Bridging the AI Agent delivery gap, Tray.ai unifies composability, AI and integration on a single platform to optimize how developers implement and maintain high-value AI agents
Saves enterprise IT from pain of coding custom agents and chaos of single-purpose SaaS app agents
SAN FRANCISCO – Tray.ai, innovator of the AI-ready composable integration platform, today announced Tray Merlin Agent Builder to speed the creation and deployment of high-value, production-ready AI agents. To make it simpler for enterprises to get started with AI agents, Tray.ai also unveiled new curated Tray Agent Accelerators that provide a user-friendly experience for rapid AI agent implementation, offering templates for knowledge agents, IT ticketing agents and customer support ticketing agents. Expanding Tray’s low-code AI agent development capabilities, the new advancements give enterprises the fastest, most flexible, and safest path for scalable AI agent delivery. Enterprises can now avoid the traps of high costs and long lead times typical in custom agent development as well as the constraints and silos created by implementing and managing single-purpose agent offers from each SaaS application in the enterprise tech stack.
“Enterprises are rushing to deploy AI agents even though they’re missing essential building blocks for real business impact. Without the foundation of a composable AI integration platform to build agents quickly, flexibly and safely, AI agents will remain limited in scope and fail to meet expectations,” said Rich Waldron, co-founder and CEO of Tray.ai. “Two common approaches derail AI initiatives: either teams try to build with custom code that creates unnecessary technical debt, or they cobble together numerous off-the-shelf SaaS products that lead to fragmentation and integration challenges, turning their CIO into an AI-referee. An AI-ready iPaaS is essential for providing the unified, scalable infrastructure needed to bridge the agent delivery gap and effectively deploy and manage high-value AI agents across the enterprise.”
Siloed approaches and single-purpose products lead to governance gaps and vendor lock-in. Forrester predicts 75% of firms that build agentic architectures on their own will fail. Instead, Tray.ai is purpose-built to handle every step of AI agent development and deployment in one unified platform. Collectively, the capabilities announced today extend Tray.ai’s recent Merlin Intelligence and Enterprise Core agent-enabling platform innovations, including data integration for AI to handle unstructured data and retrieval-augmented generation (RAG) pipelines; Merlin Guardian to ensure safety; and Merlin AI Palette for LLM-connectivity and native AI services.
Industry Data: Integration Hurdles and Security Risks Are Blocking Enterprise AI Adoption
IT is under pressure to deliver agents fast, with Gartner predicting that, “by 2028, at least 15% of day-to-day work decisions will be made autonomously through agentic AI, up from zero percent in 2024.”¹ Development teams are struggling with siloed tools, fragmented governance and limited functionality that makes promising prototypes unfeasible in production. Meanwhile, vendor lock-in threatens to limit agent connectivity and capabilities, while lack of integrated data handling and process automation results in disappointing business outcomes. No wonder Gartner states that “by 2028, 25% of enterprise breaches will be traced back to AI agent abuse from both external and malicious internal actors,”² highlighting the risks of rushing development without proper controls.
According to the “State of AI Agent Development Strategies in the Enterprise” survey of over 1,000 enterprise technology leaders released today by Tray.ai, 42% of respondents need access to eight or more data sources to deploy AI agents successfully — which is impossible when SaaS app agents are restricted in scope by the integrations to which their host applications have access. Further, nearly 90% of enterprises said they consider integration with organizational data systems (CRM, ERP, data lakes, etc.) essential for AI agent success.
As Waldron explained, “Companies have an opportunity to tackle agent initiatives more efficiently by unifying their agent development with integration and automation needs on a single platform. This way, organizations maintain flexibility to build and adapt agents to work across their entire technology stack and ever-evolving business processes.”
Notion’s Head of Business Technology Jalal Iftikhar said, “AI agents are revolutionizing collaboration and knowledge management, and they’re at the heart of our long-term strategy at Notion. Building AI agents successfully hinges on seamless, enterprise-wide integrations to support cross-departmental business processes. An AI-ready iPaaS provides the speed, flexibility, and deep integration across the entire stack that are critical to creating and managing agents effectively. This integrated foundation ensures full control over the end-to-end process.”
