AIR raises $50M to govern corporate AI agents

The $50 million round and the problem AIR wants to solve

On September 1, 2026, AIR announced the closure of a $50 million funding round. The news, reported by TechCrunch , highlights a still underserved segment: the governance of AI agents in business contexts. In fact, as organizations accelerate the adoption of autonomous systems, control over what these systems do often remains rudimentary.

The AIR platform operates on three distinct levels. First of all, it performs a phase of discovery automatic, identifying all active AI agents in the company infrastructure. Subsequently, it initiates a continuous process of vetting of the skills and add-ons that each agent uses. Finally, it is able to block in real-time any behavior that goes beyond authorized parameters.

This approach addresses a real need. Many SMEs and mid-market companies have started integrating AI agents into their operational workflows without adequate tools to monitor them. Consequently, the risk of non-compliant—or simply inefficient—drifts is high and often underestimated.

Why the proliferation of AI agents has become an operational risk

Enterprise AI agents are no longer a niche thing. The question isn't whether to use them, but how to keep them in check.

The core issue is the modular nature of these systems. An AI agent isn't monolithic software: it integrates third-party skills, plugins, and add-ons that can change over time. Every add-on update potentially introduces new behaviors, new access to sensitive data, new interactions with external systems. However, most companies lack visibility into these changes.

Beyond this, the European regulatory context is evolving rapidly. The AI Act imposes transparency and control obligations on high-risk AI applications. Therefore, having a governance layer like the one offered by AIR is not just an operational choice: in some sectors, it could become a compliance requirement.

How continuous skill and add-on vetting works

The core of AIR's offering is the vetting continuous. Unlike static approaches — which check an agent's compliance only at deploy time — AIR persistently monitors every active skill and add-on. This means that if an add-on is updated by the vendor and introduces new unauthorized functionalities, the system detects it and can intervene.

The phase of discovery is equally relevant. In medium-sized companies, it's common for different teams to activate AI agents independently, without a centralized census. Therefore, the ability to automatically map the entire ecosystem of active agents is a fundamental prerequisite for any governance strategy.

The blocking of unwanted behaviors, finally, happens in real-time. This aspect differentiates AIR from traditional audit tools, which operate ex post. Similar to what happens in IT security systems with the concept of zero trust , AIR applies a principle of continuous verification rather than implicit trust.

Implications for Italian SMEs' marketing and digital teams

For marketing and digital managers, this development has concrete implications. Teams already using AI agents for tasks like content generation, campaign management, or data analysis need to ask themselves: who controls what these agents do? What data do they interact with? What add-ons do they use and who authorized them?

We at SHM Studio we observe that this awareness is still limited among medium-sized Italian companies. Often, AI tools are adopted quickly and pragmatically, without an adequate governance framework. However, as these tools become more pervasive, the risk of incidents — data loss, non-compliant behavior, operational errors — grows proportionally.

In particular, those managing google ads campaigns or LinkedIn campaigns with the support of AI agents should carefully check what permissions these systems have on advertising account data. Similarly, those who use agents for activities of SEO copywriting should ensure that generated content complies with company editorial guidelines.

The market signal: AI governance becomes a standalone sector

A 50 million dollar round is no small signal. It shows that investors see AI governance as a market with its own growth path, separate from AI development tools or traditional security layers.

Furthermore, the timing is significant. The AIR round comes at a time when major agentic platform vendors — from Microsoft to Salesforce — are rapidly expanding their offerings. Consequently, the number of active AI agents in companies is set to grow further in the next 12-18 months. Therefore, those in governance will have an ever-expanding market to serve.

For companies, this means that the choice of AI governance tools should happen now, not when the problem is already evident. Waiting means accumulating a control debt that is difficult to repay retroactively.

What no one is saying yet: the hidden cost of unmonitored agents

There's one aspect that rarely comes up in discussions about AI agent adoption: the hidden cost of unmonitored behaviors. It's not just about security or compliance risks. It's also about operational inefficiencies that quietly accumulate.

An AI agent using an unoptimized add-on can generate lower-quality output, consume excess computational resources, or interact with external APIs in an unauthorized manner. However, without a continuous monitoring system, these issues only emerge when the impact is already measurable — and often costly.

Therefore, AIR's proposal is not just a matter of security: it is a matter of operational quality. For those responsible Digital marketing that entrust increasingly relevant parts of their value chain to AI agents, this distinction is fundamental. Investing in governance means investing in process reliability, not just risk protection.

What to do now: three priorities for marketing managers

In light of these developments, it is possible to identify three operational priorities for those managing marketing teams with AI components.

• Inventory active agents. First of all, you need complete visibility into all AI agents in use, including those adopted autonomously by individual teams. Without an up-to-date inventory, any governance strategy is baseless.
• Define authorization policies for skills and add-ons. Furthermore, it is necessary to establish which skills and add-ons are authorized for each agent, with a formal approval process for any changes. This reduces the risk of unintentional deviations.
• Evaluate continuous monitoring tools. Finally, it is advisable to explore solutions such as the one offered by AIR or equivalent, integrating them into your architecture of AI adoption . Ex-post monitoring is not sufficient in a context of autonomous agents.

To learn more about how to structure a responsible AI adoption strategy, the team at SHM Studio is available for dedicated consulting. You can contact us through the contact page or explore the full range of our digital services .

Outlook: where the AI governance market is headed in 2027

Looking ahead to the next 12-18 months, it's reasonable to expect a consolidation of the AI governance market. Similar to what happened with cybersecurity in the 2010s, AI governance will transition from a specialized niche to a structured corporate function. Therefore, vendors positioning themselves in this space today — like AIR — are building a competitive advantage that will be difficult to replicate.

For Italian companies, the risk is adopting these tools late, as has often happened in the past with other software categories. However, the regulatory pressure from the European AI Act could accelerate adoption timelines compared to other markets. In this sense, the regulatory context could paradoxically represent a competitive advantage for organizations that prepare in time.

We at SHM Studio we will continue to monitor the evolution of this market and share operational analyses on our Blog . AI agent governance is a topic destined to become central in the agendas of those responsible Digital and SEO in the coming quarters.