice example because maybe even the coffee machine research will stay... ...because i have good coffee machines myself.

00:11:33: so.. ..i like to invest my time in doing the research.

00:11:37: but when it comes to accessories honestly i really don't care.

00:11:40: But if I'm on the web page of this specific producer, or a coffee machine and then it shows me listen you could add that to the basket because he will anyway need it.

00:11:50: I am tempted to do that.

00:11:52: Because its convenient?

00:11:53: It is convenient true!

00:11:54: And it's shown directly in front where i am right now.

00:11:58: but If the customer journey completely changing or honestly speaking not really there anymore... ...because they just give tasks for an agent say Listen..I have bought This coffee machine.

00:12:09: or I want to buy this coffee machine add all the accessories that i need and try To get them as cheap As possible, And then The best quality is possible.

00:12:19: Yeah By the way it should be delivered to my house on Saturday because That's when i want to brew My first espresso

00:12:26: Exactly!

00:12:27: And i think this Is also something a lot of companies are not thinking about.

00:12:31: It might Be Something where see more partnerships developing between brands, which would be cool because they could like try to get placed together in the AI ranking.

00:12:44: Which again diversifies the market.

00:12:46: so I don't just say that trend is something negative with a lot of people who are all going to be automated away and everything you know Skynet will be done by machines.

00:12:55: but it's really also something where seed.

00:12:57: It Could also uplift smaller brands a lot if they do it right.

00:13:02: it could lead to more diversification on the market because this single brand thinking is a human convenience thing, and the agent doesn't need convenience.

00:13:12: But that means also customer loyalty completely changes right?

00:13:16: Yes I mean you think about a brand then say i have these specific product worked out perfectly.

00:13:22: of course im going to rebuy from that brand.

00:13:25: if ive like colors in the brand or ui web page Obviously, I'm using that.

00:13:33: But if i'm just giving a task to an agent and saying listen buy me this.

00:13:38: and that the agent doesn't care about band love or loyalty so That's a big big influencer.

00:13:45: Yeah, that's something we're looking at a lot of moment how loyalty changes And I think this will definitely be massively impacted by by agentee commerce because All of these earn and burn mechanisms that usually lead to a first transaction, they will become less important.

00:14:01: You need really focus on getting loyalty through personal interaction or special moments you give the customer because earn-and-burn I can get everywhere.

00:14:14: then an agent could find their best discount for me but the agent needs.

00:14:25: Did I like the brand that are not like it?

00:14:27: What did they give me?

00:14:28: what additional things do i value Like, I don't know.

00:14:31: The brand sent Me a letter for the birth of my child which some companies Do because they know.

00:14:38: you bought pregnancy related products and then I could feed That back into the agent.

00:14:44: The thing is about that discussion.

00:14:46: I don' think that its black or white right.

00:14:48: so there's whenever I speak with anyone in the direction Of authentic commerce Everybody's always arguing, yeah then the agent is only caring about price.

00:15:00: Which isn't true in end because you as a user and users will develop our usage of JetGPD And all those LLMs also changed.

00:15:11: like at beginning we prompted completely different than everyone doing right now so it learning process.

00:15:16: The same would be true for Agentech commerce.

00:15:18: when I task an agent with something there will be more attributes I would give to that specific task.

00:15:24: And they won't say just buy the cheapest, so... That's also not how it works today right?

00:15:31: You don't go in a shop and always looking for the cheapest product.

00:15:35: The same with an agentic commerce but whenever their attribute is provided It has to be sustainable brand or one of best quality.

00:15:47: agents care about proof And that's something that is heavily underestimated right now, because just adding this as the best quality product and sustainable won't be enough.

00:16:00: So there need to be certificates.

00:16:01: they need things proving it their need to articles around.

00:16:05: so a lot of know-how that is accessible through the agent.

00:16:08: also I think what gonna be interesting there is because Right Now A Lot Of Companies They Are Starting To Use AI But Using It Is completely in the wrong way for agentic commerce, because they right now use it to generate fake block entries where a fake writer praises their product.

00:16:25: There is tons of fake comments on Amazon and Of course They have improved now into the untrained ID.

