AI legal – Harvey says AI agents are taking on more legal tasks, prompting firms to rethink how they staff and price legal work.
AI agents are starting to act like the next rung in the legal workflow, and that shift could ripple all the way to how law firms staff cases.
Harvey. an $11 billion legal software startup. says its latest push focuses on “agents” that can carry out tasks on a user’s behalf. moving beyond chat-style tools into more goal-driven work.. In Misryoum’s view. the big change is not just automation. but delegation: lawyers give instructions. and AI handles portions of the routine. time-consuming steps that have traditionally been distributed across associate and junior teams.. Harvey’s CEO argues that agent usage is already accelerating inside customer firms.
Misryoum insight: For law firms, the economics of legal work are tightly linked to time. When software reduces the hours required for drafting, research, or document review, the staffing model becomes a strategic question, not a simple productivity upgrade.
Harvey says it now has hundreds of agents live across major practice areas and that customers are running a high volume of agent-powered tasks each day.. The company also redesigned its Agent Builder so lawyers can tailor agents to specific workflows without writing code.. That matters because it lowers the barrier for firms to standardize how work gets done. potentially making it easier to roll out consistent processes across matters.
The company’s approach resembles how software engineering adopted AI copilots: instead of only generating text or code on demand. agents can follow a set of objectives and iterate toward an output.. In legal settings. Harvey frames these capabilities around defined goals such as drafting memos. supporting negotiations. or running diligence workflows. with human oversight built in.. The practical implication is that more of the work moves to the “front end” as lawyers spend more effort specifying what agents should do and reviewing what they produce.
Misryoum insight: Oversight may become more structured. As delegation shifts upward in the workflow, lawyers spend less time on routine execution and more time setting requirements, validating outputs, and managing risk.
Debate around the technology’s impact has also centered on pricing, especially in practices that historically bill by the hour.. Even with time savings. firms still face pressure to protect quality and manage liability. particularly when AI is reviewing large volumes of material and producing consolidated results.. Harvey says that as agents take on more complex tasks, verification processes and quality control become more important, not less.
To address performance and reliability. Harvey says it is working on standardized testing. often called evaluations. to measure how well agents perform across specific legal tasks.. It is also building quality-control agents designed to check the work of other agents.. That focus on testing and layered checks suggests firms will likely demand stronger proof of consistency before fully trusting delegated workflows.
Misryoum insight: The winners in this shift may be the firms that treat AI like an operational system. Clear task definitions, repeatable checks, and measurable performance can help convert faster legal production into predictable outcomes for clients.
Looking ahead, Harvey’s CEO does not predict that AI will eliminate legal roles.. Instead. Misryoum notes the core expectation is a staffing rebalancing: fewer lawyers may be assigned to each matter as AI absorbs more of the legwork. while firms could handle a larger number of matters overall.. For clients. the promise is quicker turnaround and more capacity as complexity grows; for firms. the challenge will be redesigning how teams work together when agents become an everyday part of the chain of delegation.