Merlin Agent Builder Provides Flexibility for Any Agentic AI Use Case
Tray makes it simpler for enterprises to create, manage and deploy AI agents in production without compromising on flexibility and control. These agents can be conversational or deployed in a proactive asynchronous fashion – an important step toward the autonomous enterprise. Built on the Tray Universal Automation Cloud and powered by Merlin Intelligence, Merlin Agent Builder provides a guided setup of Tray Agents, incorporating visual workflow-based tools that define each agent’s capabilities and scope, including governance and integration points. Organizations can now centrally deploy conversational agentic experiences across departments through Slack and Microsoft Teams while maintaining granular control of data access and permissions via customizable tools. This unified approach eliminates the need for single-purpose and siloed agent products while guaranteeing proper governance, monitoring and direct integration with enterprise CRM, ERP, and HCM systems.
With Merlin Agent Builder, teams have complete control of RAG pipelines, systems or data events which invoke agents, chat interfaces that employees can use to interact with agents, and tools and knowledge the agents can leverage. They can track, optimize and refine agents to improve effectiveness, responsiveness and overall performance. For extra data protection, organizations can tokenize sensitive PII data with native AI capabilities like Tray Merlin Guardian prior to any agent invocation or within tools that agents rely on.
“Implementing AI agents through Tray.ai’s platform and the new Merlin Agent Builder will be transformative for our organization’s operations,” said Lee Hassan, CTO of Aprende Institute. “The platform’s robust security protocols support safe agent deployment, while its efficient data integration capabilities, including integration with Slack, minimize training needed for our staff so they can immediately get benefits from using the agent without needing to learn a new tool or platform. With Tray.ai’s unified approach, what we estimated would be a multi-quarter project was in production with just a couple days of work.”
Additional capabilities support safe and efficient agent delivery:
Streamlined AI agent testing with step isolation: Teams can test any execution point and state directly within Tray Build, replacing legacy iPaaS approaches and external testing tools. Step isolation pulls execution data for targeted testing, so users are able to save or discard changes instantly and update output schemas directly. This accelerates AI agent development and debugging while ensuring robust, trusted agent performance.
File management transforms unstructured content handling: Agents autonomously process orders, invoices, B2B files and creative assets — extracting text, creating structured data payloads from unstructured file types and validating file structures. Agents can now manage the complete lifecycle of file-based tasks, from initial upload through content manipulation and routing, reducing manual effort while maintaining control.
Pre-built Agent Templates Fast-track Enterprise AI Agent Implementation Time
Tray Agent Accelerators, built on the Tray platform, significantly lower the barrier to entry for organizations deploying agents, providing expert-built templates for rapid implementation while ensuring best practices. These ready-to-use accelerators include RAG implementation guides, AI knowledge microservices, Jira and Slack knowledge indexing and Slack agent orchestration.
Knowledge Agent – transforms internal documentation into actionable insights, reducing response times from hours to minutes and preserving institutional knowledge across teams.
IT Ticketing Agent – provides 24/7 automated support for common requests, optimizing service desk operations without increasing headcount, all while maintaining compliance with governance policies.
Customer Support Ticketing Agent – delivers instant, accurate responses across channels, streamlining costs through automated ticket classification and routine issue resolution.
“The Tray platform uniquely consolidates AI agents with process automation and data integration capabilities, providing the critical foundation for RAG pipelines and enterprise-ready testing,” said Alistair Russell, co-founder and CTO of Tray.ai. “App dev teams can take advantage of these new advancements to deploy sophisticated AI agents capable of autonomously processing unstructured data and handling complex file management across their entire tech stack without vendor lock-in or workarounds. This unified approach streamlines organizations’ IT spend while ensuring AI initiatives can scale across any business process.”
To learn more about Merlin Agent Builder:
¹ Gartner, Top Strategic Technology Trends for 2025: Agentic AI, Tom Coshow, Arnold Gao, et al, 21 October 2024
² Gartner, Gartner’s Top Strategic Predictions for 2025 and Beyond: Riding the AI Whirlwind. Daryl Plummer, Leigh McMullen, 21 October 2024
For all press and media enquiries please contact: Stephanie Floyd Bhava Communications for Tray.ai press@tray.ai +1 (760) 410-8077