00:16:33: look pretty human but The Agent will know that those are AI-generated comments, AI generated articles.

00:16:41: They can distinguish that pretty quickly.

00:16:43: so a lot of the investments that companies are making to fake relatability and to fake human acceptance basically will be irrelevant because the agent would be like nope not trustworthy.

00:16:57: we're not taking them.

00:16:59: So That's also interesting Because there The Human Factor Will Become More Important And Not Less Important Because You Need Real Content from real people.

00:17:10: And structured in a way that it is accessible to the agent, and... That's also one part that shouldn't be underestimated.

00:17:17: We over the last couple of months have read a lot about different protocols that are in place right?

00:17:24: So just speaking about UCP and ACP and A-to-A protocol.

00:17:28: so there's technically a lot has evolved now but I think most companies aren't ready for yet.

00:17:36: You mentioned at the beginning internal processes, so that's one big part.

00:17:40: That is heavily underestimated... ...that needs work but also the IT side needs to adapt to it and It's a big change that it is coming.

00:17:51: And whenever we speak about those protocols A lot of people just mention yeah its like USB-C adapter which technically right?

00:18:01: But there are more behind it.

00:18:02: Its not opening up Whatever you have right now to the agent.

00:18:06: It's also optimizing whatever you have, right?

00:18:09: Now for an egentic buyer and For the next couple of years again this won't be black-and-white.

00:18:16: it won't Be a complete switch from Human buyers two egenting buyers.

00:18:21: so there will would be a dual stream necessary.

00:18:25: But it is not enough if I just optimize my storefront for the human eye And Just ignoring that There Is A Silicon Based Consumer who cares about totally other stuff than the human eye.

00:18:37: I mean, you mentioned an interesting point and i've been thinking about this... About protocols in all of that stuff.

00:18:42: And we deal a lot with very large enterprises and those still tend to have the tendency To build everything themselves because they think it's more secure or more efficient whatever.

00:18:54: But also these protocols will be like Do you choose the right protocols for your market?

00:19:00: A good example is Microsoft Teams, Zoom and Google Meet.

00:19:06: Three different technologies.

00:19:08: and then there is conference software... You kind of can't find a conference software that can do everything in the meeting room.

00:19:14: you either go with that or your go without it.

00:19:16: And I think something similar might happen to all these protocols.

00:19:20: You need to choose very wisely in which direction you go.

00:19:24: So are you going to lock yourself into this one or this one?

00:19:26: Or how do you optimize?

00:19:28: it's also a bit similar to payment providers, what do you offer What Do Not Offer?

00:19:33: and I think nobody has thought about these decisions.

00:19:36: Also from the business perspective where we want To Go.

00:19:39: That right And that is One of The Big Fears i'm hearing In A Lot Of Conversations.

00:19:43: But I Don't Fully Agree With That Because I don't Think with Those Protocols that much of an locked-in effect if you decide to start with one.

00:19:53: Also, I don't think it will be a complete A or B decision either.

00:19:56: i go in this way and so there would be a mixture between them.

00:20:04: And the more you experiment and start early... The more you also start to optimize your data which is going to be the basis for whatever protocol needed.

00:20:18: If you ask me personally what has more chance?

00:20:21: because we see a lot of news and we see, um... What is technically done right now.

00:20:31: I would always tend to follow those companies that have proven in the past To be able do something.

00:20:40: Yeah but as i said large enterprises are like..I'm gonna do it myself And this will hopefully die out little bit And that we get more and more standards there.

00:20:52: So go away from this, We need to do everything ourselves host everything ourselves.

00:20:57: Yeah

00:20:58: But...we can already see it!

00:21:00: There is a lot of standardization already ongoing For example just mentioning UCP or the universal commerce protocol From Google.

00:21:11: I think That goes in the right direction because is publicly available to everyone.

00:21:16: It's not something they charge for, it just a plug that are providing and saying okay this basically the requirements you should fulfill as store.

00:21:27: This what tell me an agent of your stores capable off so can do transactions on an agents behalf etc.

