Meta is reportedly moving its internal communications from Google Chat to Slack because of its growing focus on AI agents. The decision reportedly comes from the company’s belief that Slack offers a stronger environment for AI agents to interact with people, tools and workplace systems.
According to a memo from Meta’s AI chief Alexandr Wang, Slack is the strongest platform available today for agents. The memo points to Slack’s conversational interface, developer tooling and extensive network of third-party integrations as major advantages.
That last point is particularly important. Meta did not simply choose another workplace chat application; it chose a platform with an ecosystem that already connects to thousands of other tools. For AI agents, those connections could become more important than the messaging interface itself.
The decision also represents a significant win for Salesforce, which acquired Slack for $27.7 billion in 2021. Meta and Salesforce have not commented publicly on the reported move, while Google has also not responded to requests for comment.
AI agents are changing what workplace software needs to do
Traditional workplace software was designed primarily for humans. Employees open an application, search for information, send messages and manually move between different services to complete tasks.
AI agents could change that model. Instead of simply helping employees inside one application, agents could potentially move between multiple systems, retrieve information, execute actions and coordinate workflows on a user’s behalf.
That makes integrations increasingly valuable. A platform with thousands of existing connections gives agents more systems to interact with without requiring every workflow to be built from scratch.
This is where Slack’s position becomes interesting. Its value may increasingly depend not only on how well humans communicate through it, but also on how effectively AI agents can use the platform as part of larger workflows.
Europe is taking a different approach
The shift at Meta also highlights a growing divide between the priorities of major technology companies and European governments. While companies are increasingly looking for interconnected ecosystems that can support AI agents, governments are placing greater emphasis on digital sovereignty and control over their technology infrastructure.
France, for example, has been moving government organizations toward sovereign alternatives to major American technology platforms. The country has set a 2027 deadline for replacing services such as Microsoft Teams and Zoom across its ministries.
Germany has taken similar steps in parts of the public sector. Schleswig-Holstein moved thousands of mailboxes away from proprietary software, while other European institutions have adopted open-source communication and productivity tools.
The reasoning is fundamentally different from Meta’s. Governments are asking who controls the infrastructure, where sensitive data is stored and whether they can remain independent from major technology providers.
The agent ecosystem problem
That creates an interesting problem for Europe’s sovereign technology strategy. Open-source and sovereign platforms can give governments greater control over their infrastructure, but many of these systems do not yet have the enormous ecosystem of integrations that platforms such as Slack have developed.
That gap could become more important as AI agents become more capable. An agent does not just need somewhere to communicate; it needs access to the systems where information lives and where actions can be performed.
This does not necessarily mean sovereign platforms will lose. Governments could eventually build their own agent ecosystems around open standards and open-source infrastructure. But doing so will require more than simply replacing one messaging application with another.
The bigger question is whether the next generation of workplace software will be judged by how independent it is, or by how many things AI agents can actually do with it.
Meta appears to be betting on the latter. Europe, at least for now, is betting heavily on the former.
And that difference could become one of the most important technology debates of the AI-agent era.
