This panel will focus on Identity's Role in Omni-Channel Customer Engagement and how to stitch it together, or stitch IT together. I would like to start with each one of you giving a quick introduction of themselves.
Ulrich, we know you already, but maybe there are people who just joined, who came in. So, a quick introduction, what do you do, and what is your relation to this topic? Sure. My name is Ulrich Herberg, I'm an engineering manager at eBay, and in Germany here, focusing on identity. In terms of Omni-Channel, on the topic, as I mentioned in my talk, we are, to this last person's question, we are constantly exploring what is the least friction experience for authentication so that while we want to remain secure, we want to offer a frictionless experience.
And that is often, obviously, across different channels, mobile phone, desktop, our platform is offered on all of them, so that's why I was interested in this panel. Thank you. Sadrick? Hi. My name is Sadrick. I'm part of the cidaas management team. We are developing a cloud identity and access management solution, in particular started with customer identity and access management, so we particularly have the same issues or same challenges as you have.
We want to provide great solution, frictionless experiences to the end users of our customers, quite active in B2B, B2C use cases, from retail media to also many B2B use cases. Hi everyone. My name is Fatima Perone. I lead Verified in the Canadian market, and that's our digital identity scheme that's managed by an organization called Interac. Interac is also our payment network for the Canadian market.
When it comes to this topic, previous to this role, I used to lead Enterprise Digital for one of the top five banks in Canada, and the Canadian banks together actually built this Verified product, trying to be very similar to the Nordic markets. I then was able to leverage Verified and embedded it across our organization for various types of use cases in the omni-channel experience, and then now in my current role, obviously both for both public and private sector, we're thinking through how you position the product well, so it can be utilized across channels and use cases. Great. Thank you.
Okay, my first question would be, so when I think of customer engagement, the first thing that comes to my mind is the user journey. What is the challenge in the omni-channel space, so to say, and what role does identity play in that regard? So why is this a topic that we need to discuss? So one thing that, when you ask the question that I have to think of, you talked about the user, but the problem is there isn't such a thing. It's not just one type of user. That's what I mean. So let's say we have a buyer. It's an individual who wants to buy us an item.
We have a seller, a CDC seller, who wants maybe to buy something, a used phone that they don't use anymore. We have small and medium businesses, a small shop that wants to sell some of their items and has maybe two employees. And we have large customers that have hundreds of employees managing eBay. And the user experience for them is very different. For example, like many users, they would like to... The large companies, they may have one eBay account, but they have hundreds of eBay employees. So how do they manage this?
And these employees are sometimes low-level clerks who should not be able to change the payout account or download the tax statements. So it's a totally different requirement, totally different experience that are expected from these users compared to some individual users. So that's what I think is interesting and challenging for us, to make sure that we find solutions that are adapted to these different user types.
Yeah, maybe to add here, I think the same way. There are many complex scenarios in that. So the identity is the foundation for the customer engagement in that use case. So as outlined, you have not one user type, you have dedicated user administration scenarios on the seller side, same on the buyer side. So if we take any retail use case, any media use case, so we know Apple family, Spotify family, so we don't have one customer maybe that might be a group of family, group of friends, a B2B customer use case.
So they all have different flows, they need self-service capabilities, and the data what we have about these help to engage. And in particular, if we take Omnichannel, we have the digital services, but we also have real-world use cases. So we have retail use cases where you come to the real world, you buy online or put something on a shopping list online, then you go into the store, you have maybe loyalty programs which spread across the family completely. So you really have, depending on the use case, different scenarios, different complexities which you need to manage.
And their identity is the cornerstone to get all the data, the permissions, and basically get an illustration of the real world to manage that. Agree with everything you guys have said. If I was to add anything, the way we have thought about some of this user experience and the challenges across different types of use cases is anytime you're able to present the digital ID component of the solution very consistently in different types of use cases, the user experience over time gets better. And when you have the end user recognize that, hey, I've done this before, you also get adoption to go up.
So we've thought about, hey, how do you position some of those user flows? Have it be if a client is doing self-serve, you know, opening an account online, or if they're at the branch and they need to pass the ID verification at that stage, how do you mimic that similar flow on both sides? And that has, over time, also allowed us to overcome that.
And then when I think about as an organization who provides a digital identity solution to the market, we are quite cognizant that we are a mechanism for various types of organizations to provide end-to-end digitization because organizations are trying to either reduce costs for their organizations or simplify the tech stack. So how we integrate into those use cases is also quite important. So it should be a simple integration, and then you're able to choose the assurances that you want to purchase, right? And that also helps simplify some of those user flows.
Yeah, I think that use case identity is also an enabler for many use cases. So you have like the assurance terms and all that, so it's quite often also an enabler.
Yeah, you brought up something that resonated with me, the simple user experience. The problem sometimes, it's not easy to say it's simple, but it is often unclear what it means. So I want to give an example from us. When we introduced pass keys as one authentication factor, as I mentioned before in the panel, pass keys are not widely known yet in the general population. And so we need to educate people. What we do, but we don't know how much education do they need. Do they want very technical details? Do they want to know how long the key length is and how secure this is?
