
Gartner’s recent IT Symposium/Xpo revealed a fundamental shift in how CISOs must approach data governance, external collaboration, and AI risk management. The most striking insight is about timing and scale: by 2028, Gartner predicts 90% of B2B buying—over $15 trillion—will be intermediated by AI agents. Imagine procurement, finance, and vendor agents exchanging data and making decisions autonomously.
The governance challenge now extends beyond Shadow IT to what I call Shadow AI Agents: autonomous systems making data-sharing decisions outside traditional security boundaries.
After two decades as a CISO and Trust Officer across life sciences and technology, I’ve navigated cloud migration, SaaS proliferation, and remote work. But today’s challenges are unique: velocity is faster, complexity is higher, and the governance gap is wider than ever.
Here’s the central thesis: Gartner’s research on agentic AI, unstructured data security, and third-party risk management all converge on one critical capability—unified external data governance. Fewer than one in four organizations have built this foundation.
This analysis explores why unified external data governance matters, what Gartner’s research reveals, and what security leaders should prioritize in the next 90 days.
As we look ahead, three mega-trends stand out, each reshaping the security landscape in ways that demand our attention and action.
We’re on the verge of seeing AI agents automate procurement, customer support, and partner integrations, exchanging data and negotiating contracts with minimal human oversight. Our legacy controls, built for human workflows, won’t keep up. Unified visibility and governance over these autonomous data flows is no longer optional.
GenAI and agentic AI rely on unstructured data: documents, emails, recordings. Legacy DLP and CASB tools are reactive and limited, often missing critical external shares across platforms like SharePoint, Teams, and custom portals. Real-time, context-aware governance across all channels is now essential.
AI agents will interact with external partners across diverse regional platforms and regulatory frameworks, creating complex compliance and accountability challenges that traditional vendor risk assessments weren't designed to handle. Ownership of third-party cyber risk is often unclear, and traditional vendor assessments are insufficient. Real-time monitoring of data flows, and clear accountability are now required.
Recognizing these trends is only the first step. Building unified external data governance that can enable real AI value requires three foundational capabilities that most security organizations lack.
Perfect policies are the enemy of scalable AI governance. Segment AI use cases by risk tier and apply tailored controls. Build operating models and oversight systems that enable visibility and accountability for high-risk, externally facing AI collaborations. Focus governance on the AI portfolio, not every conceivable use case.
Move beyond technical metrics. Connect security initiatives to business outcomes—revenue acceleration, risk reduction, and competitive advantage. Frame achievements in terms executives understand: “Secure external collaboration reduced contract cycle time by 18%, enabling $4M in incremental revenue.”
Detection always lags behind threats. Shift to real-time prevention: enforce policies before risky behaviors occur, make secure collaboration easier than insecure alternatives, and use behavioral analytics to flag anomalies instantly.
Of course, strategy means little without execution. Here are three practical actions every security leader can take this quarter to move the needle.
Inventory all external sharing channels (email, SharePoint, OneDrive, Teams, portals, third-party tools). Identify top business processes involving external collaboration, map their data flows, and create a governance heat map to spotlight high-risk gaps.
Draft a one-page decision rights matrix, define risk tiers, and clarify stakeholder roles. Pilot the model with a high-risk use case and iterate quickly—don’t wait for perfect policies.
Select a process where security is seen as a bottleneck (e.g., sales collaboration). Measure before/after states, calculate ROI, and present results to stakeholders—demonstrating security as a business enabler.
As we navigate this new era, one thing is clear: unified external data governance is the foundation for success in the age of agentic AI. The window to act is closing fast. Retrofitting governance after autonomous systems are deployed will be exponentially more complex and costly; map data flows, build a governance model, and establish metrics this quarter.
If these trends resonate with you—or if you’re wrestling with similar challenges—I invite you to join Mark Cassetta and me in our upcoming webinar. We’ll unpack these topics, share practical strategies, and answer your questions about building resilient, future-ready security programs.

Gartner’s recent IT Symposium/Xpo revealed a fundamental shift in how CISOs must approach data governance, external collaboration, and AI risk management. The most striking insight is about timing and scale: by 2028, Gartner predicts 90% of B2B buying—over $15 trillion—will be intermediated by AI agents. Imagine procurement, finance, and vendor agents exchanging data and making decisions autonomously.
The governance challenge now extends beyond Shadow IT to what I call Shadow AI Agents: autonomous systems making data-sharing decisions outside traditional security boundaries.
