Anthropic has just published over 90 named AI agents as part of Claude for Legal, all available on GitHub, each configured for a specific legal workflow. Not a generic contract review tool, nor a research assistant. Agents with names like DSAR Responder, Termination Reviewer, Claim Chart Builder, each designed to run a defined task from start to finish, and in some cases to keep running continuously on incoming work.
That distinction matters more than it might appear. A tool responds when you ask it to, whereas an agent acts on your behalf, within the parameters you have set. It watches an incoming document stream, flags deviations from a playbook, sweeps signed agreements each week, and brings you what needs your attention. The lawyer still makes every consequential decision. The agent does the groundwork to make that decision faster and better informed.
What Anthropic has done is lower the barrier to deploying that kind of capability significantly. These agents can be modified in natural language. You do not need to be an engineer to configure one for your practice area, your firm’s precedents, or your clients’ specific commercial requirements. The more precisely an agent is tuned to a particular workflow, the more useful and accurate it becomes.
For a junior lawyer, this changes what it means to be productive at work. The ability to deploy, supervise, and take responsibility for agents operating in your name is becoming a distinct professional skill, and one that relatively few lawyers currently have. Firms investing heavily in this infrastructure will want people who can work with it from day one, not people still finding their footing with tools the firm has already deployed.
The technical knowledge required to get started is lower than most people assume. The professional judgment required to use these tools well is exactly the kind junior lawyers are already developing.
AI Legal Recruitment works with lawyers who are building both.