00:21:37: And its kind like universal language will be used Whenever such standards are applied, and I often like to compare it to HTTP for example on the internet.

00:21:49: It is also a standard that started and is widely used in this forming basis of Internet right now And the same will be true for those agentic protocols whether it is agents communicating with each other.

00:22:01: So these A-to-A protocols important at something like that was created.

00:22:06: But he's also regarding how storefronts tell an agent what they're capable of doing.

00:22:12: I mean, but not a heretical question.

00:22:14: Would you give Google or chat GPT your credit card information right now?

00:22:22: That's the right question and honestly speaking personally i won't have a problem with giving my credit card to google because they can already handle safe transactions same for apple pay For example...I know that you also use it so thats not an issue Right now, I would be personally more afraid of tasking an agent with high-valuable transactions without having total influence on what the decision might be.

00:22:53: So we are kind in a step between right now and to create that just needs time.

00:23:02: but you can see it already at this stage where users start experimenting a slow market is forming.

00:23:14: For retailers, the concerning thing right now is this speed of that and how fast it's happening.

00:23:22: so especially with transactions that might be below five hundred euros or something I think will go way faster.

00:23:31: But you made an interesting point there because he said specifically for retailers but we also have a lot of B to B customers.

00:23:40: I do see a lot of potential for agentic layers there to also become multi-layered agents that do stuff, order stuff interact with one another.

00:23:50: But then the volume is not five hundred euros.

00:23:52: it might be...I don't know five million euros if in such large plant and you have agents ordering parts or something that might stop working because of predictive data And there i see an even bigger risk than those get left behind.

00:24:09: There are a lot more infrastructures needed also, not just storefront-wise to the consumer or agent but in the background.

00:24:17: A lot of systems need to be connected and talk with each other about sharing data you have potential points for failure.

00:24:25: I'm really curious how these big BtoB giants deal because they're notoriously slow.

00:24:35: Totally agree, but I would also like to add the positive side because especially those B-to-B transactions are one that is creating a lot of overhead.

00:24:45: Because they usually take time and people involved in negotiating.

00:24:49: just think about a casual example an RFP right?

00:24:53: So within B-To-B we often are confirmed or confronted with RFPs And an RFP usually takes somewhere between one to six months maybe even.

00:25:05: A lot of people are involved, presentations sharing off content evaluating on the other side comparing negotiating and then sharing again.

00:25:15: so there is a lot back-and forth.

00:25:19: A huge part of that can be taken over by agents, communicating with each other and negotiating.

00:25:27: And not in a time frame for six months but within the timeframe of a couple of milliseconds.

00:25:34: So both sides will have complete different basis on decision making In first step or second.

00:25:42: maybe it's even done before them and we just prioritize the transaction itself.

00:25:47: That is good point.

00:25:50: I have another perspective that we played around with a lot.

00:25:54: There was the topic of accessibility and DU, accessibility acting everything... And a lot of things you do to optimize accessibility on your pages are paying into the same thing as what you need for agent commerce work.

00:26:07: but if you think about it step further or something like this i'm very excited about.

00:26:12: allowing agents to perform tasks also opens up whole new layer of inclusivity for people who are currently left out of traditional shopping experiences or traditional consumer experiences.

00:26:29: And that doesn't necessarily have to be something as like a completely blind person, it could be a person with motion sickness who cannot deal with sites where a lot of carol cells are flashing and blinking and beeping all the time.

00:26:41: People with bad eyesight, people who have currently some sort of emotional restriction because they're wearing a cask And agentic also allows them again more to participate in the market, which from an ethical standpoint I really enjoy it out if you think about that way.

00:26:59: It's important because we hear a lot of AI so disruptive and everybody is sort-of scared.

00:27:06: but there are lots good things they could potentially do.

00:27:10: I totally agree, and that's one important part.

00:27:13: And the other part which might apply to everyone is just simply If it's something specifically you have a space for at home and don't know if the measures will fit.

00:27:38: You look on your product, then after twenty minutes reading through everything... ...you figure out okay to dimensions do not really fit so just wasted a lot of time energy.

00:27:50: I think.

