036 - What Is a Vertical AI Agent?

A general AI assistant can summarize a contract in seconds. Getting it to apply a law firm's own playbook, follow jurisdiction-specific rules, and work inside the systems where a case lives is a different problem, and closing that gap is what vertical AI agents are built for. This episode explains what a vertical agent is, how it differs from a general-purpose model like Gemini, Claude, or GPT, and why the specialization increasingly lives in the system wrapped around the model rather than in the model itself. The episode walks through Google's Gemini Enterprise for Legal, announced August 25, 2026, as a detailed example of that architecture: purpose-built skills for contract review, regulatory tracking, and legal research, secure integrations with tools lawyers already use including iManage, NetDocuments, and e-discovery platforms, and permission handling that carries over a firm's existing access controls instead of flattening them. It also covers Google's privacy claim that client data and playbooks are never used to train its foundation models. Three other companies illustrate the same pattern outside law. Ambience Healthcare builds its platform around reconciling medical records and generating specialty-accurate clinical documentation and coding. EvenUp positions itself as the leading AI platform for personal injury law firms, covering case work from intake through trial. Harvey describes its product as legal AI for law firms and corporate legal teams, spanning contract analysis, due diligence, compliance, and litigation. The episode explains why narrow, workflow-specific AI can create measurable business value, since automation pays off most when it's applied to an entire process rather than a single question, while noting that vertical AI products aren't automatically protected from competition, because foundation model providers are building their own industry-specific tools. Listeners will come away understanding that calling a system a vertical agent says nothing about whether its underlying model was trained specifically for that industry; the specialization usually comes from retrieval, instructions, tool connections, permissions, and evaluations layered around a general-purpose model. Useful for anyone trying to understand how AI agents are being deployed inside regulated, workflow-heavy industries like law, healthcare, and finance. Sources & References Introducing Gemini Enterprise for Legal, Google Cloud Blog — https://cloud.google.com/blog/products/ai-machine-learning/introducing-gemini-enterprise-for-legal Ambience Healthcare — https://www.ambiencehealthcare.com/ EvenUp Law — https://www.evenuplaw.com/ Harvey — https://www.harvey.ai/ Voice narration is AI-generated.