OutSystems, the AI-enabled low-code development platform, reported rapid traction with its enterprise platform for building and orchestrating intelligent AI agents, Agent Workbench, at last week’s OutSystems ONE World Tour in Miami. The company revealed that over 5,500 agents are already in development, with nearly 1,500 certified developers actively building enterprise-grade agentic applications—signaling a swift move from pilot projects to production-scale AI workflows.
Since its general release in late September, the platform has been adopted across multiple industries as enterprises move from pilot projects to production deployments and begin orchestrating multi-agent workflows with built-in human oversight.
Early enterprise deployments
Several organizations have reported moving agentic use cases into production:
Axos Bank is using Agent Workbench to support operational efficiency initiatives, including an intelligent log-analysis agent that interprets error logs and recommends actions in real time, and a document-mapping agent that automates data entry tasks.
Thermo Fisher Scientific deployed a Customer Escalation Agent that interprets unstructured customer-interaction data to automate triage and speed case resolution, reducing manual workload.
Ascot Insurance is applying the platform to assess KPIs within its Avail Claims system. The workflow interprets unstructured activity logs and notes to determine KPI status, surfacing data previously inaccessible through legacy processes.
MAGnet Auctions built an agent that reads and validates odometer photos against internal records, reducing manual image review by more than 90% and improving data accuracy. The company said the automation has delivered cost savings and enabled staff to shift from quality-assurance tasks to higher-value work.
Open ecosystem and model flexibility
OutSystems positions Agent Workbench as part of an open agentic ecosystem, enabling organizations to connect any supported large language model and reuse that connection across applications. The platform integrates with model providers including Amazon Bedrock, Azure OpenAI, Anthropic, Gemini, Cohere, Mistral, Databricks, AI2, IBM watsonx, Vertex AI custom models, and Hugging Face open-source models, allowing teams to mix or swap models per use case based on cost, accuracy, or latency requirements.
The company is also expanding a library of prebuilt agents and agentic applications designed for widely used enterprise systems. Examples include agents that automate scheduling tasks in Google Calendar; generate updates or summaries in Confluence and Notion; and manage work-intake and task lifecycles in Trello, Asana, Monday.com, and Jira.
OutSystems has trained and certified nearly 1,500 AI developers on the platform as part of its global enablement programs and is encouraging enterprises to onboard AI development teams ahead of 2026.
Woodson Martin, CEO of OutSystems, said: "The enterprise AI race isn’t won by faster experiments; it’s won by building trusted AI-powered systems that scale. With Agent Workbench, enterprises are building an agentic future that connects data, managing the full agent lifecycle, and orchestrating human-AI collaboration to transform how work gets done and deliver unprecedented business value. The early momentum we’re seeing proves that enterprises are ready to move beyond experimentation and deploy AI that truly performs at scale."
Kevin Hearn, SVP, Head of Consumer Bank Development at Axos Bank, said: "Our work with OutSystems on Agent Workbench is a significant step in our AI strategy. We’re expanding the use of AI capabilities within OutSystems to realize immediate gains, without needing to invest in specialized AI roles. With Agent Workbench, we can quickly and safely create agents for specific use cases, such as analyzing error logs and automating data entry from documents. By creating and embedding agents into our operations, we are building a more intelligent and responsive banking ecosystem that is ready for the future of finance."