Hi, everyone. My name is Ashley Kramer. I’m the director of product marketing here at Uncork, and I wanna welcome you to today’s session. We’ll kick things off in just a minute, but please use the chat to say hi and let us know where you’re tuning in from. Feel free to drop in any questions in the chat during the presentation, and we’ll try to answer as many as we can during our q and a at the end with Rich Robinson, our chief evangelist. Looking forward to the session. Okay. We’re gonna begin now. Thanks so much. Welcome, everyone, and thanks for joining us. My name is Richard Robinson, chief evangelist at Uncork. Today marks a massive milestone for us as we unveil our latest release, which introduces Uncork AI. This launch is about leveraging AI not merely to accelerate development, but to ensure systems are built with clear architectural standards with full transparency, governance, and security. Uncorked’s mission has always been to help customers reliably and securely manage complexity while controlling the cost of building and operating resilient enterprise systems. In the AI era, that mission becomes even more essential, and Uncork AI is designed to do just that. So let’s dive into what that means for this release. We’re going to walk you through how we’ve rebuilt the Uncork platform from the ground up to be natively AI first. As part of that, we will take a close look at our new build agent and explore how our rearchitected component model, smart components, provides our users with new levels of flexible hyper customization and reusability while maintaining the governance and security that matters most to enterprises. With Uncork AI, you are not merely generating new code each time like with other tools. Instead, you are creating, configuring, and reusing smart components to create your unique application each time. And lastly, you’ll experience our enhanced IDE that comes complete with improved way finding, live UI previewing, and domain filtered logic views. We’ll be showcasing all of these features and more in our demos today. Whether you’re focused on innovating faster or ensuring strict enterprise governance, you’re in the right place. So let’s jump right into the highlights of this release. Most AI tools generate code fast and leave customers to manage the risk, complexity, security, and long term changes themselves. That’s where tech debt and chaos creep in. The Uncork AI Build Agent never generates unvetted code. Instead, it configures applications and customizable agents using proven, reliable components within the guardrails of an enterprise grade secure platform. This makes software on Uncork AI safe to scale, change, and operate long term. The agent unifies the development life cycle leveraging best practices, planning, building, testing, and documenting both new and existing apps, resulting in better performing solutions that are easy to understand and maintain. It starts by generating an explicit plan that you approve. Once signed off, the agent shifts from interpreting requirements to executing against a deterministic checklist, building exactly what you asked for, nothing more and nothing less. That’s how we’re helping you eliminate the cost and the risk of managing code sprawl, drift, and hallucinations that are common with AI generated code. Thanks to Uncor’s modular architecture, you request changes using natural language prompts, and the agent makes targeted updates only to the specific components involved rather than rebuilding the entire module, which saves tokens and prevents unintended outcomes. You can audit those changes using our native module history tool to compare versions side by side. Finally, the agent automatically builds transparent test suites covering both happy paths and edge cases, which you can run as is or augment with your own custom tests. The advances of AI have unlocked the true potential of component based architecture. Instead of working from a fixed set of components, the Uncork AI build agent enables our customers to create, modify, and reuse their own custom sets of components to meet their unique needs. Simply provide your requirements in the platform’s guardrail and reasoning engine helps ensure architectural best practices, security, and governance in the creation or reuse of component and functional combinations that make up your application. Our AI build agent gets to do its magic, unlocking a new level of speed and customization that wasn’t available before. And the Uncork AI platform ensures predictable, governed results from AI powered development. In terms of rollout, smart components are only available to use if you have AI turned on in the platform. For those of you using Uncork AI, the build agent will use smart components by default for all new application builds and optimizations. Smart components are also interoperable with Centauri components. In this release, we enabled smart components to cover the essential capabilities needed for all enterprise apps, AgenTic, UI, Logic, and Query, and we have plans to expand this functionality in future releases. Let’s dive into these a little bit deeper. AgenTic components allow embedding native AI capabilities out of the box. Whether you’re using our prebuilt components for standard use cases or integrating your own custom Agentsic frameworks. UI components deliver modern dynamic presentation layers for your end users. Logic components execute complex business rules, calculations, and data transformations in a single step. Inquiry components focus on executing high performance, direct data retrieval, and real time analysis. Beyond creating individual apps faster, smart components change how you scale development across your organization. Users