Latest / CXOTalk / Ecommerce Strategy: AI-Powered Search and Discovery
Transcript
- 0:00We're discussing personalization for e-commerce and Retail with
- 0:06Sean Milani, the chief technology officer of algo Lea.
- 0:11We power it search and browse and recommendation experiences
- 0:15across 17 thousand websites all over the internet.
- 0:18Where actually the second biggest search engine in the
- 0:20world behind Google, we power over 1.75 trillion requests
- 0:26every single year. So a very exciting powering a
- 0:29lot of the fun. Datian the internet.
- 0:30I have to say that at cxo, talk. We're actually one of your
- 0:34customers we use. I'll go Leah, on cxo talk.com.
- 0:37It's great to talk to one of our customers, and it's great to
- 0:41hear that. We can help our your experience
- 0:43as well. You're focused on e-commerce and
- 0:46Retail. So why don't you give us just
- 0:48the high-level overview of that? We actually serve customers
- 0:52across a whole range of Industries, but e-commerce is
- 0:55obviously the one of the ones that were best known for the
- 0:57reason is because search is so important as part of Whole
- 1:00Discovery journey in e-commerce, people love e-commerce stores.
- 1:03And one of the reasons is because they get such an
- 1:05incredibly wide selection of products that they can shop for.
- 1:10But it's also hugely overwhelming for a shopper.
- 1:13It would be like walking into a physical store and it's the size
- 1:16of like, a huge Stadium. You would just have million
- 1:19items to choose from but it's it's very overwhelming.
- 1:22They call this the Paradox of choice and more choice.
- 1:25You provide a customer often the less satisfied they are or the
- 1:29more overwhelmed They are and their experience can actually
- 1:31deteriorate with more choices. E-commerce brings a huge
- 1:34opportunity. But it's also means that we have
- 1:36to have new tools and Technologies to be able to help
- 1:38customers and guide them through this experience, showing how do
- 1:41advancements in Ai and machine learning such as chat GPT
- 1:48effect, this search landscape as you've been describing he hopes
- 1:52that around 20-30 years now but the experience to a shopper
- 1:56still feels very much. Like it could come out of the
- 1:581990s during the.com. If a lot of e-commerce, sites
- 2:01are really just a product database or a product catalog
- 2:05that have a user experience on it.
- 2:07You still have to talk to the website and the way you would
- 2:10talk to a computer, right? You have to type in like
- 2:12specific words, you have to click on different filters to
- 2:16kind of narrow down the selection and this is going to
- 2:18change dramatically now. So a lot of these language
- 2:22models that are powering chat GPT of exploded in size and
- 2:26sophistication over the last few years and there.
- 2:29Now able to actually really understand humans in a way in
- 2:33which humans can use not natural language and understands the
- 2:36intent and the concepts behind what they're looking for.
- 2:39Sean can you elaborate on? How may I help some make this
- 2:45shopping experience, better for the user, you don't need to be
- 2:49able to talk to the computer anymore.
- 2:51The computer can actually understand you.
- 2:53So what we're seeing is people are using far more expressive
- 2:56language. Now, the number of words that
- 2:58they're typing into the search box, Is becoming significantly
- 3:01larger. It's like doubled in size in the
- 3:02last couple of years, but also people are expecting experiences
- 3:06that are far more sophisticated, and personalized.
- 3:09As I enter in searches or I click on filters or a view
- 3:13products, it expects the experience to adapt to me to
- 3:17really understand what I'm looking for to remember me, when
- 3:20I turn up to the site, Shawn you've described how natural
- 3:25language search just makes it easier for end users.
- 3:30Your Chief technology officer of algo.
- 3:32Leah, can you give us a glimpse behind the scenes of how this is
- 3:38possible? There's been a real Quantum Leap
- 3:40and what we're able to do over the last year or two, as Chachi
- 3:44beauty is really demonstrated to the world that you can interface
- 3:48with users using natural language.
- 3:50Both as the inputs to ask questions but also as an output
- 3:53to be able to answer their questions.
- 3:55But we're fundamentally changing the way in which we search and
- 3:58retrieve information with these Large language models in these
- 4:01new breakthroughs. So, instead of using the actual
- 4:04word, we're now actually taking words and we're turning them
- 4:07into something called a vector. And the vector captures, the
- 4:10concept, and the Nuance behind, either the word or the phrase of
- 4:15the sentence or your question. That means that we can then go
- 4:17and search through all the products are all the web pages
- 4:21for other vectors. We turn everything in the index
- 4:24into a vector and then it becomes this kind of exercise
- 4:27where we try to find similar vectors.