00:27:50: generally the world is getting more fast-paced with everything and we have short form content.

00:27:56: Most of people don't even are not able to watch a thirty minute video till end because they already get distracted in midst.

00:28:03: so there's just big change within human psychology, how humans behave with technology.

00:28:15: We want all information now and we want to be done with a certain task real quick.

00:28:22: And I think, right now at the stage where this is already possible it will go way more in that direction.

00:28:30: so things are changing.

00:28:33: The good thing about time would be freed up which i can use otherwise maybe out of nature away from screen.

00:28:44: But for coming back to what will change for retailers We have to understand that agentic commerce is not just automating because we had that right?

00:28:55: I mean Amazon introduced it.

00:28:57: What's five years ago?

00:28:58: even more With Alexa, Just keep repurchasing stuff That i anyway need which

00:29:04: didn't work so well.

00:29:05: i heard

00:29:06: yeah It first of all Is didn't and i think the world thought that it Will impact everything way More.

00:29:12: but we have To Understand this was like A stupid crone job so a simple automation task like every thirty days.

00:29:20: we purchase that because I will anyway need it.

00:29:22: Now you're moving in the direction where i just give an intent and the whole reasoning is done over by the agent, Do you really want to keep on using that product?

00:29:35: Is still the right supplier for it or do I get somewhere else with a better delivery condition, in a better quality and quantity.

00:29:45: Does this make more sense to order only once every six months because we will save in delivery costs?

00:29:50: Those are all considerations humans don't need anymore but the agent is taking over

00:29:58: them To go more again in the negative set.

00:30:01: How do are we sure that?

00:30:03: The reasoning behind is unbiased because That's also a big, big topic right now on GAI market.

00:30:10: like most all AI models are biased Because humans are biased and data we use to train them as biased.

00:30:17: how Do We get better?

00:30:20: I have no answer then.

00:30:24: I don't either.

00:30:25: It's a very good question and it will have a big impact on agentic commerce, where things are moving towards.

00:30:34: And to even add on top of that we did not consider the throwing in money part right?

00:30:42: Same with SEO like you can pay to be first drink I mean,

00:30:48: it's not a business model.

00:30:50: It is the businessman-model of one of biggest companies worldwide right?

00:30:55: You always paid to be on front leading page and you payed for being in one of first three or four links mentioned And nobody even clicked on page two on Google because they don't have time.

00:31:12: So the saying if you're on page two of Google, your dad or something like that?

00:31:16: Yes.

00:31:16: And right now is a fear not mentioned by I your dad but also Google this changing.

00:31:23: I mean, I don't know how much You have already played around with AI mode off Google.

00:31:29: there is no in this Link based results as more official.

00:31:34: Yeah first

00:31:35: you will get write to the answer of your question so There's No need To click On A link.

00:31:41: and IdentiCommerce is doing the same thing, but it's not only showing you a product.

00:31:47: It's checking out your products for you.

00:31:50: And Google itself is completely transforming right now.

00:31:55: But I don't think that they will just take their business model and throw them in the bin.

00:32:01: So i'm pretty sure there'll be payment part connected to it.

00:32:08: They are a business, they have to make money.

00:32:10: somehow everybody wants to hop on the train and It's

00:32:13: not just Google right?

00:32:14: Google is one example but all of the other key AI players will hop on that

00:32:18: train

00:32:21: If you want your product to be shown or outputted by an AI Or recommended by an Ai You'll eventually pay for it.

00:32:32: It's again, not a A or B direction.

00:32:35: So it will be a mixture because just recommending paid for products we'll also not solve it Because the user will quite heavily be disappointed by DAI.

00:32:45: cuz you want to have a big overview of products that are compared with each other and then You want the agent to select really the right one?

00:32:51: And not only the one where The biggest brand is behind was able to afford putting a lot of money on this ranking.

00:33:02: So, it will be a mixture?

00:33:05: Yeah I'm curious how that would go.

00:33:07: and also again usually such ways of working then also inspire small rogue alternatives to come up.

