The Pattern
A workflow automation platform called Vesence recently described something that practitioners have started to recognize: users are not just asking AI agents to help with legal tasks. They are asking AI agents to build the legal tech they wish existed.
One Vesence user automated an entire FDI compliance workflow — filling PDF checkboxes, populating form fields, redacting sensitive data across templates — not because Vesence had shipped that feature, but because the user described the task and the agent wrote the code. The agent inspected the PDF structure, identified form fields and annotation layers, wrote a manipulation script, and delivered the output. The script was then saved as a reusable skill.
This is not isolated. The pattern keeps repeating across legal tech deployments: coding agents serve as a customization layer, extending purpose-built legal tech in ways that vendors have not yet productized.
Why This Is Different From Previous Automation
Legal professionals have always been able to automate workflows — VBA macros in Word, document assembly templates, automated deal closing sets. The difference with coding agents is the interface: instead of requiring a programmer to translate legal needs into code, the agent handles the translation. A lawyer can describe a workflow in plain language and have working automation in minutes.
The composability implication is significant. The traditional constraint on legal tech was that each tool sat in a silo. Coding agents can compose across these silos because they can write scripts that orchestrate multiple tools via their APIs. The result is custom workflows that no vendor has built, assembled on demand.
What Legal Teams Need to Do
The legal teams seeing the most value from coding agent deployments share a common trait: they invested early in understanding their own workflows deeply enough to describe them to an AI agent with precision. Vague descriptions produce vague automations. Specific descriptions of specific workflows produce specific, deployable code.
The operational implication for law firms and in-house teams: the highest-leverage AI investment may not be in new tools but in the skillset to extend existing tools. A legal operations professional who can effectively prompt a coding agent to extend a CLM, connect a research tool to a template engine, or automate a closing checklist is generating value that no vendor can match. The gap between firms that understand this and firms waiting for vendors to ship the features they need is growing.