can create new highly custom smart components to adhere to organizational standards and design guidelines for others to reuse an application builds, ensuring token use optimization and organizational consistency across your application portfolio. That level of scalability becomes even more powerful when paired with the build agent. When you prompt the agent to add a new feature or create a new app, it doesn’t start coding from scratch. It scans the library first and reuses or copies and modifies existing smart components where it makes sense. This means you are automatically enforcing architectural standardization across your application portfolio, eliminating duplicated logic, conflicting queries, and fragmented design systems. Finally, the smart component library unlocks single point maintenance. Instead of managing component sprawl and code drift, you make changes once, and they automatically propagate across your entire application portfolio. With these capabilities, your component library stops being a fixed catalog you work around and becomes one you shape to your own business. To bring all these capabilities together, Uncork AI provides an enhanced modernized IDE, which delivers full feature parity with the previous Uncork designer, but now in a single unified workspace. This intuitively simplifies way finding and sets the standard for how you interact with the platform moving forward. Tab navigation and deep enterprise search let you effortlessly navigate logic, data, and architecture across your entire application portfolio. Alright. I think that’s enough slides for now. Let’s see some of these new capabilities in action. Let’s jump in and see how easy it is to build a new app with Uncork AI and to understand exactly what you’re building without having to write or review a single line of code. We are going to build a customer onboarding application with identity verification, risk screening, and manager approval, and ensure security and compliance at every step. In our workspace, we can give the Unquirk AI build agent a business requirements document with all the specs for the onboarding app. It’s a very comprehensive doc, and it’s got a lot of requirements. We can see the agent processing the requirements in the chat, as well as suggesting best practices, such as building a portion of the application in case we wanna make changes that could affect our overall build. Modifications to our base application could affect our dashboards, so it might be prudent to start with the first portion of our app. However, in this case, we’re confident in our requirements, so we’re gonna push the agent to build everything. We could also enter our own response. Before building the app, the agent creates an organized, well crafted build plan for review so we can clearly see exactly what the agent intends to build. Let’s break down each of the sections of this plan. The analysis summarizes the approach, the current state, and desired behaviors. Next, we can review the user journey, including an outline of all of the paths users will take. Now we’ll start to see how the build agent plans to turn our requirements into an application. Uncork AI differentiates itself here by taking a deterministic approach from this point forward, minimizing hallucinations, sprawl, and drift that can occur with probabilistic builds. With our component based model, we can commit to building exactly what is detailed here, nothing more, nothing less. We start with the data models that will be created, including written by and read by columns, providing full transparency into potential impacts before execution. Then we introduce smart components, Uncorq’s new AI first component model that optimizes runtime performance, UI flexibility, and memory consumption, eliminating workarounds that help enforce architectural standardization across your portfolio. To put it briefly, they unleash the speed of Adjentic AI while ensuring the enterprise grade governance companies need. The Uncork AI build agent will create the solution by assembling only secure, proven smart components, never unvetted code. In addition, the build agent will identify where shared smart components can be reused rather than creating new components unnecessarily. It can either reuse shared components or use them as templates to copy and modify where configurations slightly differ. Otherwise, it will create new ones that could be reused later by other applications if you decide to make them reusable. Under modules, we can see how the agent has translated our requirements into a blueprint of smart components within each module. Next, we have tests where the UnQRC AI agent instantly generates comprehensive test plans. We can also seamlessly refine edge cases. Then the agent will implement these tests without affecting functionality or data. The agent ensures standards and consistency in testing, which has historically been a major pain point for large enterprises. And finally, we can view the execution plan, which summarizes everything that the agent is about to do. It all looks good. So with a click, the agent starts building. Meanwhile, we can follow the execution plan, all of the modules, data models, integrations, logic between screens, access controls, and tests, which are built together. Again, Uncork AI is directly mapping the plan we approved to smart components, ensuring a secure, compliant, and optimal application that also aligns with our expectations, something that differentiates Uncork AI from other cogeneration tools. When the agent is done, we have an app that can be deployed within the same platform, not a prototype that we have to harden in the next sprint. Meanwhile, we can preview the app right in Uncork