AI Agents
Who Will Control the AI Agent Ecosystem?
Don't just read what happened. See what could happen next.
Prediction
Control fragments across OS/browser runtimes, productivity suites, and lab platforms — no single permanent monopolist in 24 months.
- Confidence
- 48%
- Horizon
- 12–36 months
- Impact
- High
- Direction
- Increasing
Prediction changed +8 points in 21 days
Short answer
Agents need identity, tools, data, and distribution. Past-year moves show Microsoft, Google, OpenAI, and Salesforce all trying to own the runtime. Expect federated control with painful integration seams.
Why this question matters
Ecosystem control sets switching costs, data gravity, and which startups can exist independently.
What's happening now?
Suite-native agents
Office, Workspace, and CRM embed agents where work already happens.
Signal · strong
Lab-hosted agent platforms
Model vendors ship app stores and tool protocols.
Signal · strong
OS and browser agent hooks
Computer-use features push control toward the client platform.
Signal · moderate
Protocol experiments
Open tool/agent protocols try to prevent hard lock-in.
Signal · weak
What could happen next?
Scenario A
Suite winners
Work software owns agents; labs become model suppliers.
Scenario B
Lab platforms win
ChatGPT-class hubs become the agent app stores.
Scenario C
Interoperable mesh
Open protocols keep multiple controllers viable.
Key companies / entities
- Microsoft
- OpenAI
- Salesforce
- Apple
- Amazon
Evidence
- company
Agent platform and store announcements
- news
Enterprise agent architecture write-ups
- research
Open agent protocol proposals
Prediction history
- Sep 22, 202648%
- Sep 15, 202645%
- Sep 8, 202643%
- Sep 1, 202640%