Hello everyone. My name is rajan shep and i lead product management, for ai and industry solutions here at google cloud. Thank you for joining us. This year has been really hard for all of us but it's been really inspiring, to see how people and organizations. Have been coming together. To help each other during these times. Now as businesses, start to refocus, on the future and things start to improve. We believe that many organizations. And industries, will be able to apply. Many of the lessons, that they learn during these times as they build their businesses. We're already working with many of you and are looking forward to working with all of you, across, industries, to reimagine. How we can successfully. Use ai, to generate, value for your business. And we're at a very exciting, point in time for ai. I've said in the past that over the next, decade. Every, organization, will be transformed, with ai. And today we're going to focus on the big question of how. From working with top enterprises, customers, around the world, we've learned that the how actually comes in two flavors. First. In solving, common problems, that businesses, face. These could be problems that are unique to your industry, that we can solve by integrating, into key workflows. Or they could be common, problems across industries, for example things like providing. Great customer, service, with your contact, centers or, keeping up with document heavy processes. And the second thing, is, by equipping, your teams to build. The best ai solutions, to solve your unique problems. And so today, i'm going to share about how ai is driving solutions, to common problems, and then i'll introduce one of our engineering, directors, ting liu. To discuss, how ai can solve your unique problems. And throughout, we're going to be sharing example, stories of how customers, are realizing, value today. So for those of you, who are hands-on, with the technology. We're going to show you some cool improvements, to our products. And ways that you can make an impact with your companies. For those of you who are trying to understand. How ai can better. Be right for you. We hope that we can spark some ideas, about, ways that you can be successful. So, before, we get into the how i want to talk a little bit about the why. Why should you be thinking about ai, and also why work with us on this.
So, Based on the trends we're seeing, it's important to build out those ai solutions, sooner, rather than later. The organizations. That embrace, ai, will have a competitive, advantage. And the window for being ahead of the pack rather than getting left behind, is closing. Mckinsey. Recently published a study saying that companies, that fully absorb ai, in their value-producing. Workflows, by 2025. Will dominate, the 2030, world economy. With a 120. Increase, in cash flow growth. Soon. Every organization. Will have an ai team just like we all, built web teams back in the 90s. And that makes the question of how we equip you even more importantly. So to truly, unlock. The value of ai, you first need to really focus on the business problem in not the technology. And it really helps to work with companies, that have already engaged with customers. To deploy, ai at scale. So they can help you overcome the roadblocks, that you're going to face in in the process. That's why we're working so hard on this. Google cloud ai can help you with those roadblocks. And we can partner with you to solve those identifying, business problems, there are really three reasons, you should be thinking about google here. The first is because of our leadership, both in ai research, through google research. And also in practical, applications, of ai, through google products. Large enterprises. Trust us to the advisors. To their most critical transformation. Projects. In addition to the innovations, we're making on the software level, we're actually innovating, all the way down to the hardware. For example, we just delivered, record-setting, performance, in six out of eight benchmarks. In the latest round of the ml perf benchmarking, released in july. The second reason. Is that we bring our heritage, and experience, of deploying. Ai, in production, across google. In products like google photos, gmail, and many more. To help enterprises. Realize, the value, in ai. Our teams and ecosystems, support, the largest, brands, in the world to infuse, ai, in production. To build new revenue streams and to drive operational, efficiency. And the third reason, is, trust from an enterprise. And the trust that they are able to have, with their employees, and customers, is a top priority, for us. We're driving, leadership, in responsible. Ai through ai principles. And recommended, governance, practices. Are things that we are practicing, internally, and we're also working with customers, on. Later king is going to share some examples, of how customers are using the platform, to identify, and resolve, complex, issues things like bias, in models during development, in the evaluation, phase, so as i mentioned. The first flavor, of how organizations, will be transformed, with ai is through solutions, to common problems. So let's get started with our horizontal. Solutions, and how business decision makers who care about how ai, can use them to gain a competitive, advantage. And optimize, their business operations. The first solution that i'm going to talk about is context center ai. We've taken our world-class. Conversational. Ai, speech, and natural language. And we've applied it to the contact center space, so that you can improve, customer, experience, and operational, efficiency, at the same time. You can try it for yourself with the multiple, cci. Demos, now available, in our next on their demo area. It first, easily plugs into your existing, workflows, and we have full integrations, with telephony, providers, like avaya. Cisco, genesis, twilio, and 5.9. So that you can take advantage, of your existing, investments, and relationships. Ccai. Also helps expedite, customer requests through virtual agents, it helps assist live agents. And it also, offers a layer of insights.