Or are they okay with just very high level, you use your biometrics to have a key? So what we do is A-B testing. So we have two versions, and for small population of our users, we enable one version, version A, and the other small population, we do user B. And we see what the response rate is, and we ask them for feedback later. Or what we also do is we have user experience researchers. So I saw some videos where we covered those volunteers who came into eBay. We let them play with sign in with pass keys. We don't help them. We just let them do it.
And then we later ask them, what was your experience? How did you like it? What did you not like? Some users had no clue what they were doing, and they were saying, oh, this is broken or this didn't work. And so but that means it's not broken for them. It means we didn't explain it well enough. I think one good thing here is to see, so we are exactly talking about customer training or user training, right? And that it shows that you have a quite heterogeneous platform with different user types, as you outlined. And depending on the user, you might experience better adoption or worse adoption.
So we really need to think about what is your use case, how can you manage it, and what guidance do you need? So if you have like a tech-a-thin person working there, they directly understand pass keys. And if you have like a young person who doesn't understand it, so it also shows the challenge in the user journey, in particular, it becomes more complex. So if you really have like eBay, anyone can buy there. So young to old. If you offer pass keys in general to all of them, so the challenge in the user journey on the one side.
But I think what we also might can discuss in that use case is how it enables then the data quality, like the information what we have in the identity for like more personalized experience. So I think it's also for you in eBay, an enabler for like personalized recommendations to buy things if you know about the person. So identification is also a key for like further actions or business values. Any thoughts on that from you, Fatima?
No, I don't have anything to add to that. It's you know, pass keys is something that we're dabbling into and still sort of exploring from our perspective. But it's interesting topic for sure. Sure. Okay. Based on your experience, what are key success factors that differentiate organizations that get omni-channel customer engagement right from those who struggle with that? I think you brought up some good point that the technology, you have people, users who are very familiar with technology.
And they will be fine if you introduce something new, an identity mechanism, let's say an authentication mechanism that is new and technically difficult. They may have already heard about it. But I would say many of our users, like you speculated correctly, are not in that bucket. We have a 70-year-old woman who's never worked with technology much, but she still wants to buy something new and still wants to be able to access eBay. And so if you're only looking at what's the latest technology, what's the most secure authentication mechanism, you will lock out a lot of this population.
And so that could mean failure for a company if they don't also explore the convenience and the friction to those factors. And talking about secure and seamless, which is the title of this panel, I think there are also two other points that you cannot forget. The one is cost. I talked about the cost of SMS OTP. So that can literally kill a small company if, you know, an SMS in Germany via one of those larger vendors where you can send out SMS is about $0.07, to my knowledge. That's a lot of money if you send out a lot of SMS just for authentication. So that can literally kill a company.
And compliance is another thing that can also kill a company if you're not adhering to, I talk about PSD2, if you're not adhering to it, and then the regulators come knocking at your door, they can take away your license. I fully agree. In particular, compliance, I don't want to touch the topic because I will not have a good sleep tonight. So we'll skip that from my perspective for now.
But I think one important thing, as you also outlined in other previous questions, is for me, never stop adapting, which basically covers a little bit also the topic that you mentioned, you explore, make A, B testing, you make like research, you always try to find out what is the current process for each of the different user types, how they react and provide them always the best. So also this continuous learning in that phase. And that's, I would say, normally a logical case. But we quite often see in our customers, they say, some of them at least, okay, what do others do?
Let's just do it like that. And then they stop. They introduce it, the current state. And then one year, two years, three years later, you're still behind again. Because the world changes quite fast. And I think this adaption to new things, learning that, it doesn't mean to directly use pass keys and force all users to do that. It means learn, okay, which users can I offer that? How can maybe authentication registration flow look like in the future?
Test it, explore it, and then see which users can utilize it. Maybe you have a target use case, like an electronic retailer, you might can offer direct to all users. Maybe you're eBay, then you need to maybe find other ways. But it's a process. And I think this never stopping to learn or to adapt to new things is quite important. And that's quite often a deal breaker. I like that. What we've done in the past, and what I would do in this scenario that you're talking about as well, is you introduce it to the channel where you get the biggest ROI.
In Canada, you know, specifically in the banking industry, our fallback when you don't pass on identification is still sending people to a physical location with pieces of ID. And that's really expensive. If you talk about 7 cents, that's like $20. So that really does kill a business, right? So it's starting with those use cases where your ROI is the highest, and then you start scaling it across. And you have lots of learning in between. So by the time you've scaled it across an enterprise, hopefully you've gotten that experience right. The only thing I would say is the feedback loop, right?
You do need a... So I'll just give you an example of some of the challenges we're having right now. So we service obviously all of Canada, public, private sector alike, and we have seven institutions that are both relying parties, but also data providers. When they are seeing any sort of issues on a user experience, and they provide that feedback back, how fast are you able to iterate and actually put those improvements into the ecosystem? Because letting those sit across all types of organizations, it could kill the adoption of digital identity.
