SnapLogic Adds MCP Builder to AI Integration Portfolio

SnapLogic launched the MCP Builder, a template that automatically generates Model Context Protocol servers from existing integrations and APIs, enabling IT teams to expose deterministic pipelines to AI agents while maintaining governance and audit trails through its Agentic Integration Platform. The move addresses challenges in operationalizing AI at scale, including ensuring reliable task completion, controlling token costs, and applying data management policies to AI agents.

SnapLogic today made available a set template that makes it simpler to build Model Context Protocol MCP servers that are integrated with its integration platform-as-a-service iPaaS environment. The SnapLogic MCP Builder automatically generates MCP Servers from existing integrations, OpenAPI specifications, and Application Programming Interface API Management services. As a result, IT teams can expose existing deterministic pipelines to AI agents that can be managed via the existing SnapLogic Agentic Integration Platform https://techstrong.ai/features/snaplogic-extends-reach-of-ipaas-environment-deeper-in-the-realm-of-ai/ that provides access to more than 1,000 connectors. Dominic Wellington, director of product marketing for AI and Data at SnapLogic, said SnapLogic MCP Builder makes it simpler for IT teams to apply a set of best practices for building and deploying MCP servers through which access to data by AI agents can be effectively monitored and governed using an AI Gateway developed by SnapLogic. A Trusted Agent Identity capability also propagates user identity and permissions across downstream systems to provide IT teams with a complete audit trail, he added. After launching a wave of AI initiatives with mixed success, more organizations are now looking to operationalize AI applications at scale. The challenge they face is finding a way to ensure that AI agents that have been assigned a task have the context needed to reliably complete it. Otherwise, AI agents will simply invoke the probabilistic capabilities of a large language model LLM to make a best guess. The issue is that many of the tasks the typical enterprise IT organization wants to automate need to be completed the same way every time in a way that can be easily audited. Additionally, organizations are also now trying to ensure that AI agents consume as few tokens as possible. Otherwise, the total cost of automating tasks using AI agents can easily spiral out of control. In effect, best FinOps practices need to be applied to token consumption, otherwise known as tokenomics, noted Wellington. SnapLogic is making a case for using the same platform that makes it simpler to extend the same policies previously defined to AI agents, he added. Unfortunately, too many organizations simply launched AI pilot projects just to build something. The assumption was they could figure out the details later, most of which revolve around issues pertaining to data access, security and governance, said Wellington. It’s not clear just how far along organizations are when it comes to truly operationalizing AI. While there is clearly a lot of pressure being applied by senior business leaders to make sure the organization is not losing ground to competitors, most organizations are struggling with the same fundamental data management and governance issues. The fact is that many data management issues that have been long ignored or simply glossed over are now raising their ugly heads in the age of agentic AI. Hopefully, AI will lead to best data management practices becoming more widely adopted. In the meantime, however, organizations need to carefully consider what tasks are being automated by AI agents based on what data. Otherwise, the output being generated is going to lead to sub-optimal outcomes that do more harm than good.