00:33:15: so i am excited for them like example Google and you have Ecosia which is free-to use equivalent based in open models with plants trees for every search and such big corporate decisions usually inspire good things in people as well

00:33:35: to come

00:33:35: up with.

00:33:36: And I'm curious to see those, yeah?

00:33:39: So...I wonder one thing because you also have a lot of conversations with big companies and big retailers and we all are aware there is a change here and we're all aware agentic commerce is approaching quite quickly.

00:33:56: why aren't more Assets freed up for that.

00:34:00: Why aren't more budgets freed up and focus put on the topic?

00:34:06: I think it's two-fold one, you still have parts of companies with no disrespect to the structure or leadership age demographic where they are like.

00:34:23: we're just gonna sit out.

00:34:24: It will be fine if we can watch this which we already said.

00:34:29: that won't work.

00:34:32: And then I think there are a lot of companies who don't want to be the first to fail because at the moment, there is no playbook.

00:34:41: nothing has written.

00:34:42: There's just.

00:34:43: you can try it out and might work or It might not... You can really look at learnings or failures from other companies.

00:34:50: Nobody wants to beat this First Person To Fail Which i personally mindset that limits you a lot because you can learn a lot from failure.

00:34:59: You basically have taken the first step to improving when you fail once, but this is a mentality problem more than it is a technology problem and I think its also how you approach your problems Because as i said A LOT of these POCs are purely IT driven When in reality especially for agentic projects, agentic commerce You also need to consider the process and people who are behind this process.

00:35:28: For example, if you change how people create product data for their commerce shop in free uptime... ...you have to adapt your job profile so that something is done with it at a time.

00:35:44: I think many companies don't see that complexity.

00:35:50: how to staff such a project internally and externally, so that both ends are met.

00:35:56: So the technical infrastructure works but you also serve the process of people behind it and get the best out of it And... That is slowly changing.

00:36:07: I know some companies starting looking at processes first then check How can we help with AI?

00:36:16: But as you said a lot stuck in this POC hell where they're just Chasing the next shiny use case that might not even fit for them.

00:36:24: Yeah, absolutely.

00:36:25: and this Big surge of looking for that specific use case like looking for a holy grail That's ongoing since years now right?

00:36:35: And it never works.

00:36:36: honestly

00:36:37: It never works.

00:36:38: we know that AI started.

00:36:40: We know that a lot has happened in the last couple of years and The speed is enormous and Honestly I really get it but its very hard to keep up.

00:36:49: I mean, how much time are we investing every day just here at an economy in researching what is possible?

00:36:56: What changed which new model was released yesterday?

00:36:59: and it's time consuming to keep up.

00:37:04: But this important... And i have the impression that a lot of business owners are kind of saturated.

00:37:12: so AI topic started and it was put on every roadmap, every agenda.

00:37:17: Then somehow the expectation went from zero to sky high.

00:37:23: so everyone expected AI to be able to automate everything to help with every process And then It kind of hit hard on them that That was not immediately the case because as you said You can't just take a broken process throw AI on top of it With small investment and then think everything is going to work out a part.

00:37:44: that did not happen.

00:37:45: And we see that in figures also, eighty percent of all the PUCs were started failed eventually and didn't lead to result from what they wanted to achieve.

00:37:57: So I think all of this piled up towards stage where everyone is aware that AI in general changes alot already will change even more than agentic AI approached.

00:38:10: Everyone said ok now thats the part.

00:38:14: things will really speed up and take up.

00:38:17: And organizations just can't keep up with that pace, the bigger the organization... The more compliance is there, you mentioned governance at the beginning.

00:38:28: ...the more governance layers are needed and the more certainty needs to be there in order to kick off a big project.

00:38:37: I understand it but still i think especially for agentic commerce, mindsets need to be changed because we don't have three or five years to start and kick off.

00:38:52: There will companies out there who are competitors starting right now that had already started six months ago And those would be the companies which succeed in the end.

00:39:05: they will build up the necessary know-how internally.

00:39:09: They will select the right partners to work with, and maybe also fail at the beginning in a couple of use cases and tasks.

00:39:19: but there would be ones who really understand agentic commerce And then end.

00:39:24: we'll take over that market portion out here.

