Customers keep their existing Adobe permissions, skip the infrastructure build, and start working with agentic AI on day one
Adswerve today announced that MiaProva, its Adobe optimization and analytics platform, has integrated Adobe's newly released Model Context Protocol (MCP) tooling, giving Adobe customers a fast, governed way to start working with Adobe's agentic AI tools, using the Adobe permissions they already have.
Adobe announced general availability of its MCP tooling at Adobe Summit in April and confirmed that customers can use these tools with any AI provider. Organizations can connect the AI provider they already use through MiaProva, or lean on MiaProva's native integration with Google Cloud, without standing up new infrastructure, contracting a new AI vendor, or requesting elevated access from a security team. Building the equivalent in-house means standing up MCP infrastructure, wiring in a provider, solving the permissions problem, and then figuring out what to actually ask — a quarter of engineering time that Adswerve customers skip entirely. Through MiaProva, customers get a dynamic working environment instead of just a project plan.
Adswerve acquired MiaProva in January 2024, deepening its long-standing commitment to helping organizations maximize the value of their Adobe investments. MiaProva has spent more than a decade helping teams scale testing and optimize programs with industry-leading best practices, automate reporting, deploy risk-mitigating alerts, and build the visibility that surfaces not only real-time test results but the analytics leaders need to understand true impact.
Adswerve has continued to expand deep integrations across Adobe Target, Adobe Analytics, Adobe Experience Platform (AEP), and Adobe Real-Time CDP. MiaProva supports optimization and analytics teams at enterprise retailers, financial services firms, and global technology companies, several of which are in early use of the Adobe MCP integration.
"Adobe gave the market a powerful set of MCP tools and the freedom to use them with any AI provider. Our job was to make that real for customers on day one,” says Brian Hawkins, Head of Optimization Technology at Adswerve. “We focused on the two things that actually slow adoption down: governance and knowing where to start. If security won't sign off, nothing else matters, so permission-aware access came first. With that in place, plus a library of tested prompts, organizations can get to work immediately, and do it safely."
The friction MiaProva was built to solve (such as bottlenecks, manual workflows, and disconnected context in a scaling optimization program) is the same friction that stalls MCP adoption. MiaProva's answer here follows the same logic. Beyond the infrastructure savings, MiaProva was built to solve two other things that typically stall MCP adoption.
First, MiaProva expanded its server-to-server OAuth integration with Adobe to take advantage of user-based OAuth already in place for Adobe Analytics. When a user initiates a session with Adobe MCP tools inside MiaProva, the same permissions governing what they can do manually in Adobe are applied to what they can do through Adobe's MCP servers.
This is a meaningful step forward for governance. Many organizations stand up MCP access using server-to-server OAuth, which typically defaults to System Administrator permissions. Left unchecked, that means someone with read-only access to a live test could type a plain-language request and turn it on or off, something they'd never be authorized to do through the interface itself. MiaProva closes that gap by enforcing the user's actual permissions at the session level, so access through MCP mirrors access in the tools themselves.
Second, MiaProva built in default prompting. Customers get access to dozens of effective prompts that the Adswerve team has found and tested to be valuable in real Adobe environments. Rather than starting from a blank box, teams start from a curated library of prompts that already work.
MiaProva also enables collaboration around prompting. Users can save their most effective prompts, share them internally with colleagues, or add them to a Pack, which serializes related prompts so they can be reused and run as a set. The result is a shared, growing library of institutional knowledge rather than expertise that lives in one person's head.
The Adobe MCP tooling integration is available now to MiaProva customers. To see it in action, watch Adswerve’s on-demand webinar where the team will walk through the integration, demonstrate governance in action, and share prompting best practices from real Adobe environments.
