>>Welcome to this cube conversation. I'm Dave Nicholson and I am joined by Lisa O'Malley senior director of product management, Google cloud, specializing in industry solutions. Lisa, welcome. Welcome to the >>Cube. Thank you, David. Great to be here. >>So let's, let's dive right into it. What makes an industry solution and what makes for a poser of an industry solution? >>Um, I think industry solutions are really all about driving business outcomes, that individual industries and individual companies within those industries really, really care about. Um, you know, uh, an alternative might be to take a horizontal solution, whether it's a CRM or an ERP and slap some industry labels on it and pose it as an industry solution. We like to do the hard engineering work, which is really going and figuring out what are the key outcomes that industries care about and spending time understanding the root causes and helping them with a Google cloud platform and all of the security data and analytics and AI capabilities that we have helping them really deeply solve those problems, um, at a, at a level that makes a difference and transforms their industry. >>So can you give us an example of something that's engineered in for a specific industry? When someone tells me they're engineering something in a, I think of a, I think of my car seat and if you're going to engineer in comfort, you better provide some controls for adjustability for me. So how, how do you strike that balance between hard engineered and sort of the bespoke services and customization that are always going to be necessary? >>Yeah, so clearly we don't want to create bespoke solutions for individual customers. We like to take, you know, industry wide problems and think about them a different way. You know, one example might be retail search. Um, you've probably all gone to a retailer, uh, typed items into the search bar and had an unsuccessful result. Um, and then maybe you've gone to Google and Googled it there and come back with the item in that retailer as one of the results. So search within individual retailers, websites historically has not been great. However, we've delivered a solution that brings Google quality search to an individual retailers, catalog and website. Um, and what we see is that this really helps them with what we call search abandonment. So it's like $300 billion a year is lost every year to people having unsuccessful, uh, searches on various websites. And so by, by delivering our product discovery solution, which incorporates, uh, retail that really solves an industry-wide problem. >>So Google is considered to be at the forefront of artificial intelligence. AI ML gets tossed around a lot, um, GCP, Google cloud, it's the real deal. Uh, how does, how does AI factor into some of these industry specific solutions? >>Uh, great question. Um, and not all of them are based on AI, but clearly, you know, when we think about Google, you think about, you know, data analytics, our ability to manipulate data and to apply AI and ML to real-world problems. Um, I'll give you an example of where we're using some of our core AI technology. And so that would be a product like visual inspection, where on a manufacturing line, you want to be able to, um, identify defects very effectively. Um, existing systems require a ton of training data. Whereas our machine learning allows us to deliver very high quality, like 10 X reduction in defects, um, with about, you know, 300 times less training data. And so that's where we've applied both our vision technology and our machine learning capabilities, uh, to come up with a great solution that fundamentally changes how inspection is done on manufacturing lines. >>So visual inspection inspection is one category, uh, recommendations is also often cited as another example. Uh, do, do you have any specific, uh, customer examples either with names or without are fine? Um, where, where recommendations come into play and, and, and, and what are some of the, the shades of difference when you talk about, um, the kind of intelligence that goes into visual inspection versus recommendations? Okay. >>So, uh, recommendations, one customer that I can talk about is Ikea. Um, they have implemented recommendations for a number of months at this stage, and they've seen an increase in click-through rate of about 30%. Um, we measure about 400% increase in, um, you know, relevant recommendations and overall that's, that's delivered at a 2% increase in average order value. Um, and so that's just one example of how recommendations and recommendations technology obviously has been with Google for a long time, when you think across search and YouTube, and a lot of the capabilities that are core to Google. And so being able to apply that more broadly to an industry circumstances is really, really powerful, um, on the visual inspection side, uh, Foxconn deployed this technology within their phone manufacturing process and that, um, uh, increased the accuracy of their defect detection by about 10 X. >>So, so you touched on this a little bit already, but if someone is trying to evaluate the difference between a real industry solution or industry cloud versus something that's just slapping a label on top of another label of, you know, for something that's generic, um, what are the sort of litmus tests that they should apply? What are the things that you look for? What are the criteria you think are important? >>Yeah. Um, I think it's really important to, you know, to really dig down, to identify as this just a horizontal solution, or has a company done the real hard engineering work to solve the problem? The way I think about it is I ask several questions, you know, do we think that these products have been engineered from the ground up to solve a specific industry problem? Are they just selling, you know, horizontal capabilities like CRM or ERP, um, and putting an industry label on it, have they actually being built for real world companies, you know, can they demo it in a real world example? Um, how much of it is, you know, original code code or is it, you know, just a reference architecture, how much, um, must a customer pay or work to actually implement that solution? Does it work out of the box or is there, you know, a big implementation with lots of system integrator, uh, spend required? And then I think lastly, and maybe most importantly is, is the, is the pricing connected to the value