00:39:29: To highlight it again three to five trillion by twenty thirty.

00:39:35: It's massive.

00:39:36: It is massive and it's a big part of the overall global market that is shifting towards agentic commerce.

00:39:42: You made an important point there when you said PUCs were done but they failed to reach the goal And then people sort of got saturated, I think this was also part of The issue that i currently see especially with agentec because Agentec doesn't just mean you have chat GPT or Gemini referring to your site on your own commerce store, which most know currently as the chatbot that nobody likes because stupid.

00:40:11: but those will evolve as well into something smart and cool.

00:40:15: But there I really see why do such PUCs fail?

00:40:19: They fail because nobody sets an actual goal at the beginning.

00:40:23: Nobody says what is the metrics.

00:40:25: we want to improve by this...I have an example a lot of companies are investing in agentics bot structures that recommend product bundles, because they think customers cannot do it themselves.

00:40:41: It's purely on assumption and they don't measure what is successful when hundred people click the bot?

00:40:49: Are we successful with a hundred transactions are

00:40:51: made?".

00:40:52: What about if ninety-nine of those hundred transactions were sent back?

00:41:01: Does success look like with agentic commerce?

00:41:04: As a question, nobody asks and then they just build something.

00:41:09: And I'm like yeah okay is this successful?

00:41:12: we don't really know actually but nobody's using it must be failure in.

00:41:17: maybe that's just the matter of perspective shift.

00:41:21: also because normal commerce was clear.

00:41:24: your intent as to people have to buy stuff.

00:41:28: That's the success metric.

00:41:30: I think with agentic layers, these metrics get more varied on when the agent thing itself is successful than just purely.

00:41:40: The product was bought because you have a lot of levers that it can take and so projects must become more strategic first in IT later and then strategy and IT.

00:41:54: very close connection um And there are a lot.

00:41:58: companies need to also invest in educating their people, to think more on goals KPIs.

00:42:08: Not a bad way not like your performance is measured.

00:42:10: but what are we working towards?

00:42:14: Are you just using AI for using AI's sake?

00:42:17: or how does... And you can break that down.

00:42:19: Sprinkle the AI and every roadmap right.

00:42:22: Yeah,

00:42:22: and it's also that why a lot of companies fail to adopt AI internally.

00:42:26: because you just say okay everybody has to use AI now.

00:42:30: Every member in my organization gets the ChatGPT Pro account or whatever... Everybody uses it.

00:42:35: but figure out yourself what do they want?

00:42:38: That is going to fail as people are people who need a North Star.

00:42:42: so we have no choice but to enable an organization with AI.

00:42:48: then people will find their way into it.

00:42:50: So we touched on fears, right?

00:42:53: A bit when speaking to business owners.

00:42:55: Fears and optimism.

00:42:56: Yeah, fears and optimism.

00:42:58: but when it comes to fears one of the base conversations I hear a lot is this fear of disintermediation.

00:43:05: so... Right now all of the retailers they own The Customer Journey,

00:43:10: right?!

00:43:11: Everyone wants users on their website scrolling through looking at newsletters really owning that relationship.

00:43:21: And when it comes to agentic commerce, I think the biggest fear is losing that relationship.

00:43:27: now i always like to compare it too.

00:43:30: my store My physical store that I have and I know there are new customers or Customers who aren't willing To take front door anymore but they would Like to come in through backdoor rather than rolling out the red carpet on the back door and saying, obviously you're also allowed in.

00:43:48: I don't close the front door... ...I open up a backdoor to allow them in.

00:43:53: They are just closing it and shutting the door completely And saying no this is money we do not want.

00:44:00: Those are users and specifically agents.

00:44:03: We do not buy on our side.

00:44:06: That's the discussion that i dont really get because nobody would ever say No to additional purchasers.

00:44:14: And it's not that we decide to close the front door of this store, It will be a dual stream.

00:44:20: Exactly!

00:44:21: Those users who like to take back doors or agents shopping for them they'll anyway do.

00:44:26: it doesn't matter if your backdoor is open another one would be.

