Shared AI agents expose governance gaps in European firms
A shift towards shared "multiplayer" AI agents is forcing European businesses to confront major governance blind spots as personal chatbots give way to collaborative digital workforces.
European businesses are shifting from individual AI chatbots to shared, collaborative "multiplayer" agents that work across entire departments. According to Gabriel Hubert, CEO and cofounder of AI company Dust, these "digital teammates" are moving beyond simply answering individual prompts to participating in shared company workflows.
In sales, for example, a shared agent can gather data, apply qualification criteria and update a CRM system, standardising a process that previously varied by employee. “The important part is that the workflow is shared, improves over time and becomes available to the rest of the team,” Hubert says.
This replaces the current standard where individual productivity gains do not scale across the organisation. “One person learns a better way to do something, but that improvement doesn’t necessarily spread,” he adds.
For European enterprises, this structural shift presents a pressing corporate risk. Data from Smarsh shows that while 55% of large EU enterprises use AI, only a quarter believe their governance models are fully equipped for implementation.
As agents become shared resources, traditional IT frameworks struggle to prevent "shadow AI"—the unauthorised use of tools that circumvent central oversight. The primary danger is data leakage, as an agent inherits the access rights of the workspace it was built on.
“The context and operating knowledge of a company should belong to the company. It shouldn’t be trapped inside a model provider, an individual conversation or a series of disconnected applications,” Hubert argues.
Dust’s approach assigns administrators to strictly manage which data sources agents can draw upon, ensuring the platform blocks restricted files from unauthorized staff. However, technology is still outpacing corporate structure. Microsoft research indicates that organisational factors like manager support and culture drive more than twice the AI impact of individual employee effort alone.
To bridge this gap, Dust relies on "AI operators"—specialised staff who redesign processes rather than just automating isolated tasks. Hubert describes these operators as running an "anti-to-do list" to eliminate repetitive work entirely.
By 2027, he expects firms to stop asking whether they should use agents and start focusing on how to manage an agent workforce. “As agents absorb more execution, judgment becomes more valuable. The asset is the loop between the agents, the context the company owns and the people who keep improving both.”