- 4:29Just like you use words and try to match them with similar words
- 4:32and web pages. But it means that you can do
- 4:34really, really interesting things to give you a really
- 4:36powerful example is if let's say you go to a website and you
- 4:40actually want to shop for a specific brand.
- 4:43So let's say I want to North Face jacket and if the site
- 4:47doesn't sell that brand, right? It's able to understand the
- 4:50concept behind the brand. So understands North Face sells
- 4:53outdoor jackets and it can Surface other outdoor jackets
- 4:56that are similar to North Face. We can actually find
- 4:59Specifically designed to solve those problems.
- 5:02Even if the words that you're using, when you describe the
- 5:05problem, don't match the words in the product.
- 5:07How is this different from traditional search, whether on
- 5:12e-commerce sites or just in just broadly, traditional search is
- 5:17literally just taking the literal words that you type in
- 5:20and trying to find products that use those literal words.
- 5:23You know, a good example is if you wanted to search for
- 5:26chocolate milk or milk chocolate, these are two terms
- 5:29of exactly the same words, but very different categories to
- 5:32search for. And so it can be very, very
- 5:35confusing sometimes for these search engines to be able to
- 5:38disambiguate the actual words. But when we can translate them
- 5:41into very specific Concepts and really understand the intent
- 5:45behind them. Then we're going to get far
- 5:47better more accurate results. And I have to say, some of our
- 5:51early customers are just seen incredible increases in the
- 5:55amount of conversions, and the amount of things that Shoppers
- 5:58are able to buy because They're able to find it.
- 6:01Sean you mentioned Vector search.
- 6:04It's a topic that has been around for a long time yet.
- 6:08It's not widely discussed. Why is that?
- 6:12And what are you doing with this?
- 6:15We've known that vectors are a better way to represent.
- 6:17Concepts Than Words, we've known vectors work from a scientific
- 6:21perspective and we can get great results in the lab but it turns
- 6:26out that they're really hard to scale.
- 6:28Like I mentioned we've Got we do about 1.75 trillion, search
- 6:33requests a year, across 17,000, different websites and to be
- 6:37able to, like, apply vectors on. Every single query is really,
- 6:42really challenging. Vectors are very, very
- 6:45computationally, heavy it. Take up a lot of memory both in
- 6:48the server but also when you're training and storing them and
- 6:50they're actually pretty slow and expensive to roll out.
- 6:53The Breakthrough that we've recently had is we figured out
- 6:57how to compress vectors, so we can actually compress Them into
- 7:00like a 10 times smaller format, keeping the same kind of
- 7:03relevancy and we can do it at extremely high speeds.
- 7:08We can take every single query across as one point seven five
- 7:11trillion and we can apply Vector search them.
- 7:14This is the big breakthrough. This Hashem technology that
- 7:16we've built means that for the real world production
- 7:19environments. When you need high speed, you
- 7:22need a reasonable cost and you need really big scale.
- 7:25So this is what makes it practical for you to.
- 7:29Use vectors in this, highly real-time shopping environment,
- 7:35every single query for our customers, we're able to use
- 7:39both factor and keyword in a single kind of API call.
- 7:43And by using both of these strategies, we actually get a
- 7:46lot more information back and we can use the extra information.
- 7:50Like how many keywords don't match but also how did the
- 7:52vectors match to do a much better ranking?
- 7:55And I think that's pretty unique.
- 7:56I no one else in the industry is able to do this at this size.
- 7:59Scale and speed that we're doing enough, because we're using this
- 8:03kind of technique called hashing.
- 8:04So, the technology is enabling a simpler easier user experience,
- 8:12that's at the same time, more effective.
- 8:14What we've seen is about 70% of all of the search queries are
- 8:18actually far more complicated one-off type queries, where
- 8:21people are asking for very specific things, they have a
- 8:23question and that kind of 70% of queers are going, very
- 8:27unanswered at the moment because they're very Uncle to answer
- 8:30with just matching words. So this kind of new AI Vector
- 8:33search is really monetizing and helping solve the customers
- 8:38needs in that 70% of a long tail case and it actually translates
- 8:42into a lot more sales. Sean we're does personalization
- 8:46fit into this landscape creating a loyal customers.
- 8:51One of the most important parts about creating a great business
- 8:54because you pay to acquire a customer the first time and if
- 8:57they either don't convert that first time because they Find
- 9:00something, or they end up, not coming back for a second or
- 9:03third shopping Journey. Then the economics of acquiring,
- 9:06the customer become pretty unattractive.