that you're bringing and is that pricing transparent? Um, and is it easy to understand for the customer and where the value, uh, where the value lies based on the pricing? >>What does the process look like, uh, within Google, within Google cloud, when you're considering what to categorize as an industry, what level of granularity do you get down to? How, how do you, how do you figure out what makes sense? Is it a level of effort in terms of engineering? Is it total addressable market? I'd love to, I'd love to be a fly on the wall. In some of those conversations, you think of the obvious categories, like financial services, retail, um, but give us an idea of what those conversations look like when you're trying to determine what constitutes an industry. >>So I think there's what constitutes an industry. And then there's what constitutes an industry that we want to build a specific productized solution for in terms of what constitutes an industry. I mean, I think those are pretty established in the market. They are things like financial services, healthcare, retail, and then there may be sub-verticals within those industries. So within financial services, you might have banking, retail banking, and commercial banking. You might have payments, you might have capital markets. Um, and I think those are, are, are pretty well established. I think you would expect the similar, the types of conversation that we have around what's the total addressable market. What's the level of technical sophistication within those industries. And then what are the problems that they're really seeking to solve? And do we have solutions that we think can make a difference there? >>So example of AI applied in an industry solution is the area around documents. Uh, and, uh, so, uh, again, if you can give us an example of, of >>Document AI, uh, how it's brought to bear, how it's different from making recommendations and product searches, and maybe a customer example, if you have one. Yeah. So, um, document, uh, AI is really, really phenomenal space in that the, the efficiency gains and operational efficiency, essentially around taking paper out of the process or taking people who are reviewing paper out of the process, uh, the opportunity is immense. You know, when you think about mortgage applications and the hundreds and hundreds of pages that we all have to sign up for, whether it's a refinance or a new buy, um, and then some poor person within the institution has to go and review all of that documentation. Um, we can turn that into, you know, something really phenomenal by using the document AI technology that Google has developed over time and training it on mortgage documentation. For example, um, Mr. Cooper uses our document AI technology, and they were able to, uh, increase their efficiency by about 400%, uh, in terms of their mortgage application process. So that's pretty phenomenal, but, you know, documents, don't just show up in mortgages nor do they just show up in financial services. There are documents all over the healthcare system. There are documents all over a public sector system. Um, and we believe that there's immense opportunity to take, uh, to take much of that paper and that re manual review of paper out of the system. >>So Lisa what's next, what kinds of industry solutions are you're working on that you can share a glimpse into? We're not asking for secrets that we can't, they can't be shared here just to be clear, but what's on the what's on the horizon. >>So I think there's some exciting things happening in our retail environment. Um, you can imagine that, you know, in the post COVID world, retail is very different from where it has been. And so the ability to bring your online and offline business and your consumer journeys through that business, um, really together is, is going to be super important. And so we're working on a lot of things there around understanding a full 360 view of your customer and how we might help them through their shopping journeys. Um, on, in, uh, in healthcare, we have, um, some phenomenal products in the market like our healthcare data engine, which helps take, um, sources of data from the many silos that exist across our healthcare system and bring them into one longitudinal view of the data. And so you can imagine that there would be many, um, diagnostic and operational opportunities to use that data in a much more efficient way than there are than the, than it's being used today, >>Specifically in healthcare, has that, have we seen a pivot, um, uh, because of the pandemic? >>So I don't know that it was specific to the pandemic. I think that the, um, the healthcare industry is, uh, is undergoing a lot of change, uh, in general across the board. And so the, the realization is that with, um, uh, I think the, what the, what the pandemic has done is it has accelerated some existing trends around movement towards telehealth, um, movement towards dispersed, um, healthcare within communities, as opposed to big centers. Uh, and so, you know, we find then that the, the data becomes even more fragmented and becomes more siloed and lots of, um, companies are solving small pieces of the problem. And so what Google would like to be able to do is to bring all of that data together, harmonize it, understanding all of the regulatory and compliance issues and opportunities that there are within the healthcare area and enable not just Google to build solutions on top of this data, but also to enable partners, um, and, uh, and, and providers our, uh, our payers themselves to, uh, to build solutions on top of the data. >>Lisa, it looks like it's time to wrap up. Do you have any final thoughts on, especially a, you know, where, where does AI progress us in industry solutions moving forward? >>So, you know, I think that AI is a tool that we should use wisely. I think it's something that we should understand how we, you know, understand that the, um, uh, our customer's deep needs, um, their business. I would come and sit there hoping to drive and where careful application of AI and machine learning can really benefit everybody in transforming their industries, whether that's through increasing top line revenue, taking cost out of the system, or generally being more efficient. >>Fantastic. Lisa, thank you for joining us for this cube conversation from the cube until next time. This is Dave Nicholson. >>Thank you, David. It was a pleasure. >>Thank you, Lisa.