We're Continuing, to invest in ccii. And today we're launching, agent assist for chat, in addition to voice interactions. And we're also announcing, the ability to create a custom, voice that represents, your unique brand. Now dialog, flow, is the technology, that's the core of ccai. Today, i'm excited to announce dialog, flow cx. This new version of dialogflow. Offers customers, the best virtual, agent tool on the market, it's optimized, for enterprises, with large contact, centers, designed, to support complex, conversational. Architectures. And it's truly omnichannel. So you can build this once and deploy, anywhere. Both in your contact, centers as well as digital and social channels. Dialogflow. Cx, is truly state of the art and will make it possible to give your end customers, an improved, experience, with more intuitive, conversations. So one customer, that's really redefining, the possibilities, of ai powered conversation. Using contact center ai, is verizon. I had a chance to sit down, virtually, of course with said the vet, who's the executive, director, of customer, service technology, from verizon. And i'm excited to share the conversation, with you now. Thanks for being here today to share your experience. I, am looking forward to our conversation. So first of all can you tell me a little bit about why you were considering, this transition, of using ai in your contact centers. Happy to be here and thanks rajan. We at verizon, have consistently, focused on improving our customer, experience. And to that end, have been leading the charge on digital transformation. For a few years now. We consider this transition, as a way to improve, our customers, experience, but we also realized, that we had to find an efficient way to do that, there are two areas that we focused, on, the customer, facing. Experiences. On digital, chatbot, and ivr, and the agent-facing. Experience. With assistant, guidance capabilities. We were looking to cloud ai to deliver a more consistent, conversational, experience for our customers. And, better equip our agents with the tools and information, to support our customers. So what were some of the key factors. In your decision to select, google in particular, and contacts, in your ai. Yeah. We anchored on the two areas that i mentioned, the customer, facing digital experience.
And The agent experiences. On the customer, facing experiences. Our journey started several years ago. As we look to, enhance, the support for our customers, with the introduction, of a chatbot. As the bot evolved, we saw a huge opportunity. To take that experience, to the voice portal. And make it a great experience. And also for our agents, we saw agent assist, was able to help, our agents help the customers questions, we evaluated, several solutions, over a span of two to three years, and ultimately, picked google, for our contact center transformation, program. For a few key areas. An advanced, nlp. Synthetic, voice, that can help in the voice portal. And ability, to automatically, create, dialogues. And lastly. We also wanted, to make sure. That from a scalability. Standpoint. That the partner that we chose, is able to scale to our needs, and that's where, we thought google can come in and help us we've started the journey, and, we are seeing, good results. That's great and so what was the implementation. Like. We had to look at several factors. And we actually came up with a i would say a rather meticulous, plan around it. Given the breadth, of our customers, needs, we had to pick and choose the right use cases. And. On the technology, side of the house, we had to decide. Which, experiences. Have. Systems, already, tied to it like do we need new apis, to be. Created. Or if you can use existing, apis, so that was a big decision, complex decisions had to be made. When it came to, measurements. We had very stringent, analytic, requirements. And lastly. Business continued to evolve, which meant we had to continuously. Adjust the plan. It was almost like they say you know building the engine, while running it as well. That makes sense. So what kind of results have you seen so far, after implementing, ccai. We are in the process of rolling it out to a wider scale, but we've onboarded, a few use cases, already, and early results are promising. On the digital, bot. The auto dialog creation. Engine, is able to create, five times. The depth, of the conversation, so it's a lot more on human-like. And as a result of that, customers, are engaging with our bot, lot better. On the agent assist side, about 30 percent of the agents, have you started using the tool. And we are seeing almost 75, percent of those, agents giving us very high rating for the experience. I mentioned these are early days, but we're looking forward to more improvement. In. All our, uh, kpis. So now as we look to the future, what's your, overall, vision for customer experience. We want to be. Sure, that the customers, have the best possible experience.
