An AI agent may need live data from a mainframe running a critical business application. Moving that application to the cloud simply to give the agent access can add cost without improving its performance. Massachusetts-based unified data management leader Precisely offers another route. Its software can extract live data from a mainframe and feed it into an agent or cloud environment while the original application continues to run. That lets companies make existing information available for AI without replacing the systems that produce it.
The newly unveiled Precisely Platform aims to make that information easier for AI agents to interpret and use. Governance rules guide how teams and agents can use it, giving companies a way to prepare data for AI across separate systems.
“People often assume that modernization means migration, but I firmly disagree with that assumption. I believe the starting point should be the organization’s goals. What are you trying to achieve as a business, and what risks do you need to address? Those questions should guide your approach to modernization,” says Ashwin Ramachandran, Precisely’s senior vice president of product.
The company serves more than 12,000 organizations, including 92% of the Fortune 500. The company plans to move its new platform from preview to general availability in early 2027. Developers can use AI Studio’s ready-made agents and development resources, while new Model Context Protocol (MCP) connections let compatible AI assistants access capabilities across existing Precisely products.
A Unified Data Foundation for Enterprise AI
The platform combines a shared data catalog with consistent business definitions to help AI agents understand information across enterprise systems. Ramachandran explained that much of this unification happens at the logical layer, giving companies a common view of their data while allowing the underlying systems to remain separate. Teams can access the platform’s capabilities through its Gio AI assistant.
For mainframes, that approach means making live data available to agents through an interface that keeps them from operating directly on the transactional systems running the business.
“We want to ensure that this interface exists, which brings us back to the role of the data and infrastructure layer. We have high-performance mechanisms that allow us to extract live data from a mainframe platform and feed it into an AI agent or a cloud-native stack,” says Ramachandran. “This gives agents access to the information they need while allowing them to work at the data level, without operating directly on the mainframe itself.”
Stewart Bond, IDC’s vice president of data intelligence and integration software, says organizations face increasing pressure to operationalize AI, yet many struggle with fragmented data environments, inconsistent governance and limited visibility. Those gaps become harder to manage once agents act on the data. “The industry’s shift toward unified data intelligence platforms reflects the need to bring integration, quality, and governance together in a single offering. Organizations that establish this foundation will be better positioned to successfully scale their AI initiatives,” he added.
Giving AI Agents Governed Access to Enterprise Data
Precisely provides third-party data that adds business context to an organization’s existing records. Its linked datasets span more than 250 countries and territories, covering location and business attributes. Customers can combine their own information with wildfire risk data or demographic insights to inform decisions that require a broader view than internal records alone can provide.
“Customers can compare these datasets without having to build custom routines to connect them. That is an important part of how we make this information useful. We cover the data they already own across the enterprise and provide the third-party context that helps them interpret it more fully. Together, those sources give them a broader foundation for making informed business decisions,” Ramachandran says.
Address verification and geocoding connect records to validated addresses and geographic coordinates, giving AI agents a clearer understanding of the locations behind the data. According to Precisely, the platform can attach quality and governance scores to individual data assets. Its embedded AI helps teams generate data quality rules and identify opportunities for enrichment, while subject-matter experts establish the business standards that guide those workflows.
“You can take the rules your business teams have defined and embed them directly into the processes you use to correct and improve data. That means everyone works from a shared understanding of what good data looks like, established by someone with the authority to define that standard. From a data engineering perspective, we inherit those requirements and use them to guide our data workflows,” Ramachandran says.
Establishing the rules once reduces repeated configuration and helps prevent standards from drifting as different teams put them into practice. The MCP servers extend workflows to compatible AI assistants, including Claude, Microsoft Copilot and ChatGPT. Callers authenticate using an API key, while the platform enforces the permissions administrators have assigned.
“Even when a tool for creating a data quality pipeline is available to an agent, the platform still checks whether the caller has permission to use it. If that caller does not have permission to author pipelines, the platform blocks the request. Making the tool accessible does not change the permissions required to run it,” says Ramachandran. “We manage and enforce those controls according to the access rules and permissions administrators have established within the platform.”
Across the portfolio, Precisely Automate connects agents to SAP workflows, while EngageOne limits communications requests to each user’s permissions and records those requests in audit logs. For customer-managed Syncsort, access relies on controls within the customer’s mainframe environment. The company plans to introduce Syncsort’s MCP capabilities in the coming months, with MCP servers available by subscription to customers with qualifying products.
Modernizing Without Forcing Migration
AI Studio gives developers starting points for building with enterprise data, including a property analyzer app and tools for configuring data replication. Customers can explore these resources through a free trial and adapt them to the requirements of their own workflows.
“We do not assume that the skills and agents we provide through AI Studio will meet every customer’s precise requirements without further adaptation,” Ramachandran says. “The initial collection gives customers a place to begin, and I would like to see it develop into a community over time. Customers could contribute the skills they have built for their own domains and industries, sharing how they have adapted those capabilities and where they are finding value. That is something we will need to grow into beyond the initial launch, but it is the kind of participation I would like to see as customers build on what we have provided.”
Jimmy Duchesne, director of solutions innovation and presales at Precisely partner Korem, points to the difference natural-language access has made in putting the technology to work. “Being able to interact with data and engines from Precisely with natural language, and easily build GeoAgentic solutions, has been a game changer,” he says, in a written statement. “Time to value used to be calculated in months. Now, it can be calculated in days.”
Ramachandran plans to track MCP server usage as customers adapt and expand the initial collection. He expects agents to make some migrations more practical by reducing the need to recruit specialized expertise for a full application rewrite. “In other cases, agents will simply increase the demand for data,” he explained. “Each enterprise will assess that differently, based on its appetite for risk and the costs it is prepared to carry. Those considerations will shape how an organization responds to that growing demand and whether migration makes sense for its business.”
Precisely’s approach gives companies a way to put existing data to work for AI while evaluating each migration on its business merits. Systems that still perform well can support new applications without a wholesale rebuild. That gives enterprises greater control over where they invest, with the value of the next business opportunity determining which systems need to change.