AI, experiencing it exactly as customers would, moving through its screen, submitting records, and watching the data flow. So in a matter of minutes, we have an application that is production ready, secure, compliant, easy for us to understand, and aligned with what the business requested. Of course, software development doesn’t end at launch. Regulations and customer expectations are constantly changing, and to stay agile, the business will always need changes to solutions over time. So now let’s see how Uncork AI empowers us to quickly update the application. With most AI tools, we’d have to review code and figure out how the app works before making the change. Uncork AI empowers us to make the change without having to be an expert in all of the implementation details. We can quickly navigate to the dashboard we need to update. Uncork AI offers us two different visualizations. In the UI view, every element on the page is interactive. We could select any of these elements to see how they connect to the app’s underlying logic, but we could also ask the build agent to explain the element in a way that’s easily understood. And logic view shows us exactly how the app functions without having to review a single line of code. This means we can quickly find and drill into the exact context as needed, which will make it easier to craft an effective prompt and make a change. Starting in UI mode, we can click the approval rate KPI, the element we know we need to change. Note that it’s currently 32%. Since we first built this app, the requirements have changed. The denominator should reflect only resolved cases, not all cases in progress. So let’s ask the agent to fix this. Before making the change, the build agent outlines every impacted dependency, then brings us back to the plan and waits for our approval before making the change. This simplifies the ability to evolve and update your application without needing to extensively review the code base. We can accept the changes in the application updates, going through the execution steps and the same flow that we saw before. Then we can go back into the module and see that the KPI has successfully updated based on the new calculation. But we are not done. We have been told to add the experience in this application. It’s something everybody wants and every boss is demanding by the end of the quarter. So we are going to add a conversational agent right into the app. Let’s ask the agent to add this as well. Again, we get a plan that provides a full understanding of what the agent is going to build so we can confidently accept the change. Uncork AI starts making the change right away. There are no integration headaches, no separate platform to stand up. When we navigate back to the module, we can see the finished product, and we can even test the agent. For example, if we ask who has the most open cases, now we know who we need to follow-up with. But, also, we got some insight into some unassigned cases that we may need to remediate. The build agent goes beyond creating and changing applications. It can find potential issues and suggest fixes, analyze legacy applications for performance improvements, and create documentation for applications all in one unified experience. Best of all, these visuals become the solution. We can understand the app and its logic at every level, quickly making changes and integrating powerful features easily just like a principal architect, but without needing to understand code. At the end of that demo, we use one class of smart components called Agenci components, allowing you to embed fully customizable Agenci experiences directly inside your applications while powering them with your preferred enterprise LLM. At launch, we’re introducing two powerful agents with more on the way. First is the conversational agent. This isn’t just a basic chatbot. It’s fully context aware. It can answer questions about live submission data, explain why a user was declined, run what if scenarios, or clarify business logic. Users can even upload documents like pay stubs, and the agent can automatically prefill on screen form fields. To keep human control front and center, the user always does the final submission. Plus, users can download complete chat logs for Audit. Second is the summarizer agent. Leverage this agent to synthesize complex information from uploaded documents, web links, on screen data, and database query results into clear actionable summaries. Both agents give you a turnkey way to deliver secure intelligent experiences right out of the box. In addition, UI smart components modernize the way you render and maintain enterprise app front end experiences. Architecturally, UI smart components remove historical layout limits and allow for more performant, dynamic experiences right out of the box compared to previous versions while still maintaining the governance, security, and architectural standards you require. Fully integrated with the Uncore AI build agent, enterprise teams can rapidly translate design artifacts and mock ups into on brand, accessible, and responsive applications, ensuring brand compliance without custom front end engineering overhead. Next up are logic smart components, which simplify how business rules are built while expanding the logic complexity Unquote AI supports. Historically, you had to jump between three different tools, decisions, data workflows, and calculators, stitching them together just to process a single piece of business logic. Smart components consolidate all three of these into one unified experience. Instead of jumping through fragmented modules, you get a clean, imperative, chain like flow that gives you human readable explanations for every single step. Moving