Whether your business is in banking, insurance, the public sector, manufacturing, or another sector, as a leader, you’re feeling the pressure to move faster with AI while you keep operations running. To that end, you have a legacy system that everyone depends on. It carries decades-old business rules, logic, and dependencies, while the people who understood it are long gone. Maintaining these systems eats the time and money you’d rather spend on what’s next. Plus, they are almost impossible for AI agents to know and understand, because they have no way to see inside it.
Meanwhile, your teams are using AI coding tools, hoping for fast results, but each one works from a different picture of your systems. Without shared context, standards, and oversight, every agent builds in its own way. The result is more rework, inconsistency, and technical debt. Enterprises need one approach that brings AI coding tools, enterprise context, and governance together before fragmentation sets in.
Today, at OutSystems World Tour in Las Vegas, Nevada, we announced the general availability (GA) of Agent Experience, which opens the OutSystems platform and Enterprise Context Graph to any coding agent. This includes Claude Code, Cursor, Codex, and Kiro.
Before Agent Experience, using AI on legacy systems was a challenge. Now, AI agents can have live, deep visibility into enterprise systems, data, and logic so businesses can overcome the challenges of using AI for legacy systems modernization. And our customers and their development teams can cut rework, reduce token spend, and reach production faster under one governed platform.
Let’s take a look at what the availability of Agent Experience means for OutSystems customers, along with OutSystems Agent Workbench, Agent Spotter, and Agentic Industry Solutions, which were also announced today.
One of the biggest reasons modernization projects stall is because teams don't fully understand what the legacy system does. It may have been around even longer than they have. With Agent Experience, you have what you need to find the business rules in your legacy code, identify what you need to keep, and discard the rest.
Once you’ve established what matters, your team can start designing with any coding agent. Agent Experience connects it to OutSystems, and Mentor brings it to life with enterprise controls built in. Throughout these processes, the Enterprise Context Graph grows with you. Your agents get more context for the next change, and you preserve the business knowledge you’ve built over decades while giving your team a faster way forward. From modernizing legacy systems like COBOL, Lotus Notes, and AS/400 to deploying new agentic systems, you can accelerate delivery without rework or risk.
Let’s dive a little deeper into what using an agent of your choice looks like.
Recent developer survey data shows that 74% of AI-generated code requires significant post-generation rework or is never deployed to production. Agent Experience eliminates this bottleneck for regulated industries by enforcing an architectural blueprint that radically accelerates release timelines. AI leaders can take advantage of the speed of coding agents while keeping full control.
OutSystems Agent Experience gives regulated enterprises open coding choice without technical debt, security vulnerabilities, or unbounded token costs. It unlocks new platform capabilities that enable teams to:
With sovereign deployment options, predictable AI spend, and architectural coherence built in, Agent Experience delivers reliable, production-ready software and legacy modernization for even the most heavily regulated industries. And we couldn’t have built it without help from our customers.
After the Agent Experience Early Access Program was announced, more than 143 customer and partner organizations started building, and we are grateful to every one of them. They made more than 900,000 calls to the OutSystems platform through Agent Experience and close to 99% were created without error.
These organizations didn’t cut OutSystems any slack, either. They put Agent Experience through its paces, refactoring old monoliths, wiring custom MCP servers straight into SAP, and locking down zero-trust security gates. The impact included a major system refactoring with zero downtime, and integrations that take five minutes instead of an hour.
By adopting OutSystems Agent Experience, YESCO safely combines AI-driven productivity with enterprise-grade governance to deliver rapid business outcomes. Using the OutSystems secure platform layer to develop applications with coding agents such as Claude, YESCO has compressed software release cycles from six months to just two weeks.
This accelerated pace enables the company to safely deploy cutting-edge agentic apps. For example, their AI estimating tools save hours of design time, provide customers with instant pricing, and drive business growth in a trusted, fully governed environment.
“OutSystems Agent Experience is game-changing,” said Mike Neyman, Director of Software Integration and Operations, YESCO. “The platform’s open, model-independent approach helps us delight executives and customers with AI innovation while maintaining the governance and security required for enterprise use.”
Lowenstein Sandler is building on OutSystems with Claude Code and Agent Experience and using the Enterprise Context Graph to pioneer agentic legal applications. The firm’s development team is closing the vulnerability gaps of standalone AI tools by using Mentor's governance framework. “The Enterprise Context Graph is paramount. It’s what lets us build the right tools with full access to our data,” said Chris Palka, a software engineer at Lowenstein Sandler.
For example, Lowenstein Sandler built a production-ready Litigation Matter Tracker in under three hours. The new workflow cut application finishing times from one hour to just 5–10 minutes, all while maintaining the strict enterprise security, compliance, and auditability required in the legal industry.
