How OutSystems is evolving to give enterprises leverage and digital sovereignty in the AI era

According to McKinsey, more than two-thirds of organizations say their AI initiatives remain stalled in experimentation or pilot mode. The investment shows up clearly in one place on the P&L, and that place is the expense line. Revenue impact and margin improvement, for most, are still aspirational.

How do you close the gap? At our 2026 ONE Conference in Amsterdam this week, OutSystems unveiled the platform we have been building to do exactly that: the Agentic Systems Platform. Now every organization has an onramp to enterprise AI, while preserving the autonomy and control they need to meet their regulatory, operational, and financial imperatives.

You should not have to choose between moving fast on AI and preserving your IP, market advantages, and control. Read on to learn what that means in practice, the innovations we just announced, and what your team can do this week to get started.

The agentic era demands a new kind of platform

The scope of what OutSystems delivers has grown so much that it is no longer accurate to call it a low-code platform. It is now an open Agentic Systems Platform, powered by the OutSystems Enterprise Context Graph, with a governed, secure foundation for agentic innovation.

This evolution reflects the heightened strategic value of a governed platform in the age of AI, especially for organizations that are realizing the spiraling costs and risks of AI-generated code and agents across their business. As disparate AI tools proliferate and every team builds its own stack, proprietary data is leaking into markdown files on employee devices, and institutional memory is getting locked inside agents that operate outside IT oversight.

And the economics are turning volatile. Inference costs are climbing month over month, and the major AI vendors are still discounting heavily to win adoption. As that pricing shifts toward consumption, teams that built directly on frontier models face wild swings in operating costs, and architectures so wedded to one model that swapping to a cheaper option means a rewrite. Leaders, meanwhile, have more questions than ever about vendor lock-in, security, cost trajectory, and audit readiness. But until now, they have lacked real choice and leverage in the AI marketplace.

This is not a sustainable situation for any enterprise—and demands a new kind of platform. A trusted, unbiased platform that acts as a neutral AI broker, giving you the ability to choose and swap out AI models as newer, more purpose-built and cost-effective ones become available. This is what the OutSystems Agentic Systems Platform is uniquely designed to deliver, because it already works with your architecture and enterprise context.

The Agentic Systems Platform brings the entire agentic lifecycle into one governed environment, with new capabilities for engineering with any AI coding tool, orchestrating multi-agent workflows, and new ready-made solutions for highly regulated industries.

Unlocking your AI advantage with the Enterprise Context Graph

Most AI systems operate on prompts, APIs, and documents. But enterprises run on deterministic workflows, rules, decisions, and the interactions between systems. AI, by nature, is probabilistic— it generates answers based on patterns. Without operational context, that gap creates risk: AI can make suggestions that sound helpful but produce outcomes that are generic or disconnected from the real business.

The Enterprise Context Graph closes that gap by giving AI a working model of the enterprise applications you have already built on OutSystems. It turns those workflows, integrations, and governance into a structured context that AI can reason with. Agents see relationships, dependencies, and what might break when they make a change, so they build with greater precision because they understand the architecture they are working inside.

AllianceCorp Manufacturing (ACM), a global manufacturing technology leader and longtime OutSystems customer, is using their Enterprise Context Graph to move beyond siloed reporting and make production data actionable for agents across the business.

Exposing context to a broader agentic ecosystem

But what if you want to use this enterprise context outside the OutSystems platform?

That is exactly why we built the Agent Experience, a new platform layer that exposes a suite of A2A and MCP tools and services that enterprise developers can trust to build, orchestrate, and govern their agentic portfolio. Rather than keeping this data trapped, the Agent Experience exposes the Enterprise Context Graph alongside core platform capabilities to external AI agents via three core categories of natural-language services:

• Context Services: Allows external agents to query your entire application estate, schemas, and relationships before building or modifying code.
• Platform Services: Empowers agents to safely trigger lifecycle operations like deployments, rollbacks, and quality validations under strict ODC Portal approval flows.
• Mentor Services: Outlines the exact blueprints so agents can delegate platform-native generation of entities, screens, and logic directly to Mentor.

This open architecture bridges the gap between external AI tools and unified, governed lifecycle delivery. Enterprises can now see, decide, build, and ship in a single, continuous prompt chain, all within the AI tools developers already prefer.

Delivering agentic systems for the enterprise: Expanded AWS & OutSystems partnership

AWS delivers the world’s most advanced AI infrastructure. Seamlessly integrating with these capabilities, OutSystems provides the enterprise-grade agent and app development layer required to turn that core technology into mission-critical agents and apps that deliver real business impact. By expanding our partnership, we are bridging the gap between AI potential and production-ready reality. Together, we’re delivering native product integrations and new agentic solutions that accelerate legacy modernization, enable model flexibility, and provide the AI scalability and enterprise governance businesses demand.

Accelerate legacy modernization with AWS Transform

Legacy modernization has traditionally been slow, manual, and expensive. Engineering teams have spent months assessing legacy code, manually rewriting logic, validating changes, and stitching together disparate tools just to move critical systems forward.

With OutSystems Agentic Legacy Modernization, that’s no longer the case.

Teams can now harness the power of agentic AI to automate and accelerate the most tedious aspects of legacy migration. Using our native agents alongside AWS Transform, teams can automatically analyze and decode massive legacy estates like Lotus Notes, COBOL, and Java. From there, developers use AWS Kiro and Mentor to easily re-architect those specs into modern, agentic applications that land directly inside OutSystems. By leveraging your unique Enterprise Context Graph, this human-in-the-loop approach ensures your new agentic system automatically aligns with your business logic, ensuring it is governed, secured, and delivering real value from day one.