- 9:09How do I create an experience for a shopper where I recognize
- 9:12that there are a customer who's been to the store before?
- 9:15How do I make it feel? So that when a customer comes to
- 9:18my store, that they're going to get a superior experience, the
- 9:21second and third time versus going to a competitor store and
- 9:25that's all about personalization.
- 9:27Once I've seen what a customer prefers Purrs either through the
- 9:31searches that they make, on the website, or the products that
- 9:34they click on the products they buy, or the products that they
- 9:37click on a. Don't buy, you can observe the
- 9:39entire behavior of the customers, they're shopping and
- 9:42you can learn from it, and you can do that in two ways.
- 9:46The first way is when they actually appear for the first
- 9:49time, you can really hang on every single, click every single
- 9:54word that they type in and category that they look at.
- 9:57And in real time, even for a shopper that you've never Series
- 10:00never been to the store before. You can start to adapt.
- 10:03This is the called real-time personalization, and it's kind
- 10:06of like a cold start problem because you really don't know
- 10:08anything about the customer, but I can pretty quickly find out if
- 10:11their brands that they are clicking on more or if their
- 10:14price points that they're gravitating towards.
- 10:16And then I can start to show and put more of these types of
- 10:21products around brands or prices or categories in front of them
- 10:25as they click and discover around the site.
- 10:27The second really important thing is once they've actually
- 10:29Bought from you, the next time they come back, you acknowledge
- 10:32that. They are a customer that you've
- 10:33seen before and you're able to offer them again.
- 10:36Similar personalized experience and we've seen that these types
- 10:40of personalization algorithms and personalized, experiences
- 10:44can drive substantial increases in conversions.
- 10:47On the second and third visit, as well as this kind of
- 10:50real-time when they're shopping the first time.
- 10:52So I think it's a very important part of the whole customer
- 10:55lifecycle from acquisition first shopping.
- 10:59Third, and creating a long-term loyal customer.
- 11:02So you're talking about two kinds of personalization one.
- 11:06Is the cold start? As you describe when a visitor,
- 11:09First shows up at your site, but then remembering that visitor,
- 11:14when they come back the second time, the third time and
- 11:18creating an even more tailored and personalized customer
- 11:21experience for them. What is this?
- 11:23Do to the overall customer lifetime value?
- 11:29You A lot of people when they first do their online experience
- 11:33and move online, they start to focus really on that very first
- 11:37experience of my put some advertising out.
- 11:40It cost me X, those Shoppers turned up to the site and they
- 11:43bought why, you know, here's my return on investment.
- 11:46But as you start to build a more loyal following and you start to
- 11:49become a little bit more sophisticated in the way, you're
- 11:51thinking about your online business, you start to look at
- 11:55the return over a much longer Lifetime with a customer.
- 11:58You take the whole Time value. And then you set that against
- 12:01the acquisition cost and what it means is, if you create loyal
- 12:05customers, you can spend a higher dollar to acquire a
- 12:08customer. And so this is really like how
- 12:10to create a scalable customer, acquisition business model
- 12:14online is trying to get your cost per customer acquisition
- 12:18and your customer lifetime value.
- 12:19And the ratio of these two numbers, you want to get that
- 12:23very high. So high customer lifetime value
- 12:25low customer acquisition, how is this different from the
- 12:29traditional? Additional concept of customer
- 12:32Journeys and looking at the total lifetime value of a
- 12:36customer. Typically people will look in
- 12:39session. So the only look at the specific
- 12:42shopping session that the customer is in Manila.
- 12:45Tribute basically all of the value against the acquisition
- 12:49cost for the customer but as you start to create much more a, I
- 12:53powered experiences and personalization, you can start
- 12:56to account for a much larger amount of It's been created when
- 13:01the customer is acquired, over more and more sessions.
- 13:03So, the technology enables this broader Horizon and more
- 13:11accurate understanding of the customer Behavior.
- 13:15So, that in effect, you can amortize your initial customer
- 13:20acquisition cost over this longer broader time Horizon, a
- 13:25great e-commerce stores one where you spend, you know,
- 13:28potentially even more money acquiring customers because
- 13:31those customers are going to spend over a series of months
- 13:34and years with you and they're going to become loyal and
- 13:37generate long-term profitability on an individual customer
- 13:41perspective. So think of it like a PL for a
- 13:44specific customer, right where your profits are the long-term
- 13:48value that they're going to generate through loyalty?
