The new system automatically deploys coding agents for tasks like bug detection, security reviews, and incident response—reducing the need for constant human oversight.
Cursor Introduces Automation Layer for AI Coding Agents
As agentic coding tools become more powerful, software engineers increasingly find themselves supervising dozens of AI coding agents at once.
To manage that complexity, Cursor has launched a new system called Automations, designed to automatically trigger coding agents inside its development environment.
Instead of manually prompting AI agents, engineers can now set up automated workflows that activate when specific events occur.
Examples include:
- A new code commit
- A Slack message
- A scheduled timer
The goal is simple: reduce the growing attention burden on developers managing multiple AI processes.
Moving Beyond the “Prompt-and-Monitor” Model
Most current AI coding tools rely on a prompt-driven workflow.
Developers launch agents manually and then monitor their outputs.
Cursor’s Automations framework flips that model.
Instead of humans initiating tasks, agents run automatically and only involve engineers when needed.
Cursor engineering lead Jonas Nelle described the concept as an AI production line.
“It’s not that humans are completely out of the picture… They’re called in at the right points in this conveyor belt.”
This approach aims to make AI agents persistent collaborators rather than tools triggered one prompt at a time.
From Bug Detection to Incident Response
Cursor has already tested the concept internally.
One early version of the idea is Bugbot, a feature that automatically reviews newly submitted code.
Every time a developer adds code, the system triggers an agent to:
- Scan for bugs
- Identify potential issues
- Suggest improvements
With the new Automations system, that capability expands into more complex tasks.
Examples include:
- Security audits that analyze deeper vulnerabilities
- Incident response agents triggered by PagerDuty alerts
- Automated queries of server logs via MCP connections
In some cases, AI agents can immediately begin diagnosing infrastructure issues before a human engineer intervenes.
Automating Development Operations
Cursor says its system already runs hundreds of automations per hour across its internal infrastructure.
Beyond debugging and security, Automations are also used for operational tasks.
For example:
- A weekly agent generates Slack summaries of codebase changes
- Agents review system activity logs
- Models analyze code updates for performance implications
In theory, anything an engineer could manually ask an AI agent to do could instead be triggered automatically by system events.
A Heated Race in AI Coding Tools
Cursor’s launch comes during intense competition in the agentic coding market.
Major AI players including OpenAI and Anthropic have recently released upgrades to their own AI coding agents and developer tools.
Despite the competition, Cursor continues to hold a strong position.
Data from Ramp suggests that about 25% of generative AI customers use Cursor in some capacity.
The company’s growth has been rapid.
According to Bloomberg, Cursor’s annual recurring revenue recently surpassed $2 billion, doubling in just three months.
The Next Stage of AI Development Workflows
The rise of automation systems like Cursor’s reflects a deeper shift in software development.
Developers are no longer just writing code.
They are increasingly orchestrating networks of AI agents that write, test, and maintain software autonomously.
That raises an interesting question:
If AI agents can automatically review code, respond to outages, and audit security—does the developer become more like a systems supervisor than a programmer?
Cursor appears to be betting that future engineers will spend less time coding—and more time managing AI-driven workflows.
TL;DR
Cursor has launched Automations, a new system that automatically triggers AI coding agents based on events like code commits or Slack messages. The tool helps engineers manage dozens of AI agents simultaneously, handling tasks like bug detection, security audits, and incident response.