to the data layer, query smart components change how the platform handles data retrieval. Previously, fetching data meant configuring legacy read plug ins in dealing with standard API overhead. Query smart components consolidate that entire workflow into a direct high performance database querying engine that reduces the configuration complexity of retrieving data. This allows you to build expressive, highly complex queries across a massive datasets, providing deep data analysis capabilities without needing to understand SQL or write a line of code. Let’s dive into how this new component model can improve the performance of your apps. We’ve seen how Uncork AI turns requirements into secure, governed, full stack applications. Now let’s see how it does it all without sacrificing speed or performance. Let’s look at a situation where we need to identify where cases are failing regulatory checks within the onboarding pipeline and which intake channels are producing the most compliance exceptions. This is a large dataset. It’s upwards of around 50,000 records. With smart components, we can process all of these records in less than a tenth of a second without any special effort or deep knowledge of the platform. However, we know that serving up 50,000 records isn’t enough. The team needs to be able to interact with it efficiently. So as part of our previous requirements, we asked Uncork AI to build a view that enables managers to interact with the data by searching, filtering, and drilling down into multiple levels of nesting and aggregation. They also need to analyze that data from a different angle, or to put it another way, they need to pivot to exception analysis. That’s no problem. So smart components are performing, but will they hold up under pressure? Uncork AI can process half a million records in under half a second and scale to millions. The UI renders, filters, searches, and updates at fast responsive speeds no matter how large the dataset. In other words, with Uncork AI, enterprise guarantees remain intact even as performance continues to scale. So what does all of this actually mean for your overall development life cycle? We have always been focused on delivering an integrated development life cycle for efficient enterprise application management. Uncork AI still provides the ability to launch apps to production faster with integrated deployment and hosting in the effortless platform management we are known for. This release really focused on enhancing the build experience for clients, providing them with an AI native tool that they could trust to deliver enterprise grade modern applications for their customers and users. Uncorked AI does not stop at raw code output, leaving you to manage infrastructure, firewalls, and deployment dead on your own. Uncorked AI carries your application all the way into production and then manages it for you over its life cycle, handling the necessary evils of enterprise ops, DDoS protection, encryption, backups, and scaling, so your team can focus on what actually differentiates your business. If you don’t have Uncork AI turned on today, please reach out to your customer advocacy manager to discuss in more detail. Thank you, and now on to the q and a. Hi, everyone. So thanks for joining us for this webinar. Hopefully, that was very insightful and informative. Happy to take any questions that you have. Feel free to add them into the chat as we sort of go through this. I see that we have two that were sent, I know, prior to this. So, Robert, thanks for for sending those. And so one of those questions is, do the agents, that I have have have access to my data? And the agent and the answer is technically yes, but they’re attached to your LLMs. Right? We don’t want you to have to necessarily get uncorked LLMs to have access to your data and go through that approval process. A lot of firms already have a strategy around which AI can read their data, whether it’s nonpublic material information, PII, or other sensitive information. And so this ensures that you can still be aligned with that motion that you already have, whether it’s direct with the LLM firm, whether it’s to, like, the enterprise or, Azure Foundry, and ensure that the data from your production system only ends up with your approved LLMs not going through Uncork. And then we got another question here, Rich, of what impact will it have on existing modules? On existing modules, the, what the LLM is or sorry. What the ability is designed to do is understand everything about Uncork. So both modules you have built five, six, seven years ago and, modules that you decided to build today. And so you can either ask it to continue to build those modules using those components, or you can ask it to actually convert a lot of that logic into smart components. A lot of our early customers who were testing this found that they were able to see massive performance improvements because using that new smart component, architecture, they are able to take a lot of the existing modules they have and utilize smart components not only for speed and and basically get improved load times, but they’re also able to do a lot more than they could before because of the complexity smart components offers from a resiliency and just a capability standpoint overall. And then another question we have here is, does the QuerySmart component also integrate with external APIs, or do the developers still have to follow the existing process to integrate them? So you would still follow the existing process to to integrate them. Right? So external APIs are always gonna be sort of way to pull data in from a a third party source, whether that’s directly from that API or a middleware. But the QuerySmart