- 13:51And the initial outlay the costs are how much you could spend to
- 13:55acquire them you've touched on this.
- 13:58But what is the impact? Act on the customer or the
- 14:02benefit that the customer receives to encourage them to
- 14:08come back and therefore provide that loyalty to the retailer
- 14:13from a Shoppers perspective. The benefit is really clear,
- 14:15which is that they're able to find like better products and
- 14:19they're able to do it in a much shorter time period.
- 14:21And it's as simple as like when I turn up to the site, do I even
- 14:24feel like this is a shock, that is kind of tailored towards my
- 14:28needs and to my Tastes so for Shoppers, they're able to find
- 14:32the things they want. But they also have a sense of
- 14:35feeling of being welcomed and a feeling that the place that
- 14:38they're shopping is the type of place that they want to continue
- 14:40doing business with, because it's it reflects their tastes
- 14:43their style, their values. I think the benefit of this type
- 14:47of Lifetime customer value. Analysis is obviously crucial
- 14:55for retailers, but it leads to the next question.
- 14:59In which is how do we implement this.
- 15:04You need to have a partner who can help guide you through the
- 15:06process, their companies like out go.
- 15:08Lea and other people who have solved these problems and who
- 15:13have really focused on making the integration and the data
- 15:17collection pieces that are needed to power, these AI
- 15:19algorithms, extremely simple with, a few lines of code,
- 15:22you're able to integrate with. Like I'll go Lea, unlock our
- 15:25search API working with vendors and working with folks who have
- 15:29dedicated Dated large amounts of resources, time and expertise to
- 15:33solving these problems tends to be the first starting point.
- 15:35And often the best solution, the long run.
- 15:37What are some of the best practices for implementing
- 15:41personalization? Across all of the touch points
- 15:45as customers interact with your site?
- 15:48The more touch points in the more data that you're able to
- 15:52capture about how your users are using a product, the better, all
- 15:56of these algorithms and reading need to be able to understand
- 16:00Views conversions, but also the product data customers are
- 16:03looking at and having something that can indicate which
- 16:07customers which between sessions.
- 16:10So if they can log into your website, that's the best, but if
- 16:13you can add cookies to your website so you can track them
- 16:15between visits. That's even that's good.
- 16:18As well. What about the buying funnel?
- 16:21We are customers are in their journey in their relationship to
- 16:25you as a retailer. Customers will turn up on the
- 16:28homepage for example and And you'll have very, very little
- 16:32intent about what they want in the shopping Journey that the
- 16:35kind of high level of the funnel where you're not sure what
- 16:38they're looking for. You want to showcase a breath to
- 16:41them of things that they can find on the store.
- 16:44And if there are returning customer, you want to have a
- 16:47strong personalization bias on this home page, so that they can
- 16:50see things that are at their taste or style or price points.
- 16:54But as they go and they searched and they start giving you more
- 16:58information about their intent. Really want to focus on the
- 17:01relevance, and on the accuracy of the results, you're showing
- 17:05them because they're giving you a very specific request.
- 17:09And then, once they land on a product detail page, for
- 17:12example, you have an even more specific piece of information
- 17:15they're interested in this product.
- 17:17So again, you want to become even more tightly focused on
- 17:20your recommendations, you maybe might not want to use as much
- 17:25personalization when you're on a product landing page.
- 17:28Because actually the The Shoppers told you specifically
- 17:32something that they're looking for.
- 17:33And so as a customer goes in through the Journey, you've got
- 17:36to think about how much intent you really have and either
- 17:39dial-up personalization. When you don't have as much
- 17:41intent or dial-up relevance accuracy.
- 17:45When you have high intent Sean what is the role of real-time
- 17:49data in all of this process. You got to think that a large
- 17:52number of customers coming to your website.
- 17:54It may be the first time that you've seen.
- 17:56It's their first experience. You want that first experience
- 17:59to be great. So as soon as they click on a
- 18:01single item or they type in a search request, you want that
- 18:04data to flow back into your systems and adapt, your
- 18:08algorithms and real-time you have to have systems that are
- 18:11extremely fast at both ingesting and getting the data in but also
- 18:15like retraining and adapting the algorithm.
- 18:18So I think real-time data is very important for that first
- 18:20experience but also what you sometimes find is customers come
- 18:24back and they shop slightly different way to their taste.
- 18:28So you also want to be able to adapt In real time, even for
- 18:30customers, that you know, a lot about, then you use this data to
- 18:34make the decision around what to show next to that user.