We Can continue, to improve. And iterate, on how customers, and agents can reach the information. They need quickly. And easily. To that end we are exploring, additional, ways our assistant. Can enhance, the customer and agent interactions. And bring the convergence, of physical, and digital channels together. Ultimately, we want our customers to have the best experience, and what they call as ambient experiences. And that's our goal and that's what we're working on at verizon. Well thanks so much for your time today cyan i really appreciate the conversation. Thanks for the opportunity. Appreciate, it. It's so great to hear about the value, that verizon, is seeing. Earlier, this year we had a chance to use contact, center ai to help organizations. That are overwhelmed, with the influx, of demand to their contact centers. When we launched a rapid response, virtual agent we made it possible to quickly build and implement. A customized, contact center ai virtual agent to respond, to the common questions that people had due to covet 19. Over chat, voice and social channels, so to learn more about ccai. Check out the ccai. State of the union breakout. The next horizontal, solution i'm going to talk about is document, ai, document, ai makes it possible to unlock, insights, from your documents, with machine learning and use these to their fullest degree, by extracting, structured, data from these unstructured, documents. Customers, are able to make better business decisions, whether it's things like processing, invoices, more quickly and accurately. Knowing. Each of your customers. Or reducing, things like mortgage, processing, time and many many other use cases. One company that has reimagined, their document, strategy, by leveraging document, ai, is the industry-leading. Mortgage services, provider, mr cooper. They're challenged, to manage, domain-specific. Mortgage, document, types, and they've classified, over 100, million pages. They've trained, over 130. Critical, mortgage, document, labels and they've achieved over 92, percent, accuracy, in these classifications. Using doc ai, they can trust their content will be accurately, classified. Earlier, this year we used the power of dot ai. To help the response to kova 19.. We developed the ppp. Lending ai solution, to help lenders, accelerate. And automate, processing. Loan applications. For the paycheck, protection, program. The document, ai, ppp, parser, api. Made it possible, for lenders to help with the large intake, of volume, of ppp, loan requests.
By Using ai to extract, critical, information, from the loan documents, submitted, by applicants. We also released some exciting, updates for document ai this year, ranging from form parser, which is a new way to extract, text as well as spatial, structures, from documents. To the invoice, parser, which is a specialized, api, that's highly accurate. In processing, invoices. To learn more about document, ai, take a look at the document, ai overview, breakout, or try the unlocking, insights, from document ai demo. On the next on their site, and also the demo for document ai on our website. These horizontal, solutions, are solving problems that are common across industries. But sometimes. Needs can differ, and we need to consider, industry-specific. Solutions. As thomas, mentioned in this opening keynote. We're investing, in specialized, solutions, across, industries. And have worked with customs like macy's, fox, and cardinal health. In addition. We expect that the movement to e-commerce. Will continue to accelerate, with a consistent, omni-channel, experience. Moving, from a nice-to-have. To a must-have. Our customers, are struggling to optimize, their merchandising. And supply chain as well as the product discovery, to drive conversion. We're building out solutions, to address these needs including things like demand forecasting. Product search, and recommendations. Ai. In july, we announced that the recommendations, ai is now available, in public beta, and it's already being used by companies like sephora, and haynes. In healthcare. Cobit 19, has forced institutions, to rethink, how, they can provide, high. Patient-centered. Care, outside, of the traditional care settings. We've seen and we expect to continue, to see, a huge uptake. In the use of remote collaboration, and telemedicine. Tools. Covet 19, has further highlighted, the need for better infra interoperability. Of healthcare, data to enable remote, care, as well as the ability to plan, forward, how healthcare organizations. Need to respond. We formed a strategic, partnership, with mayo clinic, combining, our cloud and ai capabilities. With mayo's clinical expertise. To jointly develop solutions to transform, healthcare. In financial, services, we see a strong, need to streamline, processes. Particularly. Those with complex, document management, processes, such as lending for example. Today we're excited to announce two new solutions, designed, to make this easier for financial services, organizations. The first is lending doc ai. A solution, built specifically. To help financial, institutions. Process mortgages, more quickly and efficiently. It automates, many of the routine, document, reviews, so people can focus on the more complex, decisions. The second, is prepare to pay doc ai. Ai for procure to pay. Helps, enterprises, automate one of their highest volume and highest value, business processes, which is the procurement, cycle. We provide a group of ai powered parsers. Starting with invoices, and receipts. That take documents, in a variety of formats, and return, cleanly, structured, data. Another, area that companies like hsbc. Are starting to look at to transform, and modernize. Is their end, customers, experience. And also, helping better manage risk. And we're working with them on solutions, like aml.
