AI agents are becoming real users of software. Here’s what changes when products need to support humans and agents working together on shared data in realtime.
Join thousands of companies using Liveblocks to build multiplayer experiences for people and agents.
Today, we’re kicking off Unveil Week with a broader vision than human-only collaborative features. Software itself needs to change for the age of agents.
For a long time, software was designed for humans acting independently. For most products, it worked well enough; a user clicked a button, changed some data, and moved on. Maybe another person would pick it up later. But underneath it all, the product was still built around humans acting sequentially.
Then, products like Figma, Notion, and Google Docs introduced a different model—one that revealed how powerful software becomes when multiple people can work together. No more late nights merging v1 with v2, only to realize v3 has the latest changes. They made collaborative software mainstream, and thus realtime infrastructure with synced state became foundational for multiplayer experiences.
The shift in AI from question-answering chatbots to contributing agentic collaborators materially changes what it means to build software that works. AI Agents are real users of software, in many cases they’re becoming the main users. The ROI of investing in AI solutions isn't just a genie answering questions, it’s an army of specialized users, purpose-built to get work done more efficiently than humans alone.
They won’t just answer questions. They’ll update records, edit documents, trigger workflows, leave comments, and coordinate work across systems. And the scale will be massive. There may soon be far more agents using software than people.
Whether you think the number is 10x or 100x… we’re going to have some order of magnitude more agents than people.Aaron LevieCEO at Box
That’s the shift. Realtime infrastructure with sync was already core infrastructure for any products built for human collaboration. Now, it’s becoming core infrastructure for all software, because agents multiply the number of entities that can interact with shared data in parallel.
When people and AI agents work in the same product, three challenges show up immediately: concurrent updates, coordination, and visibility. These were already important in collaborative products. But as agents become users of software too, they become much more universal.
People and agents update the same data at the same time. A user edits a document while an agent updates a section in the background. A workflow runs while someone else reviews the output. Another agent leaves a comment before the first one finishes. Without the right infrastructure, changes get overwritten, lost, or applied out of order.
That is why sync matters so much. In the AI era, software needs to handle many actors working on shared data in parallel, not just one user at a time.
AI works best when it feels like a collaborator, not a black box. Users need to trigger actions in context, review what happened, and stay in control. The workflow needs to happen inside the product, not across disconnected tools and hidden backend jobs.
Once AI starts taking action instead of just generating text, coordination becomes a product requirement. The experience needs to make it clear how work moves between people, the agent, and the overall system.
As soon as multiple actors are active at once, visibility becomes part of the experience. Who is active? What changed? What is the agent doing right now? What finished? What still needs review? Users need answers to those questions in realtime, where the work is happening.
That is what makes AI feel trustworthy. The more active agents become inside products, the more important it is that their work feels visible, understandable, and easy to follow.
These are not edge cases anymore. They are becoming foundational product requirements for any software where people and AI work together on shared data. That is what Unveil Week is about. Over the next five days, we’ll be sharing the building blocks we’re releasing for products where people and AI work together in realtime.
Stay tuned for more updates throughout the week.
