Skip to content
Fast Company Executive Board

The Fast Company Executive Board is a private, fee-based network of influential leaders, experts, executives, and entrepreneurs who share their insights with our audience.

Your AI governance strategy may be stuck in the desktop era

Reliance on traditional security controls creates a massive blind spot.

[Photo: Getty Images]

Profile photo of Jim Dolce

CEO · Lookout

The modern workplace has undergone a fundamental shift, driven by the simultaneous rise of artificial intelligence (AI) and the widespread adoption of mobile devices.

In the U.S., IDC’s recent Resource Center & Analyst Forecast Report notes that 93% of employees spent at least a third of their time away from a desk, relying on mobile devices as their primary interface for personal and business email, text, calls, and work. In this mobile-first, AI-enabled reality, smartphones are the new operational nerve centers of the modern enterprise.

Yet a critical disconnect lies at the heart of this transformation. Despite rapidly embracing AI to boost productivity, organizational governance models—the policies and systems an organization uses to monitor and secure its data and employee behaviors—remain rooted in the past.

These tools, meant to drive efficiency, create a growing, largely invisible risk that many IT security and risk teams are ill-equipped to manage.

The rise of autonomous agents

In essence, centralized security checkpoints like firewalls acted as digital security guards that could open every package and inspect every line of text leaving the office network. Today, mobile devices and cloud applications completely bypass those checkpoints, rendering an increasing share of AI activity virtually invisible. 

Organizations believe they have comprehensive AI oversight, yet a significant portion of employees’ real-world usage remains unseen and unmanaged. For example, an employee might use their phone to snap a photo of a confidential presentation slide or whiteboard strategy session, then upload it to a free mobile AI assistant to quickly write a summary report.

The business risk intensifies as we enter the agentic AI era. Unlike generative AI, agentic systems are “autonomous actors” that plan, decide, and execute multi-step workflows independently.

According to Gartner’s strategic technology trend forecasts, agentic usage is growing rapidly, with more than 40% of enterprise applications expected to include task-specific AI agents this year. These systems can autonomously initiate communications, trigger financial transactions, and modify records without human oversight.

Agentic AI fundamentally changes the mobile risk equation. When embedded in mobile environments, these agents operate with full user authority, leveraging corporate access, authenticated sessions, multi-factor authentication, and third-party cloud applications linked directly to an organization’s main corporate account.

Because a mobile phone consolidates identity, access, and data into a single, always-on interface, it becomes the ideal launch point for autonomous action. The result is a new attack surface: a single gap on a mobile device can enable data theft, privileged access to application program interfaces (APIs), and manipulation of business processes at machine speed, beyond the reach of traditional, desktop-centric controls.

Compliance is no longer a choice

For executive leadership, compliance is no longer a matter of choice; it’s a regulatory mandate. For example, the European Union’s Artificial Intelligence Act (EU AI Act) is the first comprehensive, legally binding framework for AI.

Global standards such as ISO/IEC 42001 and the NIST AI Risk Management Framework further reinforce the need for formalized AI governance.

These frameworks require complete visibility, control, and traceability across AI usage, including activity on mobile devices. Organizations must be able to follow the complete, step-by-step history of a piece of data, a process, or a product from its very beginning to its final destination.

A company overlooking the mobile dimension can appear fully compliant on paper yet remain exposed on the devices that employees use daily. Small, often unseen gaps compound into material regulatory and legal risk, leaving enterprises vulnerable to non-compliance, data protection failures, and an inability to scale AI securely and with confidence.

The true cost of ignoring mobile-native security is an immediate operational bottleneck. When an employee on the road uses an unmanaged, free mobile AI app to create a quick summary of a confidential project strategy, that intellectual property is permanently cast into a public database.

The moment that data exposure comes to light, the corporate dominoes fall fast. The enterprise faces crippling regulatory fines for privacy violations, followed by an immediate “innovation freeze” in which panicked leadership bans the use of all generative tools.

New approaches

As a result, organizational solutions must continuously monitor personal-phone mobile AI activity and generate audit-ready, standards-aligned evidence. This “audit-ready evidence” is a clean, automated digital receipt proving that a company’s data guardrails are working, not an invasive log of an employee’s personal text messages or photos.

The strict privacy-by-design model on personal devices ignores private apps and personal web browsing entirely. Instead, it triggers an audit log only when corporate data interacts with an AI tool—generating a tamper-proof record that shows, for example, that an encrypted corporate document was blocked from being uploaded to an unapproved public LLM.

Compliance regulators can review the exact, immutable proof of data protection required, while keeping the employee’s personal life entirely private. The approach enables clear, defensible oversight, closes visibility gaps, and supports confident compliance with evolving regulatory requirements.

As another example, a lightweight guardrail operating directly on the mobile device can help identify employee use of AI applications, monitor autonomous behavior, and enforce data security.

For example, imagine an employee attempting to copy and paste sensitive financial figures from a corporate app into an unauthorized mobile AI chatbot. The system instantly intercepts the action at the screen layer, blocks the transfer, and displays a friendly notification suggesting an approved enterprise alternative.

Simultaneously, the solution works in the background to catalog every hidden AI software library embedded within downloaded employee apps, automatically cutting off an app’s access to corporate cloud accounts if it detects an autonomous AI agent attempting to silently harvest data over a cellular connection.

A new leadership mandate

AI is increasingly seen as an unstoppable force, so impractical attempts to ban it or restrict employee mobile access stifle the innovation that organizations depend on. The real solution is to modernize governance.

As traditional security walls collapse, organizations must extend visibility into mobile devices, where the real work happens. In an AI-powered world, the greatest threat is the governance gap that leaves mobile AI unseen and unmanaged.

Jim Dolce is the CEO at Lookout.