Anti-money, Laundering and kyc, know your customer. In media and entertainment, and gaming, we see customers, with a large amount of data and especially things like video, and media, files, companies like fox sports, are starting to use, and make use of massive, amounts, of content. And with their video content they want to make it possible, to search, and also do interesting. Work with that content. In manufacturing. We expect to see more distributed, manufacturing. Lines, and to make it safer for workers. And also more efficient for organizations. Companies like kyocera. Are using our ai, in defect, detection, and we're creating offerings, like adaptive, controls, to drive energy efficiency, and optimization. In buildings and factories. Now that i've covered how ai, can solve common problems that your business is faced with, both problems, that we see across, verticals, and those that are unique to your industry. I'd like to pass it over to ting liu, one of our amazing, engineering, directors, to talk about equipping, your teams, to build the best ai solutions. And to solve your unique problems. Thanks rajan. As russian shared, we recognize, that, there are unique business, problems, that you need to solve. And we are invested, in bringing you the best of google air technology. So you can build solutions, to solve those problems, as well. Customers, tell us that their teams working on ai are composed, of various skill sets. From ml engineers, to data scientists, to developers. So we have built tools for everyone on your team. I will highlight a few ways that we enable these different groups to build out unique ai solutions. First ml engineers. Developing, ml into production, involves, more than just building the model, it involves data collection. Feature extraction. Resource, management. Model versioning. Monitoring. And much more. The set of operations, that require, to deploy, and manage production, models, is referred to as, ml ops. And it can easily become the bottleneck. For an organization. As they build more models. We enable organizations. To scale, and streamline, their ml development. In a repeatable, and reproducible. Manner, with our ai platform. Today. We are announcing, new capabilities. And services. Within our ai platform. That will simplify. Ml ops. Starting with ai platform, pipelines. We announced, a hosted, offering, for building, managing, ml, pipelines, on ai platform, earlier this year. We are pleased to now share that. A fully managed, service, for ml pipelines, will be available by october, this year. With the new managed, service. Users can build ml pipelines, using pfx. Pre-built, components. And templates. That significantly. Reduce, the effort required, to deploy, models. Today. We offer, a continuous, evaluation. Service, in our platform. That samples, prediction, input and output, from deployed. Ml models. And helps users, assign human reviewers, to provide. Ground choose labels. We are also pleased to announce, a continuous, monitoring, service. That will monitor the model performance. In production. To let you know if it's going stale, or if there are any outliers. Skills, or concept, drifts. This will simplify, the management, of models, at scale. And will be available, to our users, by the end of this year. The next announcement. Is the foundation. Of all these new services. Our new ml metadata, management, service. This lets ai teams track all the important, artifacts. And experiments. They run. Providing, a curated, ledger of actions, and detailed, model lineage. This will enable, users to determine, model provenance. For any model trained, on ai platform. For debugging. Audit, or collaboration. Our ml metadata, service, will be available, in preview, by the end of september. Our customers, have told us how important. Ml ops is to them. And we will continue, to invest in mlr's, capabilities. Such making it easy for customers, to organize. Share and serve ml features, at scale. Second. We are improving, the agility, of ai data scientists, and developers. Giving them one place to go with the ai platform. Today. We are excited, to announce that by the end of september. The ai platform, will include, automl. As an integrated, function, in the workflow. This combines, the best of no code and code based options, to build custom, ml models, faster. With high quality. To hear a deep dive on this, take a look at the creating, value with ai platform, breakout.
