May 2026 has been an eventful month for legal AI, and we view the recent momentum as a positive and encouraging development.
Anthropic launched Claude for Legal with 12 practice-area plugins, 20+ MCP connectors, and integrations designed for legal research and workflow management. OpenAI is building its own legal offering. Gemini has also been appearing in legal workflows with results that are increasingly difficult to ignore.
The market has finally caught up to where in-house legal teams have been heading for years. The question now is how much of that potential actually translates into real-world legal and compliance outcomes.
For one of our clients, a major tech company, the answer was a 99.5% reduction in the volume of content requiring human review when responding to a legal matter. That came from pairing advanced legal AI with workflows built specifically for how legal and compliance operations actually function.
Claude for Legal is Anthropic’s most significant push into the legal industry to date.
The platform includes 12 practice-area plugins covering Commercial, Employment, Privacy, Corporate, AI Governance, and more. Each plugin is designed around a specific legal role and workflow.
When legal teams implement Claude for Legal, the platform conducts onboarding interviews to learn how the organization works, including playbooks, escalation thresholds, governance policies, and preferred drafting styles. The goal is for Claude’s outputs to reflect the practices of the legal department rather than producing generic AI-generated responses.
Claude for Legal also integrates with a growing ecosystem of legal research, document management, and workflow tools through MCP connectors. This gives legal teams access to court opinions, internal repositories, operational data, and workflow automation directly within their legal AI environment.
That said, every legal AI deployment eventually runs into the same issue: does the model understand how legal workflows actually operate in practice?
The GDPR response clock starts the moment a request arrives. A PCAOB production deadline may allow only 14 days to preserve and produce evidence. A FINRA examination may require preserving the exact state of a web-based platform at a specific moment in time in a defensible format.
The same Slack message may simultaneously need to be preserved under FINRA Rule 4511, included in a litigation hold, reviewed for DSAR obligations, and assessed for privilege and proportionality, all under different regulatory requirements.
This is where the workflow layer matters.
A legal AI assistant like Claude for Legal can generate summaries, and draft legal content. But defensible legal operations require much more than generation capabilities.
Legal teams still need to connect to enterprise data sources, apply governance and access controls, freeze collection scopes immediately, preserve dynamic collaboration and web content, maintain chain of custody, protect privileged information, and produce responsive material under FRCP 26 and FRCP 34 requirements. FRCP 37 becomes the consequence when those workflows fail.
That’s why the success of Claude for Legal depends not only on the model itself, but also on the workflows, and institutional knowledge surrounding it.
“Domain knowledge” gets used loosely in legal tech conversations. In practice, it means understanding the operational realities behind legal and regulatory obligations.
A dynamic website, personalized client portal, revised privacy notice, or social media profile may require active preservation that goes far beyond screenshots or PDF exports. FINRA and PCAOB obligations may also require WORM-compliant preservation and defensible collection processes.
Responding to an employee DSAR during an active employment dispute creates overlapping obligations involving data privacy, litigation risk, preservation requirements, and privilege review. Handling those workflows separately is often where problems begin.
This kind of institutional knowledge is built through years of real-world investigations, audits, regulatory examinations, and litigation matters.
When combined with platforms like Claude for Legal, that expertise allows legal and compliance teams to reduce manual review volumes, accelerate regulatory and legal response times, make earlier decisions about scope and exposure, improve defensibility, and scale legal operations more efficiently.
The AI model provides automation power. The workflow and regulatory intelligence determine whether the output is usable when it matters most.
If you’re evaluating Claude for Legal or other legal AI solutions, focus on a few critical questions.
Claude for Legal is an important signal of where legal AI is heading. The adoption of AI in Legal has been a major positive change for the past couple of years.
But getting value from Claude for Legal requires more than deploying a model. Organizations still need domain expertise, defensible workflows, and technology purpose-built for scale and complex regulatory environments.
That’s where Hanzo makes the biggest difference.
Hanzo has spent years developing the institutional knowledge and workflows that allow enterprise legal and compliance teams to move faster, reduce risk, and handle preservation, investigations, DSARs, regulatory response, and high-stakes data inquiries with confidence.
Content Team 
