Guidewire Keynote Chapter 2: Diego Devalle

Welcome. Oh, right there. Yeah, I got to talk. Sorry. As we, as you know, we are tech professionals, not really kind of actors and so on. So we kind of spent the last two weeks kind of jumping into this, where normally we do other things for a living. So you give us a little bit of leeway. Go back to this, right back to that. No, no, okay. All right. I'm glad to be here. First of all, thank you. It's a lot of rehearsal that went into this, as you can tell. And just to set some context, one of our main priorities in GWCP was to use best-in-class to address performance monitoring. And about two years ago, we selected Datadog and natively integrated it into InsuranceSuite.

How would you further investigate from here?

Oh yeah, I see the error message: 'Failed to retrieve fraud details.' And then we drill down even further on the message itself and see that the integration is timing out. If only all troubleshooting was that easy.

That's right. This is phenomenal. But now let's assume we have a fix for this issue and we want to deploy this into production. This typically will trigger many steps: promoting across multiple stages, database backup, and ultimately deploy. Until today, this would have been a ticket for the Guidewire ops team. As of today, we're making this available as a self-service. G, back to you.

Great. But now in a real case scenario, to test this code effectively in pre-prod, you will need to have some production data, I guess.

That's exactly right. We're going to skip the testing here, like any great developer, and go directly to the prod deploy.

Although, just a note, most of our internal teams prefer to use either a CLI or an orchestration framework rather than a graphical user interface, which is exactly why we're exposing all of this functionality via APIs. And again, all this capability is becoming available today.

It's true, it's true.

Hi Diego, I'm excited to be here today. Today marks the culmination of three years of effort, and we're going to show you the new InsuranceSuite update experience. This, as Mike mentioned, is going to be a little bit technical. For everyone in the audience, especially if you are not used to and familiar with code repositories, if you don't understand what's happening, ask your engineer or ask Ben later. I'm happy to help, and that's why I'm here. So let's get started. Let's pretend that we're a Future State InsuranceSuite developer. We want to update to the latest version of Claim Center. What we're going to do is we're going to start with the Claim Center configuration. As part of the new InsuranceSuite update process, you're going to find two branches in your repository that are delivered by Guidewire. They're going to be called 'guidewire-releases' and 'guidewire-update.' 'guidewire-releases' is the out-of-the-box product release. This has been automated since Aspect. You might recall from last year's keynote. The 'guidewire-update' branch is new, and it's even more, it's better than that. This is where we've done a lot of the magic behind updates and how we're changing how we deliver software to you. This branch contains the InsuranceSuite FL release, it's with your configuration pre-installed, pre-validated, and ready to deploy. And this is why we're saying today upgrades have become updates.

This looks almost too easy.

How do we actually ensure this will work? To better answer that question, let's look at this dashboard that we use internally within engineering to continuously validate and monitor updates. So now, this is what we call Snowcat. And so when a Guidewire developer makes a change to an InsuranceSuite application, we instantiate a process that ensures that every customer implementation continues to compile, build, deploy, and run. This includes automated testing that covers both technical and functional correctness. We've built a significant amount of validation to date, and we continue to add more. As you see here, we are validating Cortina to FL, Aspect to FL, and Elysian to FL. All those validations are now part of our release criteria.

We will be delivering InsuranceSuite FL to your repository within the next two weeks, and all of our other customers will follow shortly.

Thank you.

Welcome. Thanks Diego. I'm really excited to show you the full platform in action, starting with Guidewire Home, where all our demos started today. Let's jump into the Jutro App Admin. Jutro App Admin is the one-stop shop to create and manage your digital experiences. I'm going to create a new Jutro app. We are role-playing as Future State Insurance who are launching a direct-to-consumer quote app for personal auto. The first choice you see here is between micro-frontend and standalone app. Micro-frontend allows you to embed the same app in multiple websites. Here, I'm selecting the personal auto quote template, which contains the components, business packs, patterns, and floor plans to build this experience, all thanks to the Jutro design system. Jutro platform has a DNA of reusability and modularity. Micro-frontend and app templates enable those principles. So once we've hit create, what's happening behind the scenes is we are creating the infrastructure on GWCP to build, deploy, and operate this application. This includes the CI/CD pipelines to test and promote all the way to production, just like G showed us earlier.

So Future State started building their product by going into Marketplace and grabbing the GO Personal Auto product. They then brought this product into Advanced Product Designer, APD. APD is a cloud-native application allowing you to design and update product models. So here, we are using what I introduced before: we're using the API, we're using cloud-native extension. But furthermore, the externalization of the product model has been a critical stepping stone for the Guidewire Cloud transformation. The product model includes the coverages, terms, risks, and rules, which builds the application data models, generates the APIs, and enables digital experiences. Customers can start with the GO product or can start directly designing in APD. It's shaved off months of implementation time. So looking at the screen, it looks like the Jutro application build is not completely complete, so let's switch to the pre-baked app.