Give developers freedom while maintaining enterprise control: Kiro integration

OutSystems now natively integrates with Kiro, the agentic development environment from AWS, bringing spec-driven development into the enterprise application lifecycle. Instead of relying on manual handoffs and disconnected coding workflows, teams can use specifications, APIs, and schemas from Kiro to guide how legacy applications are rebuilt, managed, and governed in OutSystems. Developers get to work in Kiro, the advanced coding environment they prefer, while IT maintains the enterprise context, governance, and lifecycle control required to modernize critical systems safely and at scale.

Build agentic workloads using the AI model of your choice with Amazon Bedrock

As organizations look to scale their agentic portfolio, maintaining strategic control and cost efficiency across a fragmented market is critical. With deepened support for Amazon Bedrock, which provides a wide range of AI models with a single set of credentials, teams can orchestrate agentic systems with the right models for the job and scale solutions fast, ensuring cost efficiency and high performance without changing their underlying system.

Securing and optimizing the agent workforce with Agentic Enterprise Orchestration

When you accelerate legacy modernization and give teams access to powerful system integrations, you unlock the ability to deploy AI agents and automate workflows at scale. But when anyone in your organization can build and deploy agents, it can quickly lead to a messy, fragmented landscape with runaway token costs, security blind spots, and potential data exposure. Agentic Enterprise Orchestration addresses these challenges, providing teams with one place to manage agentic orchestration end to end.

Powered by the Enterprise Context Graph and featuring Amazon Bedrock, it is the next generation of Agent Workbench, and it lets organizations coordinate and govern agents, apps, and workflows across the enterprise with:

• Advanced agent evaluations that validate the outcomes your agents are responsible for, the way unit tests validate code
• Precise guardrails that enforce safety and compliance across agents, including PII masking before data reaches a hyperscaler or foundation model provider
• Agent2Agent (A2A) connector, so you can integrate external agents and extend
your agentic ecosystem
• Model swap with Amazon Bedrock, so you can continuously optimize with access to the most cost-effective and high-performing models

Ultimately, teams gain strategic control with Agentic Enterprise Orchestration, allowing them to curb runaway costs, protect data, and ensure agentic integrity at scale.

These capabilities are available to existing customers with Agent Workbench. Every OutSystems customer gets 10,000 free executions per month included with OutSystems Developer Cloud, so your team has a runway to get it right without renegotiating the budget.

Scaling quickly and confidently with Agentic Industry Solutions

Most templates and starter kits look good in a demo and run out of road the moment you try to put them into production. OutSystems Agentic Industry Solutions are different. A new portfolio of pre-built solutions from OutSystems and their partners, they streamline the deployment of agentic technology, de-risking AI adoption while sharply accelerating time-to-value. These are working systems that solve real problems, not sample apps. Because they're built on OutSystems, they're designed to be entirely customizable, so your team can make them their own. Each solution extends your Enterprise Context Graph so AI agents can reuse trusted components and reduce token burn. Because they also draw on Amazon Bedrock, they benefit from a library of domain-specific models, including models trained for finance, law, and tasks like document processing, to optimize performance and cost at scale.

At ONE, OutSystems unveiled its end-to-end Banking Solution for Loan Origination, which ships with the foundational work already done. Out-of-the-box agents handle document processing and policy retrieval, and there are connections to core systems like Jack Henry and FiServ, a full auditable workflow, and an industry-standard data model. Your team customizes the parts that differentiate your bank instead of architecting intake agents and compliance workflows from scratch.

Interested developers can start today with the OutSystems Banking and Financial Services Agent Kit. The first end-to-end Banking Solution will be available early Q3, with solutions for insurance, manufacturing, and the public sector on the way, and partners shipping additional industry solutions alongside us through the Elevate Partner Solution program.

Axos Bank, which has built agentic systems on OutSystems for years, recently delivered a 60% productivity improvement in some of its core banking workflows using these agentic capabilities.

Running AI workloads where your business requires

Speed and innovation are only half the answer. The other half is control and sovereignty.

More than 100 countries now have some form of data sovereignty or localization regulation, and financial services, healthcare, telecom, and energy each face their own industry-specific requirements. Your AI workloads need to run where your regulators, customers, and CFO require them to run.

OutSystems announced a new self-hosted deployment model for OutSystems Developer Cloud, letting customers run their runtime and data on their infrastructure, while benefiting from the modern OutSystems cloud development experience.

Deployment paths now include:

• Customer data center/On-premises
• Private cloud
• Regulated or sovereign environments such as AWS GovCloud and AWS European Sovereign Cloud

What your team can do right now

The capabilities we announced at ONE are platform layers your team can start using right now to bring agentic work under one governed environment. Here is where to start:

• Fire up the latest Agent Workbench and begin orchestrating your agents with the new guardrails, evaluations, and A2A connector. The 10,000 free executions per month give you room to experiment.
• Download the banking agents from Forge. Even if you are not in banking, you can study how the agents are architected and adapt these proven patterns to your business.
• Try Next-gen Mentor, generally available today with included credits until September 30.
• Learn more about our expanded partnership with AWS.
• Talk to us to help you define your digital sovereignty strategy.

The future of enterprise AI is governed end to end. Agents reason with full operational context. Models swap based on cost and quality. Sovereignty requirements are met by design. And because the platform acts as your neutral AI broker, executives get clean answers about cost trajectory, risk, and where the AI investment is paying back. That future is what the Agentic Systems Platform delivers, and your team can start building it this week.