We Also offer some tools, designed, for data scientists, to give them the resources. To achieve a competitive, edge. One example, is notebooks. Which is now ga. And we run the largest, number of notebook, instances. In the world. Notebooks, now supports. Important, enterprise, security, features, like, vpc, service, controls. Access transparency. And cloud iam. And has a smart analytics, framework, to run spark jobs and manage hadoop clusters, on dataproc. We have also integrated, kaggle kernels, to our notebooks, so kaggle users, can now seamlessly. Move their notebook, instance, to ai platform. For more compute. And storage, options. Third, we continue to hear from customers, and analysts, that it can be a challenge to hire ai talent. Though we offer developers. With limited ai expertise. Great tools. Through our platform. Such as pre-trained, apis. And other mail technology. To enable an organization's. Existing, teams non-ai, experts. To build ai applications. With high quality, and realize, the value of ai. Both customers, and analysts, have highlighted, our leadership, in state-of-the-art. Models. And we are continuing, to invest, in these models. For instance. We recently, upgraded. Our ocr, model, which achieves, high quality recognition. As well as the automl. Image classification. Model. Customers, are doing amazing, things with these developer, tools. From making maintenance, of their wind turbines, safer, by automating, the evaluation. Of equipment. To translating. Highly domain specific, financial, analysis. As quickly, as the markets, move. Many organizations. Across regulated. Industries, and those seeking hybrid, and multi-cloud, solutions. Have expressed, interest in using our air technologies, on premises. Last week, we announced that, our organizations. Can now use entos. To enable, stp, capabilities. On premises. Bringing our air technology, to you, in a more flexible. Way. As organizations. Increasingly, serve global, users, it becomes, imperative.
To Overcome, language, barriers. And we at google have a long history, of empowering, users, across the world. Through our technologies. Thousands, of nations. Use our language, and speech services. To connect, meaningfully. With global users, at scale, with ease. Earlier this year, we announced. The media translation, api. Which organizations. Can use to translate, audio, from streaming, input, in real time to provide a seamless, experience, for users. One customer, that's already, started, using this api, is oneplus. A large chinese smartphone, manufacturer. Who is using it to provide, real-time, streaming translation. For video chat, to ensure, their customers, feel effortlessly. Connected. One company, that has been putting, all of this together. And using ai to build a more curated. Shopping experience. For their customers, is etsy. Their marketplace. Includes, more than 65. Million, seller-generated. Listings. They are using ai to build sophisticated. Workflows. To help buyers, find exactly. What they are searching for, and to deliver, enriched, recommendations. That better reflect, their buyers, unique styles, and tastes. Lastly. Across, our portfolio. We're committed, to our investment. In explainable, ai. And, have been releasing, things like feature attributions. On ai platform, prediction. Notebooks. And automl, tables. In november, last year. We launched. Explainable, ai and model cars. This april. We launched, a new image attribution, method, called x-ray. X-ray, is a new way of displaying. Attributions. That highlights, which regions, of an image, most impacted, the model. Instead of just individual, pixels. This means the air platform, explanations. Now, offer two methods, for getting attributions. On image models. Integrated. Gradients. And x-ray. We will continue, to invest, in explainable, ai, so that, you can trust the platform, and the solutions, you build. We're committed, to continuing, to build the best tools, for all our users, to solve your unique problems. Back to your writing. Thank you tay. You know i remember. 10 years ago, brainstorming. With some of my colleagues, about what computing, would be like in 2020. At the beginning of this next decade. And we used to think about, all kinds of crazy, futuristic, ideas. Like imagine, if a computer can answer questions, for you and help you find what you need, imagine. If a computer, can help you do things that are right now manual, and repetitive. Imagine, if a computer, can use, all the information, that you have, to help you make a great decision. At that time these all seemed like science fiction. Now. This is a reality. And we've seen this in action with many of the enterprises. We work with and that you've heard about today. But, this is only the beginning of the transformation. We're looking forward to helping, you get value out of ai today. We also want to partner with you to imagine, what your business looks like, next year, in five years and in 10 years. Every, organization, will be transformed. By ai, in the next 10 years. And the question that we all need to answer, is how. And how can we help.
2020-09-12