You don't want to be in a position where someone is saying, hey, this is really frictionful. We don't want to now scale it to the next use case. So the quicker you can get in iterating on that and fixing that will allow the adoption and the scaling of those experiences. All right. Thank you so much. We didn't touch AI yet, everyone's favorite topic. So do you see any AI support that could help organizations with... I would say AI is embedded in most things, right? In the Canadian market, in our digital identity solution, we even offer the, take a picture of a physical ID and do a liveness test.
Without AI, that would just be full of fraud, right? So there's instances, and I would say almost every use case where you do have some AI embedded in, have it be back office or validation, and even from a user service perspective, self-serve, et cetera, now you can use generative AI to summarize things like complicated technology talk and simplify it to the end user so they can get it. So as much as you can adopt from that, I think there is a lot of place for AI to live.
Yeah, I think what I in particular like is that you separate generative AI and I would call traditional AI right now, because as you outlined in many use case today, we already have a traditional AI or also older machine learning things and all that included where you have like, we also have an ID validation or ID verification solution. We use machine learning in our fraud detection where we try to detect brute force or identity theft scenarios and all that, what you have there in your classic authentication processes.
So identity threat detection capabilities, and you will also have more and more generative AI use case. So I think it's important for AI to identify the use case, use the proper technology there, not because it's the best way with AI, but just because it really gives added value. And I think that's, we also, to be fair, we also announced generative AI capabilities, but quite often you don't see it in the market. They just announce it because it's a good best work with marketing capability.
But we don't talk about where AI, even if it's not generative AI, is already useful and is already applied, not only with our solution, but also with many other solutions out there. Yeah. I think you brought up this distinction I like because it's too often mixed media. Everything is now AI and gen AI, I would not use it in an authentication system to determine whether a user is legitimate or not. I think nobody, I hope nobody does that.
But yeah, we definitely also use fraud detection. We have a ton of data. I don't even know how many giga and terabytes of log data and behavioral data. And then sign in behavior data that we have. No human and no simple system can just go through it. So the attackers are more and more clever and have new attacks that are not always known to us in advance. So having systems that investigate that and at least flag it and then we can look into it more are highly useful.
Gen AI comes, these agents that you see also in our platform more and more are of course extremely helpful also for identity, not just for the customers, you brought that up, but I want to add one point for our customer service. So eBay, we have thousands of customer service agents and those are not highly trained engineers as you can imagine, at least not the ones that are on the phone with the customers.
So yes, traditionally they have their knowledge base and they have a search bar where they can search based on the behavior. But if they have a chat agent that helps them finding out what could be the problem, they could more easily find the solution and direct the user. Maybe the user has been, there's a risk block which blocks the user, that could have happened. And we don't always tell the user for various reasons because it could be a nefarious actor. But at least the customer service agent would then know how to react to this and for example verify the identity in a different way.
All right. We are running out of time already. Time flies. Do we have any questions from the audience? Yeah. There you go.
Ulrich, I'd like to encourage you to keep going with PASCE experiments because I can admit we are a little prioritizing on that, on your work. One of my slides for our business, why we should use PASCEs for customers in several segments is because big players are doing this and it's a safe bet, you know. So let's keep going. So we also quite often have in customer cases where we then outline, okay, one of the big players, they're also using whatever, identify a first approach, PASCEs, whatever.
So it helps to give, because many, if you have a reference which you see it works out, that helps the other customers or helps others to understand it. So true case. All right. We have another question here in the back. I was just wondering whether you had any thoughts on the part of the identity world that no one likes to talk about, the marketing technology piece that is obviously very prevalent in terms of how you place ads and who you target with your marketing budget.
So do you see an omni-channel strategy in the customer and in the access space take that in, or do we keep that to the side because it's a bit iffy? Maybe I can take that.
So we do have marketing and communication departments, as you can imagine, and the education that I brought up in this previous panel is part of what we're doing, and the marketing is sending out emails, mass emails to all users or certain users in a country or so on a regular basis, updating our documentation, our web pages, that we can educate people on these, how to use our customer experience in terms of for digital identity. So yeah, we definitely address this as well. It's not just from engineering or product management that we worry about this.
And maybe I'll take a different angle because just the organization is quite different. So personally for Interact, we would spend our marketing budget more on the B2B side because we want it to scale across the board. And then we depend on that business, on the marketing to the end user, because that experience is going to be very tailored to what they're rolling out. So it becomes sort of like a two-prong approach there. Maybe a short statement from our side.
Obviously, we also want a B2B side, but to also give you some outlook. Our tool, obviously, since we are a customer identity management solution, you have quite some interesting identity management data, and that's like probably you are doing at eBay too. This data is used to have like targeted and personalized interactions, engagements with customers, if that's good or not. From my perspective, I like that because I also personally like to get personalized address on an online shop or something else. How it's used depends obviously on the customer and the use case. All right.
Thank you so much. We are running out of time, but I can encourage you to get in touch with our panelists. We have coffee outside, some networking opportunities. Thank you so much for your time.