Normal is using OutSystems Agent Experience to bring governed agentic AI directly into its software development lifecycle. By combining external coding agents like Claude with live enterprise context and built-in architectural guardrails, developers can analyze full application dependencies without re-specifying prompt context.
The result is the delivery of higher-quality analytics, comprehensive code reviews, and automated documentation. By feeding process markups directly into the OutSystems native coding agent to make applications production-ready, a single developer built a "Santa's Workshop" management application in just two weeks—slashing development time from a month and a half while maintaining strict enterprise security and governance.
“The difference between Agent Experience and standalone AI tools is the enterprise governance and architectural guardrails built into the platform," said Thomas Eldrup, solution architect at Normal.
Modernizing, building, and scaling agentic systems is only the start. How do you orchestrate and govern them to drive business impact?
Agent Workbench provides a single environment for teams to build, run, and control AI agents and to modernize existing processes. Since launching Agent Workbench one year ago, the product and engineering teams have shipped continuous innovation to give organizations the tools they need to transform their workflows. Here are just a few of the capabilities that are helping teams secure, align, and scale agents:
Together, these capabilities put you in control of your agentic workforce. Guardrails protect your data, evaluations keep agents on target, and model choice puts the right model on every job. Your teams operate at their best while costs stay in check.
These are excellent benefits, especially when combined with Agent Experience. And, if you’re like many other leaders, you’re ready to reap them all, but you haven’t figured out how or where to start. Today, OutSystems also introduced a solution that can help.
Agent Spotter helps you get started with agentic AI, using the applications you already run on OutSystems. Existing applications make strong starting points because they already have the knowledge and logic agents need. Agent Spotter analyzes your OutSystems application portfolio, including apps, objects, modules, and workflows and benchmarks against industry peers to identify the agentic opportunities that fit your business.
These insights then feed into a prioritized list of high-impact agents that can transform your existing business processes. Every recommendation comes with an agent spec pre-filled and ready to build in one step so you can fast-track agents to production.
This is how you turn your existing OutSystems portfolio into a continuous engine for AI-driven modernization. Now the question is how to move the needle on the core business, and do it while managing risk. OutSystems has an answer for that.
Ready to get started? You can schedule a session with our account teams to explore your top agent opportunities.
AI investment keeps climbing, while impact on the P&L stays close to flat. That line moves when agents run the work a business depends on. In banking, that means the loan. In insurance, it means the claim. These processes hold the most value, and they are the hardest place to get agents into production.
Take an insurance claim. An agent can read a photo and draft a note. Running the claim takes a full system. It needs a policy system connection, an industry data model, document handling, and an audit trail for regulators. Everything around the agents often takes longer and costs more than teams plan. In these scenarios, teams usually weigh two paths. They build the entire system themselves or adopt a packaged product built on a vendor's data model and roadmap.
Agentic Industry Solutions are the next layer of value in the OutSystems platform, and they help you realize returns that much faster. They offer a third path, with the system pre-built for the process that defines your industry.
Each of the Agentic Industry Solutions arrives mostly built, with applications, workflows, governed agents, industry data models, and connections to core systems. The designs draw on patterns OutSystems customers have run in production for 25 years with the same governance and audit trail.
Teams choose their own models and coding tools, and with Agent Experience and the OutSystems native coding agent, they can change any part of the system. Each company keeps control of what makes it different, including its products, rules, and customer experience. The payoff shows up in the process itself. A claim arrives complete before an adjuster opens it. A loan application reaches the officer already verified.
Insurance now joins the Agentic Industry Solutions portfolio. The Insurance Agent Kit is available today on OutSystems Forge. It includes two governed agents for claims work that fit into the systems insurers already run. Agentic First Notice of Loss, a full solution, follows in November. Along with the recently launched Agentic Loan Applications for banking, these insurance offerings accelerate AI adoption in regulated industries.
Fragmented tools and ungoverned output slow the path from AI investment to business results. Legacy systems that are largely untouchable for agents make it difficult to improve operations with AI. These releases address that problem across the lifecycle. Agent Experience lets teams transform and deliver fast with any coding tool. Agent Workbench helps teams run and manage agents at scale while Agent Spotter helps them prioritize their next agentic project. Agentic Industry Solutions accelerate outcomes in the processes that drive an industry’s business.
These announcements are a testament to our commitment to making OutSystems the agentic systems platform of choice. This comes from a firm foundation of customer trust and third-party validation. G2 users rank OutSystems number one in customer satisfaction and name it a Leader in AI Agent Builders. OutSystems was also named a Leader in the Gartner® Magic Quadrant™ for AI-Augmented Low-Code Platforms (AILCAP) and earned a #1 spot for AI-Driven Application Modernization in the 2026 Gartner® Critical Capabilities for AI-Augmented Low-Code Application Platforms report.
