Upcoming Feature
Link Custom Agents
Linking Custom Agents lets lawyers chain existing agents together to produce repeatable, multi-stage outcomes — turning individual agents into building blocks for more complex, end-to-end workflows.
What's Coming
Builders can now reference an existing custom agent inside another agent's instructions using the "@" tag, enabling agents to chain together and produce outcomes across extended, multi-stage tasks. This makes complex orchestration possible without rebuilding logic from scratch.
Feature Highlights
- Chain custom agents together by tagging "@" an existing agent inside another's instructions
- Turn individual agents into reusable building blocks for multi-stage workflows
- Produce repeatable, end-to-end outcomes across drafting, analysis, and review tasks
- Supports complex orchestration across litigation, transactional, and in-house use cases
FAQs
Q: How does bundling work technically?
When building a new custom agent, a builder tags "@" an existing custom agent in the instructions. The new agent can then reference and trigger the tagged agent as part of its workflow.
Q: Are there any limitations at launch?
Yes. Multi-level bundling is currently not supported — if Agent B is already linked to Agent C, Agent A cannot successfully tag Agent B.
Q: What kinds of workflows does this unlock?
Bundling supports complex, multi-stage tasks across practice areas — from chaining complaint analysis into answer drafting for litigators, to assembling full financing packages for EC/VC teams, to streamlining lease review workflows for in-house counsel.