component is made for the internal data that’s stored in Uncork. So what you would do in those situations is do an external API call, a query call, and then merge that data together. And that’s one of the things smart components can actually do very well is be able to pull that data, and basically do an ETL process very quickly, much quicker than our old data workflow, so it’s a much more seamless integration. And then another question we have here is existing will existing components, be compatible in nine dot o? Absolutely. So the nine dot o really introduces a lot of new features and really centers our our whole platform over the Centauri, build. So new smart components are not a separate engine. They’re gonna be built alongside Centauri components, which is 99% of our applications. And so you can use smart components today alongside your existing applications. So combining this question with the one about existing modules, smart components and your normal components can live together in harmony and operate in harmony without doing anything special. Perfect. And then let’s see. Can you talk a little bit about how Uncork AI differs from other coding agents out there on the market? Absolutely. So Uncork AI’s goal is to get your app to production. Right? Our goal is not to just get you to, a POC or to some indicative code base, that that you have to then deploy yourself. Right? And that’s what a lot of AI coding tools would do, in those situations. Uncorked AI not only is focused on that componentry model to make sure that we’re not manifesting embedded code. We’re basically making sure that code is always up to date, secure, and patched, but we’re making sure it’s production ready and and actually has a production instance to go to. And so Uncorq’s core business is running your applications, and the build agent is that necessary piece to maintain that AI speed that everyone’s demanding. Perfect. Thank you, Rich. One more question here. We have, can uncork AI generate the entire application while referencing, modules in it? Absolutely. So we’ve done tests with lots of our different clients where they have very specific architecture, the the guidelines. Maybe they’re not best practices. Maybe the you know, Uncork AI could turn them into best practices. But what it can certainly do is reference modules either for the information within it or the architecture of which it’s been built out, and you can ask it to mimic that architecture or to make sure that there’s compatibility with those existing modules. So whether it’s going to create net new ones that work with your existing or actually change everything across the board to make it sort of one cohesive solution, you existing modules is a key part to being able to make changes, expand your application, and, obviously, also update some of the components that you’re on today into that new smart component architecture. Perfect. Then I’m seeing a few questions here around regarding performance. Do we need to replace our existing components? Or there was a question here about leveraging an an accelerator that was built. How does that all work together with Uncork AI? And then replacing existing components to smart components to achieve performance gains? Yep. So so smart components are sort of those new class of components that are both the conduit to increase complexity as well as performance gains. Right? And so for anyone with existing applications, nine point o isn’t just gonna make applications faster by doing nothing. Right? Legacy components are legacy components for a reason. Technology has obviously advanced. And so the the way to get the the most benefit performance wise is to use those smart components, and the Uncork AI build agent can get you there, as as far as that’s concerned. There’s no, you know, pressing piece right now to to make that conversion. We’re not sunsetting, like, the old components anytime soon. I know in the future, we plan to have more of a migration path to those. So for those of you who are worried about that, it might be worth waiting till our nine two release to look at the migration piece. But if there are any specific pages or sections of your application that could really benefit from performance improvements, we’ve seen pages have a 30 x improvement in performance just because there’s so much stuff that’s loading on the screen. Between the UI components, the logic, and the query components themselves. They will they were able to take twelve second page loads down to a half a second. And so those are some of the, the the things you may wanna test out, with the version today just to see if those, you know, easy wins are available while we, you know, have that migration strategy of, you know, further down the road. I’m on mute. And to piggyback off on that, we are also hosting a series of deep dive webinars and sessions with our, technical advisory and sales engineering team coming soon. And what you can sign up for those on the community now, but we’ll cover how to and where to make precision, updates around performance in those sessions. So we’ll we’ll provide more guidance on that, specifically as well. Yeah. To to and and to call out some of the the other questions I saw that we have here real quick. So there’s no reason to go back to classic view. The new IDE is full parity. It can run as a seed debugger. You can check all of the underlying JSON in the background if you want to. So, no necessary reason to go back and forth between those. And Uncorked AI is built for applications, not just modules. And so for those of you who may have been part of our beta program, I saw some of those questions come through. It is building full applications. It is not building just