After two decades as a CISO and Trust Officer across life sciences and technology, I’ve navigated cloud migration, SaaS proliferation, and remote work. But today’s challenges are unique: velocity is faster, complexity is higher, and the governance gap is wider than ever.
Here’s the central thesis: Gartner’s research on agentic AI, unstructured data security, and third-party risk management all converge on one critical capability—unified external data governance. Fewer than one in four organizations have built this foundation.
This analysis explores why unified external data governance matters, what Gartner’s research reveals, and what security leaders should prioritize in the next 90 days.
As we look ahead, three mega-trends stand out, each reshaping the security landscape in ways that demand our attention and action.
We’re on the verge of seeing AI agents automate procurement, customer support, and partner integrations, exchanging data and negotiating contracts with minimal human oversight. Our legacy controls, built for human workflows, won’t keep up. Unified visibility and governance over these autonomous data flows is no longer optional.
GenAI and agentic AI rely on unstructured data: documents, emails, recordings. Legacy DLP and CASB tools are reactive and limited, often missing critical external shares across platforms like SharePoint, Teams, and custom portals. Real-time, context-aware governance across all channels is now essential.
AI agents will interact with external partners across diverse regional platforms and regulatory frameworks, creating complex compliance and accountability challenges that traditional vendor risk assessments weren't designed to handle. Ownership of third-party cyber risk is often unclear, and traditional vendor assessments are insufficient. Real-time monitoring of data flows, and clear accountability are now required.
Recognizing these trends is only the first step. Building unified external data governance that can enable real AI value requires three foundational capabilities that most security organizations lack.
Perfect policies are the enemy of scalable AI governance. Segment AI use cases by risk tier and apply tailored controls. Build operating models and oversight systems that enable visibility and accountability for high-risk, externally facing AI collaborations. Focus governance on the AI portfolio, not every conceivable use case.
Move beyond technical metrics. Connect security initiatives to business outcomes—revenue acceleration, risk reduction, and competitive advantage. Frame achievements in terms executives understand: “Secure external collaboration reduced contract cycle time by 18%, enabling $4M in incremental revenue.”
Detection always lags behind threats. Shift to real-time prevention: enforce policies before risky behaviors occur, make secure collaboration easier than insecure alternatives, and use behavioral analytics to flag anomalies instantly.
Of course, strategy means little without execution. Here are three practical actions every security leader can take this quarter to move the needle.
Inventory all external sharing channels (email, SharePoint, OneDrive, Teams, portals, third-party tools). Identify top business processes involving external collaboration, map their data flows, and create a governance heat map to spotlight high-risk gaps.
Draft a one-page decision rights matrix, define risk tiers, and clarify stakeholder roles. Pilot the model with a high-risk use case and iterate quickly—don’t wait for perfect policies.
Select a process where security is seen as a bottleneck (e.g., sales collaboration). Measure before/after states, calculate ROI, and present results to stakeholders—demonstrating security as a business enabler.
As we navigate this new era, one thing is clear: unified external data governance is the foundation for success in the age of agentic AI. The window to act is closing fast. Retrofitting governance after autonomous systems are deployed will be exponentially more complex and costly; map data flows, build a governance model, and establish metrics this quarter.
If these trends resonate with you—or if you’re wrestling with similar challenges—I invite you to join Mark Cassetta and me in our upcoming webinar. We’ll unpack these topics, share practical strategies, and answer your questions about building resilient, future-ready security programs.

Balancing collaboration speed with strong governance is the top challenge. Features like Teams/SharePoint external sharing can create oversharing and audit gaps if unmanaged. Pairing Microsoft Purview with a guest-less external collaboration layer like eSHARE keeps data in-tenant, applies existing controls, and gives CIOs/CISOs the visibility they need without slowing work.
Balancing collaboration speed with strong governance is the top challenge. Features like Teams/SharePoint external sharing can create oversharing and audit gaps if unmanaged. Pairing Microsoft Purview with a guest-less external collaboration layer like eSHARE keeps data in-tenant, applies existing controls, and gives CIOs/CISOs the visibility they need without slowing work.
Balancing collaboration speed with strong governance is the top challenge. Features like Teams/SharePoint external sharing can create oversharing and audit gaps if unmanaged. Pairing Microsoft Purview with a guest-less external collaboration layer like eSHARE keeps data in-tenant, applies existing controls, and gives CIOs/CISOs the visibility they need without slowing work.