OutSystems reports rapid adoption of Agent Workbench, with over 5,500 agents in development and enterprises transitioning from AI pilots to production-ready, multi-agent workflows.
OutSystems, the AI-powered low-code development platform, announced strong early traction for its Agent Workbench, an enterprise-grade platform for building and orchestrating intelligent AI agents. Following its launch on September 30, the platform is rapidly moving enterprises from pilot programs to full production of agentic AI systems.
The momentum, showcased at the OutSystems ONE World Tour Miami conference, includes impressive adoption metrics:
- 5,500+ AI agents are currently in development across the platform.
- A community of nearly 1,500 certified developers are actively building enterprise-grade agentic applications.
“Faster experiments don’t win the enterprise AI race; it’s won by building trusted AI-powered systems that scale,” said Woodson Martin, CEO of OutSystems. The company’s goal is to manage the full agent lifecycle and orchestrate human-AI collaboration to transform business operations.
From Pilots to Production: Enterprise Case Studies
Enterprises across various industries are already deploying Agent Workbench to achieve measurable business outcomes, moving beyond simple experimentation.
| Company | Industry | Agent Application | Business Impact |
| Axos Bank | Finance | Log Analysis Agent, Document Mapping Agent | Reduces need for manual analysis of error logs; improves accuracy and eliminates repetitive data entry. |
| Thermo Fisher Scientific | Science/Research | Customer Escalation Agent | Automates triage of customer issues by interpreting unstructured data, reducing manual intervention and improving time to resolution. |
| Ascot Insurance | Insurance | Avail Claims KPI Agent | Intelligently interprets agent activity logs and notes to instantly assess Key Performance Indicators (KPIs) from previously inaccessible unstructured data. |
| MAGnet Auctions | Automotive/Auctions | Vehicle Entry QA Agent | Automatically reads odometer photos and flags discrepancies. Reduced manual human review by over 90%, delivering significant cost savings and enabling resource reallocation to higher-value work. |
Kevin Hearn, SVP and Head of Consumer Bank Development at Axos Bank, noted that the platform enables them to “quickly and safely create agents for specific use cases” without requiring investments in specialized AI roles.
Also Read: How Is Neat Transforming Hybrid Work with AI and Inclusivity?
Open Ecosystem and Model Flexibility
OutSystems promotes an open, agentic ecosystem by enabling enterprises to integrate and deploy AI agents across their critical workflows using the Large Language Model (LLM) of their choice.
- Model Flexibility: Teams can connect a model once in Agent Workbench and reuse it across multiple agents and applications.
- Broad Support: The platform supports a wide array of models, including Bedrock, Azure OpenAI, Anthropic, Gemini, Cohere, Mistral, Databricks, IBM watsonx, Vertex AI custom models, and Hugging Face open-source options. This flexibility allows teams to route, mix, or swap models based on a use case’s requirements for accuracy, cost, and latency.
The platform provides a fast-growing library of ready-to-use agents for common enterprise systems, ensuring agents can be deployed where work happens every day:
- Productivity: Google Calendar for scheduling and follow-ups.
- Knowledge/Collaboration: Confluence and Notion for summarizing and sending reminders.
- Work Management: Trello, Asana, Monday.com, and Jira for automating task intake, assignment, and progress tracking.
OutSystems has emphasized its commitment to making enterprise AI adoption accessible by training and certifying nearly 1,500 AI developers on the platform. Agent Workbench is now generally available worldwide.