All right. So Connie's bringing up the quote application. This application is right from the personal auto template that we selected. It contains all the pages, styling, and layout that is provided in the template that we previously selected.

Yeah, similar to GO products, it's in the template, so it's like a jump start. You can modify and enrich every aspect of this app with Jutro.

So what about the insurance product? It's actually metadata coming from the APD app, and let's not forget this is natively wired on top of Policy Center.

Yeah, it's pre-baked in the template via auto-generated cloud APIs. As you can see, as we're hitting continue and next, we're submitting quote data into Policy Center. So going back to our agility team and product lifecycle, now let's say that Future State used Guidewire Explorer to monitor the performance of the product and they need to modify it. What's next?

So on the right in the APD app, we can quickly add the rental reimbursement coverage to attract more customers, provide some technical details, and hit save. And this is why the product externalization has been critical. This change will flow through the rest of the cloud platform.

Indeed, that's the value of APD as a source of truth for the insurance product. Now Connie will do a quick sync with Policy Center to get the rental coverage replicated into Policy Center. This is enhancing the Policy Center data model and regenerating the cloud APIs. Now going back to our quote Jutro app, so Connie is going to do a quick refresh. A developer would usually restart their app locally or redeploy it on GWCP. Once we've refreshed this application, we're going to go through this wizard again, select the same values, and hopefully we'll see that rental reimbursement coverage show up.

Ta-da! And wow, there it is. There it is.

Oh, I see you've improvised the description. It said 'I hope this works.' It did work.

Yeah, it's always a little bit stressful to do live demos at Connection.

So starting with FL, the APD coverage change didn't just update the Policy Center but also flowed out into the Jutro app. And this is why we believe the Jutro platform combined with InsuranceSuite will be a game-changer in the insurance product lifecycle. So now we can complete the flow and actually quote the policy. So here we've got a quote. In the real world, rating would have to be enhanced to include this new rental coverage that we added, but we're going to skip that for this demo. So let's jump back to the right screen and look at the quote. So the same policy that we have there, we have it in Policy Center now. But now, last step, let's navigate down and see if the coverage is correctly reflected in Policy Center. So we go into the submission that we just created for that policy. There was a vehicle created, and we can see the rental reimbursement coverage that was selected for the policy. This is great. So this is the power of a product-driven development experience. So now, a business analyst can introduce product model changes continuously while in development. They can be more iterative with APD and Jutro and more effectively react to market needs. With product changes after going live, insurers can rapidly build, deploy, and update digital experiences for any user. With FL, APD app is generally available, and the Jutro digital platform is in early access. To learn more, we launched a complete new developer.guidewire.com website. It contains a bunch of information to learn how to use all these new capabilities and the platform. So it includes one-click access to cloud APIs, integration framework guides, and the APD app guide. But since you are all here at Connection, you should take advantage of our breakout sessions and join Sid in some of his breakouts.

Yeah, I'm looking forward to seeing all of you at the breakout sessions. Thank you, Diego.

Hey Diego, I'm so excited to show you what we're all building. So from the same Jutro App Admin screen that Sid showed you earlier, this time we're going to pull out the auto physical damage FNOL template. This contains a predefined flow that can jump-start your claims digital experience. Build your micro-frontend, which you can fully customize.

Yeah, so a lot of our insurers, it costs them about two and a half million dollars to build out that personal auto digital FNOL. We want to make it faster, better, and more cost-effective. Basically, this was the perfect app use case, the perfect first app use case. Here we have pre-created the out-of-the-box micro-frontend. Let's launch it. So what you're seeing is the unstyled out-of-the-box flow. It contains the most common questions that an insurer would take in when taking in an auto claim. So we'll go through the first few questions. Since this was created as an embeddable micro-frontend, let's see the same app embedded in a full-service website that Future State has built out.

Yeah, that's a common use case. So imagine Future State wants to leverage a new distribution channel and it's called Future Mutual. But this time, let's say that Future Mutual is not built on Jutro, it's built on a completely different, non-Jutro application. So on the center screen, what you'll see here is a new micro-frontend which looks totally different, but functionally it is the same flow.