- 18:41And of course, it's happening. Essentially instantaneously from
- 18:44the user perspective. Instantaneous, we're talking
- 18:46about milliseconds here that matter in the shopping
- 18:49experience. The faster, the experience is,
- 18:52the more likely customers are going to continue.
- 18:55We've seen 100 200 milliseconds worth of delay.
- 18:58Actually reduces the conversion. And customers end up, leaving
- 19:01the website. So speed is really really
- 19:04important in the Commerce world are there.
- 19:06Common implementation challenges that retailers face when they're
- 19:11trying to implement this kind of personalization.
- 19:16One of the hard things is making sure that you are capturing this
- 19:19type of 121 data around. This is the user that's logged
- 19:23in. This is the cookie ID that they
- 19:26accepted, this is the session ID for the shopping session.
- 19:30Ian and making that information available across every click
- 19:33every product view in every single time they come to the
- 19:36site so that the algorithms can do their work, the algorithms
- 19:40can only do so much, they need to be able to identify users.
- 19:44So that's that's really one of the bigger areas where Ecommerce
- 19:48companies need to make sure they've implemented Google
- 19:50analytics pretty well or whatever their analytics
- 19:53packages. What's the solution to this very
- 19:57common problem using off-the-shelf analytics
- 19:59products? But really focusing on doing a
- 20:01comprehensive job and implementing them.
- 20:03Also, making sure that when you choose a vendor that you really
- 20:07take the extra time when you do the integration to send all the
- 20:10events, send all the clicks and conversions with the right
- 20:13fields and everything. How do you address the ethical
- 20:17or the privacy considerations that people think of, when we
- 20:21talk about this kind of hyper personalization, one of the
- 20:25things that we have to understand first is that a lot
- 20:29of customers Workers and a lot of Shoppers are very happy to
- 20:32make the trade-off of getting a better experience by allowing
- 20:36the merchant or the site that they're shopping on to be able
- 20:40to collect this type of data. And when you think about it,
- 20:43that way, you want to make a transparent to them what you're
- 20:46doing. And you want to allow them to
- 20:48make that trade-off for customers who don't feel
- 20:51comfortable with that. You want to enable them to have
- 20:53a very clear trade-off as well and provide a great default
- 20:56experience. I think it's just important
- 20:58about transparency But also telling customers with the value
- 21:02of it is so they know ahead of time when they click the kind of
- 21:05accept cookies button or decide to log in and create an account
- 21:08for example that by doing. So they're really going to get a
- 21:12far better experience throughout the journey Shawn.
- 21:15Let's talk about measurement and kpis what are the best
- 21:22mechanisms or approaches to measuring the results of these
- 21:26kinds of personalization efforts?
- 21:29We use a A/B Testing where we will make a change to the
- 21:32website, like, we will update our algorithm and we will send
- 21:36some of the customers down the, a channel, which is the old
- 21:38experience. And some customers down, the B
- 21:40Channel, which is the new experience.
- 21:42And using statistics, we can find out whether or not the
- 21:46customers who are getting the new product or new experience
- 21:49convert at a higher rate, or spend more money per purchase or
- 21:53have a higher revenue and total and we can figure out the
- 21:58statistical significance of So that when we actually decide to
- 22:01roll out the change, we're confident that it's actually
- 22:05going to improve the experience for customers.
- 22:08So it's exciting. It's a bit like scientific
- 22:10experimentation in this experimentation, how much is
- 22:14managed by the retail shop owner versus?
- 22:18How much is managed by algo Lea behind the scenes, we have A/B
- 22:23Testing built into everything that we do on the platform
- 22:26because we think it's just like one of the highest velocity ways
- 22:29of improving. Proving your experience.
- 22:32So it's definitely built into Al kolya, we can run manage all the
- 22:36A/B experiments for you, show you the data.
- 22:38So you feel confidence in making changes Are there specific
- 22:43metrics that reflect customer Revenue but also customer
- 22:47loyalty and customer lifetime value.
- 22:50There are a lot of really important metrics that you need
- 22:52to keep track of when you're running an online business and
- 22:56it really depends on kind of where in the buying funnel, I
- 22:59think you are. So, for example, when a customer
- 23:02first arrives in the site, it's important to look at the bounce
- 23:04rate. How many clicks do they get into
- 23:06the product before they? It will experience it.