00:44:31: Yeah and I think its also wrong statement because you're not losing customer journey.

00:44:38: The journey just different and you still very much are in control of that journey.

00:44:43: You're in control how to structure your data, what information so attractive it's.

00:44:52: just the packaging must be multi-dimensional?

00:44:56: And there is a commerce shop at front where you can be pretty and blinky with shiny lights but then you have data on background especially tailored for agentics.

00:45:10: It's basically, you're treating the agent like he would any other customer persona and give it its own journey.

00:45:17: Its not stealing somebody else's journey but has to get his own

00:45:21: journey.".

00:45:21: And here I used to optimize for journeys right?

00:45:25: How many millions are invested in marketing optimizing campaigns of a specific user group?

00:45:32: Organizations free up so much money because they know how it influences buying decisions.

00:45:38: Now, we know that data structure influences buying decisions or consideration decision of an agent.

00:45:44: But we don't free up the budgets to deal with it because... Exactly!

00:45:48: ...we are afraid.

00:45:48: And

00:45:49: also its not rocket science Because you already do all this stuff.

00:45:53: You have your persona for I dunno Your groups fifteen-to twenty five The young students and teenagers.

00:46:02: You give them attributes, you have it for families.

00:46:05: I don't know young fathers... You have all of these extensive personal profiles.

00:46:10: just do the same for an agent and in you can also multiple agent profiles because that's going to be a future where not only this one interface as also consumer but i will have one travel agent who is like my personal advisor.

00:46:30: I have a beauty agent, whatever.

00:46:34: For example you could really spin that and those are just additional customers in your journey who need their own

00:46:45: journey.".

00:46:46: Those are customers right now might not be using agent e-commerce or delegating their buying tasks to agents but they will for one of two years.

00:46:57: so we already see this market is slowly changing, or actually not slowly.

00:47:03: It's very quickly changing and rather than forbidding the re-entry of those users who might have purchased with you before and always came in through the front door saying we're not allowed to come in with your agent anymore I don't think it was a right approach but Manu... We are gonna

00:47:22: close our shop like that.

00:47:24: Konami will say let us get rid of your UI, nobody needs it anymore.

00:47:29: That's not going to happen... There will

00:47:30: always be human buyers still and that is the thing.

00:47:34: AgenteCommerce is approaching its changing how e-commerce fundamentally works but it isn't taking away everything we have optimized or done because they are still needed.

00:47:46: We need a dual approach in future.

00:47:48: We need focus on good looking and working UX a customer journey that really applies to my personas and the people I want to focus on.

00:48:02: But, on the other hand i also need to put a lot of effort and focus on this backdoor or on my databases ,on my connectors to agents so allowing them in as well.

00:48:16: Yeah

00:48:16: And As I said before it might help you as company free up resources too.

00:48:20: focus more On your actual product now because The buyer might not be a human, but your user always will be a Human.

00:48:30: The person wearing using sharing your product is A Person.

00:48:36: so focus on that and just allow them to get To Your Product Faster.

00:48:42: Totally agree.

00:48:43: So Manu maybe some closing words.

00:48:46: If you're now sitting in front of one of those big retailers, business owners and C-level what's one advice he would give them right now?

00:48:57: One advise.

00:49:03: Just get started don't wait try learn repeat that endlessly.

00:49:11: keep an open mind And also Don't stop because I think a lot Companies say if we change this one thing all will be fine and everything is golden, but it's just its endless optimization.

00:49:28: And there might be things in two years with a genetic commerce that we can't even imagine right now Could totally be.

00:49:35: I mean five years ago.

00:49:37: I don't think we imagined using ChatGPT as much as we do know almost

00:49:42: like

00:49:44: or Gemini.

00:49:46: So always Be aware That something might hit you and as harsh it sounds your business model is never safe.

00:49:58: And you?

00:49:59: I would say don't close that back door of your store, roll out the red carpet for all of the agentic buyers.

00:50:10: optimize and Don't lock it because you are afraid that additional buyers or different buyers might come.

00:50:19: You're not taking away your adding And I think that should be the mindset.

00:50:24: Keep that door open and roll out a red carpet!