← Back to all posts

Key Takeaways:

The AI governance market just passed $1.5 billion in Q1 2026 funding alone. Everyone is building walls. Nobody is building the bridge.

The Rush to Govern

Something extraordinary happened in the first week of March 2026: over $275 million flowed into AI agent governance companies in a single week. JetStream Security raised $34 million backed by CrowdStrike's CEO. Guild.ai secured $44 million from Google Ventures and Khosla. ServiceNow acquired Traceloop for an estimated $80 million. WorkOS landed $100 million at a $2 billion valuation.

The message from the market is unmistakable: AI agents need governance, and the industry is willing to pay for it.

But here is what nobody is talking about: every major platform just built its own control plane. And they all stop at their own walls.

The Control Plane Explosion

In the past 90 days, four separate control planes have launched or been significantly upgraded:

ServiceNow AI Control Tower — Manages AI Specialists within the Now Platform. Tracks agent performance, enforces policies, handles escalations. Recently strengthened with Traceloop for observability and Veza for identity. Powers their Autonomous Workforce, which handles 90%+ of L1 IT tickets.

Microsoft Purview + Copilot Agent Dashboard — Monitors Copilot agents across Microsoft 365. DLP policies, auto-blocking, auto-deletion of inactive agents, governed Agent Store. Shipping end of March 2026.

UiPath Maestro — Brand new. A unifying control plane for AI agent governance. RBAC, PII detection, MCP servers as governed gateways, Agent Score for performance, fleet-wide dashboards. Full GDPR/HIPAA/FedRAMP compliance. Integrated with their AgentOps framework.

Salesforce Agentforce — Einstein AI agents with Trust Layer. Manages agents within the Salesforce ecosystem. $2.9 billion ARR and growing.

Each of these is genuinely impressive. Each solves real governance problems. And each operates within a single platform's boundaries.

The Enterprise Reality: Five HR Departments for One Workforce

Here is what actually happens inside a Fortune 500 company:

That is 45 AI agents managed by four different control planes with zero visibility across them.

When the CISO asks "how many AI agents do we have?" — nobody can answer.

When the CFO asks "what is our total AI agent spend?" — nobody can answer.

When the compliance team asks "are all our agents following our data handling policies?" — nobody can answer.

This is not a hypothetical. According to the AIUC-1 Consortium (Stanford, MIT, and 40+ CISOs), only 21% of enterprises have full visibility into their AI agent operations. The average enterprise runs 1,200+ unofficial AI applications. Shadow AI breaches cost $670,000 more than governed ones.

Control Plane vs. Control Tower

The aviation industry solved this exact problem decades ago.

Every airline has its own operations center — a control plane for its fleet. American manages American flights. United manages United flights. Delta manages Delta flights. Each one excellent at what it does.

But air traffic control — the control tower — sees every aircraft in the sky, regardless of airline. It manages the shared airspace. It prevents collisions. It ensures the system works as a whole.

A control plane manages what is inside. A control tower manages what is across.

Control PlaneControl Tower
ScopeSingle platform's agentsAll agents across all platforms
IdentityPlatform-specific credentialsUniversal agent identity
PoliciesVendor's policy frameworkOrganization-wide governance
VisibilityOwn agents onlyComplete agent inventory
Cost trackingPer-platform spendTotal AI workforce cost
CompliancePlatform-specific auditCross-platform compliance posture
Who builds itServiceNow, Microsoft, UiPath, SalesforceNobody. Yet.

Why Nobody Has Built It

The answer is simple: platform vendors have no incentive to build cross-platform governance. ServiceNow wants you to run everything on ServiceNow. Microsoft wants you in Azure. UiPath wants you in their ecosystem. Salesforce wants you on Agentforce.

Cross-platform governance is structurally opposed to their business model. It is like asking American Airlines to build air traffic control that also routes United flights.

The security vendors — JetStream, Zenity, ArmorCode, Teramind — are closer. They can see across platforms. But they are security cameras, not HR departments. They detect threats. They do not manage performance, track costs, enforce business policies, or help agents improve.

The market has created three layers:

  1. Platform-native governance (ServiceNow, Microsoft, Salesforce, UiPath) — excellent within walls
  2. Infrastructure governance (Bifrost, Kong, Cloudflare) — routing and monitoring plumbing
  3. Security surveillance (JetStream, Zenity, Teramind, ArmorCode) — threat detection

What is missing is the fourth layer: cross-platform workforce management. The control tower.

What a Control Tower Actually Does

An AI workforce control tower would:

See every agent. Regardless of whether it runs on ServiceNow, Azure, UiPath, Salesforce, or custom infrastructure. One inventory. One source of truth.

Enforce universal policies. Data handling rules that apply across all platforms. Escalation thresholds that are consistent. Compliance requirements that do not depend on which vendor built the agent.

Manage identity centrally. Today, 45.6% of enterprises use shared API keys for agents. A control tower gives each agent its own identity, tracks its permissions, and enforces least-privilege across platforms.

Track total cost. Not ServiceNow cost plus Microsoft cost plus UiPath cost reviewed in three separate dashboards. One view: "Our AI workforce costs X, delivers Y, ROI is Z."

Enable performance management. Which agents are performing? Which need retraining? Which should be retired? This is HR for AI — the job that no security camera can do.

ServiceNow's CEO Made This Argument for Us

Here is the irony. At ServiceNow's recent earnings call, CEO Bill McDermott said:

"Without cross-enterprise workflows, so-called agentic AI is just another one-dimensional chatbot."

He is right. And his platform — despite being the most aggressive single-platform AI strategy in enterprise software — still only governs agents within ServiceNow's walls.

Industry analysts confirm this. Techzine reports "limited visibility into third-party agents." CIO notes ServiceNow "struggles in disconnected systems." Diginomica finds "credibility tested beyond idealized service desks."

McDermott sees the problem clearly. His platform, by design, can only solve half of it.

The Math

The numbers make the case:

Organizations that deploy governance see 12x more agents reach production. That is not a governance tax — that is a governance multiplier. The control tower does not slow things down. It speeds them up.

The Market Is Telling You Something

When $1.5 billion flows into agent governance in a single quarter — from security to orchestration to identity to observability — the market is screaming that this problem is real.

But look at where the money went: each investment funds one layer. One piece of the puzzle. Nobody funded the cross-platform unifier because the venture market, like the vendor market, thinks in categories.

The enterprise does not think in categories. The enterprise thinks: "I have 45 agents across 4 platforms and I need one place to manage all of them."

That is not a control plane. That is a control tower.

Sources:

Ready for cross-platform AI governance?

iEnable is building the AI workforce control tower — cross-platform governance, identity, and performance management for every AI agent in your organization, regardless of where it runs.

Learn More About iEnable →