Yeah. To get this micro-frontend styled, Future Mutual would have worked with their designers to define the tokens for font, buttons, and background color to match that existing website. So now, continuing with the theme of agility and lifecycle, I know everybody at this point will ask: what if I want to add a new question to improve user experience or claim efficiency? And I guess you might have guessed the answer is going to be APD. We are working hard to expand this capability of the APD into claims. So we prepared a video of what's in development right now. You're looking at a familiar APD screen, but this time it's about claims. So you're seeing questions about the vehicle incident, and this is where Claim Center houses all the data about damage to a vehicle. We'll add a new question: 'Is your vehicle safe to drive?' We'll tag it for the FNOL template and we'll save it. So similar to what we showed before, we will sync this with Claim Center. And once again, just for clarity, this feature for APD we are working on, this is not part of FL. There is a lot of debate if we can push it to the G release or the one after, so stay tuned.

Yeah, so that's the power of micro-frontends, which are reusable building blocks across multiple experiences. So we have two digital experiences running on top of the same Claim Center instance.

Yeah, you can. So because the FNOL template comes with cloud APIs to connect to Claim Center. Can you show us it here? There we go. Okay, so on the right screen, Sid is bringing up Claim Center and he'll search for that claim number, and here's the claim that we just filed.

Yep, that's right, Diego. So the application team really loves building on the Jutro platform. The FNOL application template is available as early access in FL. To join and learn more, please join us at our breakout session or come visit me at the claims booth. Thanks a lot, Diego.

So before we jump into the demo, let me walk through the use case. As an insurance business analyst, I want to keep a pulse on a new product offering, monitor its performance. So I already set an alert based on the opportunity-to-bind rate for personal auto, and I want to receive an alert if it dips below 30%. So here it is, below 20%. I follow up on that 28% and I jump into the application. So now I'm in the Submission Conversion dashboard, one of the 16 out-of-the-box dashboards available in Guidewire Cloud. And as you can see, if I navigate back to home, I can see the remaining 16. But let's go back to the Submission dashboard where my entire scenario started. So as you can see, this is a good month. We have increased the number of submissions. We have like 687, it's going up. But at the same time, our binding rate is going down to 28%. So now in this scenario, let's imagine that you want to drill down and let's imagine you don't have the data structure available. Traditionally, what you would have done, you would have gone back to a technical analyst so they can dig into the data and get back to you with a few new views, which can take time. But now with the new Explorer, I can get that context myself dynamically without waiting. As an example, let's click search and start analyzing. The only thing that you need is knowledge of the data. That's why Tom is here. I can do the search by, for example, count of underwriting issue by new business flag. New business flag is like an indicator if this is a renewal or if this is a net new policy. And as Tom is typing, you could see that the Explorer canvas guides you toward the relevant insights, just like Google. So this first search doesn't really tell me much besides, as I said, two-thirds of those are new business. So let's remove new business flag and try underwriting issue description. And here we go. The coverage term outside of authority looks like is blocking many policy binds. So I like the very intuitive way that you can canvas the table. You can also add sorting instructions in the text field. As you can see now, I just entered the sorting instruction and the table is being recanvased based on that instruction. But now let's continue with the investigation and go one level deeper and further filter by underwriting issue type. You can see by right-clicking, again it knows the data, and boom, it looks like the issue is happening with collision deductible. So maybe we're missing some coverage term for our product definition. From here, the demo can be infinite. You can continue to drill down. But Explorer now offers machine learning guidance pre-baked in the system. A hotspot called Spot IQ. I can click on the underwriting issue type, choose the type of recommendation, and in the background, Explorer will automatically detect the main driver for me and provide me a few recommendations. Let's wait for that to happen. And as we said, it's a live demo, and here we go. As you can see, one of them indicated that one of my top-tier captive agents, in this case Armstrong Premier, is responsible for most of the submissions that have this deductible issue. So after discussing with Armstrong, I might have learned that if we offered a collision deductible of $250 or less, we could increase the bind rate. So at this point, we could go back to APD and make a product change, and Explorer will dynamically update the data so that you can quickly measure the impact of the change. So Sid, let's come back. I'm joking, don't worry. Let's roll a video of one of our customers that is using all this capability.

Today, real-time changes to policies. You really can't imagine the power of going from this used to take days to this takes seconds. In the past, we had some times where we were spending months doing upgrades, where now we did it in days. When you're spending months, you're not innovating, you're not gaining efficiency. Prior to working with APD and Guidewire, little projects would take me one to two years. Using APD, I was able to do five whole products. It was great. I had a change request that came in. I presented that to the Policy Center team and they came back with their estimate was about 20 hours. I took that to the team who doesn't use APD, they came back with 200 hours. Prior to Explorer, the process to get actionable insights out of our data would have taken at least a month. With Explorer and Data Studio, the dashboard takes minutes to create. Guidewire will allow us to implement new lines quicker, implement new ideas faster. So faster is better.