- 23:09Whereas, if someone's made it all the way to a landing page,
- 23:11for example, you probably want to be more focused on the
- 23:13conversion rate and whether they're actually going to check
- 23:16out and buy something and then after they've actually bought
- 23:19something, you really want to be looking at that lifetime value
- 23:22number. How often do they come back to
- 23:23the website? Visit again, one of the things
- 23:25that I think is extremely important, particularly for
- 23:27search is the Mission it, which someone has clicked on a result
- 23:32when given a set of search results.
- 23:34If you have, let's say, 10 rows worth of search results, the
- 23:39results that appear on the very first row will always get
- 23:42clicked on a far higher rate than the ones that appear on the
- 23:44tenth row. So I really think that making
- 23:48sure that the products that you're getting into that first
- 23:50row. We're second row are extremely
- 23:53relevant. We see it makes a big
- 23:55difference. All of these very granular.
- 24:00Sets of reference data, become the building blocks, or the
- 24:05components of understanding the consumer, and then you can
- 24:11respond, you need to understand the products really well that
- 24:14you're selling. And the kind of options you have
- 24:17to present. Secondly, you have to understand
- 24:20the Shopper who turns up through personalization through profile
- 24:24building through, looking at all, their past interactions,
- 24:28and then thirdly, you need to understand and Real-time the
- 24:31things that they're telling you that they're looking for the
- 24:33intent. How do you manage to different
- 24:36kinds of Shoppers you have folks who show up and they leisurely
- 24:40want to browse your catalog and then you have other people who
- 24:44just, you know, I don't have any time and I just need to buy this
- 24:48product and if you don't have this product right now, I'm
- 24:52going away. And if you do have it, I'm going
- 24:54to buy. There are a couple of shopping
- 24:56modes. The first one which obviously
- 24:59you need to be. Incredibly good at capturing a
- 25:02high intent, Shopper is come to buy specific product and you
- 25:07need to do that with Incredible search results and really great
- 25:09ranking of those results. Then you have a second set of
- 25:12Shoppers who really enjoy the experience of coming to a site
- 25:16and being inspired and discovering this site has to
- 25:20offer. And these are Shoppers where you
- 25:22really want to create a browse experience which is far more
- 25:28interactive and far more. Or inspirational.
- 25:32So I'll tell you at the moment. Most sites have a browse
- 25:35experience which is something along the lines of, okay, a
- 25:38customer clicks on. The dress category, we have
- 25:4130,000 dresses, you know, give them 100 Pages worth of dresses
- 25:46and order that by the most popular and it's just like
- 25:50entirely overwhelming. And often the most popular
- 25:53products are some of the least inspirational Shawn, where is
- 25:58search going over the next few? Here's one of the big things is
- 26:01going to happen to search. We talked about how vectors and
- 26:04large language models are really transforming the way that we
- 26:07searched. We're not just matching keywords
- 26:09anymore. We're actually understanding
- 26:11customers understanding the concepts and coming up with much
- 26:14better matching algorithms, but I think one other major change
- 26:19is going to be the way that we have more conversational
- 26:24approach to shopping. So when you think about the
- 26:27offline shopping experience off, When people go into a physical
- 26:31retail store because they want to get some assistance and they
- 26:35want to talk to someone who has some expert knowledge that used
- 26:39to be an experience. That wasn't very good online,
- 26:42it's very static. It's like here just read the
- 26:44information. You figure it out for yourself
- 26:46but with the power of chap GPT and these large language models,
- 26:50we're going to be able to have much more expert, personal
- 26:53assistant like conversations at scale.
- 26:56Sean what should eCommerce retailers?
- 26:59Do now to take advantage of all of these capabilities you've
- 27:03been describing. So I think there are three main
- 27:05things that eCommerce operator should be thinking about the
- 27:08first is the vendors, they work with, they should be working
- 27:11with their existing vendor today to take advantage of some of the
- 27:15AI capabilities that are coming or to you know select or add
- 27:18vendors. Who have these AI capabilities.
- 27:22The second thing is that I think they really need to focus on
- 27:24their own analytics. So really capturing all the data
- 27:28correctly about their Customers whether that's clicks and
- 27:31conversion data. But also like making sure the
- 27:34capture the cookie IDs and who's logged in ETC.
- 27:37So that you can personalize the experience.
- 27:40And then thirdly, I think they should be thinking about the
- 27:43metrics that matter to them as a business.
- 27:45So they can understand how the AI algorithms are actually
- 27:49creating a more sustainable profitable and long-term loyal.
- 27:52Customer base. We've been talking with Sean
- 27:55Milani, Chief technology officer of algo, Leah, For more
- 28:00information, check out www.google.com.