modules. And then, Rich, do we have any, constraints or recommended limits on module size to ensure, optimal performance? And is there a maximum module size we should adhere to, again, around performance? We we don’t necessarily have a maximum module size per se. Right? And I think smart components are you’re gonna realize that smart components are gonna be bigger components and smaller modules, so to speak. But, you know, what we we will say is that a browser can only handle so much. Right? You are limited to what your end user is able to load on page. And so just from, you know, any application standpoint, you don’t wanna over inundate a a customer where everything’s loading to the browser at once. You may wanna break that up. And Uncork AI is built to architecturally figure out where those breaks should be. Another question here around, the ability to integrate with, LLMs. Can we specifically integrate with OpenAI, Azure, OpenAI, Anthropic, or custom LLMs? Yes. So there there’s two main ways to do this. We have this idea of bring your own key, and that’s for the most popular ones. Right? And so Gemini, OpenAI, Anthropic, and then the cloud counterparts of those. So Bedrock, the old Vertex, now Gemini Enterprise, and Azure Foundry. Those are available out of the box. If you’re doing anything custom beyond that, you can always use our plug in component to API into those agents. So you’re not locked into only our bring your own key compatible ones. If you want, you can build a completely custom solution. If you’ve built, say, a full agent in some, you know, cloud service or some service that we’ve never heard of, it’s effectively a REST API integration into our system. Perfect. And, there’s a question here that I can answer, which is what training or certifications are available for existing teams. And we just launched a new academy experience. If you haven’t checked it out, I would encourage you to do so. And there’s a few new courses in there for everyone to get certified on the new Uncork AI platform. So would highly recommend going through that certification and and training process. Hey. One more question for you here, Rich. Uncork AI, I believe, will have the same SSE debugger feature that we have currently in, the previous versions of Unquirk? Yes. So SSE, the debugger will is also in the new IDE. Perfect. Let’s see. Can the AI build agent take an existing JSON schema or data model and automatically generate the corresponding uncork form components and validation logic? Yeah. Absolutely. So we’ve actually tested the build agent with tons of different artifacts, whether they’re requirements documents, existing code. I’ve thrown COBOL into the system and had it jump come up with a form, screenshots, of either Figma’s or, you know, inspiration websites that you’ve seen online and say, hey. I wanna mimic the the look and feel of the the site. Even React, I’ve given it React as well to say, hey. Can you mimic some of this React functionality? I’ve even built Tetris in the system, using some prompting. And so lots of different ways that you can use the build agent and the artifacts for it to understand, exactly what you’re providing and what your end goal is. It’s built on frontier models. Right? And so it knows a lot about the world. It knows about JSON and Node. Js and Python and Cobalt. You know? So so you can feed a lot of these different artifacts, and and see what you get out of it, and you’ll be pleasantly surprised. I certainly have been in the, you know, now eighteen months of testing this back from when I created the first iteration of our AI builder, and what has become is, you know, a thousand x, a million x more than what I ever imagined. Perfect. And then, last question here we have, how is RBAC applied to the the AI agent? I believe that question is asking about our own build agent. Yes. And so role based access control is based on the user, not necessarily the build agent. So for example, the build agent will not go randomly unlock modules that that user who’s using it has permission to. At the same time, if I’m read only, the build agent could document an application for me and help me debug, but it’s not going to make changes because I’m read only into that particular workspace or in general. And so the build agent is going to effectively piggyback off of whatever the user’s RBAC is, and they can’t override it. So for example, if I need to make a change and I have permission to, I need to manually unlock those modules to give it explicit permission. The AI agent is not gonna go over my head or try to circumvent our RBAC, to make changes in the background. Perfect. K. I don’t think I see any other questions coming through. There was one. There’s one about UIF, which I can answer. So, yes, UIF is a framework that some of you may have from from that, from the UIF offering. That that is compatible with UIF. Right? And so it will read UIF as if it was an existing build. So everyone has changed UIF in many different ways. And so we didn’t build it for the base UIF. We’re building it to look at your current version of how you’ve made the insurance framework your own, and it’s going to either mimic or suggest changes of improvement, to that framework. Perfect. Okay. So I think with that, we will conclude the webinar here. Of course, if you have more questions or things pop up, please reach out to your, sales rep or customer success rep as well as on the community. And, we like I said, this is the first of many webinars coming. We will be starting a series of ask the experts, coming in the next few weeks. So please join those to dive deeper on each of these concepts, and, we look forward to hosting you again soon. Thanks so much for coming today.
