Skip to content
AnthropicAnthropic

How Anthropic uses Claude in Legal

Legal teams lose days per week to routine tasks like contract redlining and marketing reviews. Mark Pike, Associate General Counsel, shares how Anthropic's lawyers uses Claude to build workflows that cut review times from days to hours—no coding required. Check out the full case study to learn more: www.claude.com/blog/how-anthropic-uses-claude-legal Stay tuned for more stories in the "How Anthropic uses Claude" series.

Mark Pikeguest
Dec 8, 20253mWatch on YouTube ↗

CHAPTERS

  1. 0:00 – 0:30

    A playful demo: a “legal lamp” controlled with Claude Code

    Mark Pike introduces a desk lamp project that blinks Morse code based on instructions he types. The demo is used to illustrate how Claude Code can turn plain-language requests into working behavior, even for someone non-technical.

    • “Legal lamp” blinks Morse code as a visible Claude Code demo
    • Shows an interactive “it’s thinking” feedback loop
    • Demonstrates rapid translation from intent to execution
    • Frames the tool as accessible to non-engineers
  2. 0:30 – 1:01

    Why legal teams need automation: reducing dull, error-prone busywork

    Mark describes the repetitive nature of legal work and how it can lead to mistakes and procrastination. He positions Claude as a way to offload low-value tasks and free up attention for higher-judgment work.

    • Repetition makes legal work dull and time-consuming
    • Busywork increases the chance of mistakes
    • Tasks often get deferred because they’re cognitively unrewarding
    • Claude helps shift effort toward higher-value legal thinking
  3. 1:01 – 1:31

    From last-minute marketing requests to a structured review workflow

    Mark explains a common pain point: marketing asks for legal review right before launches. He shares that he asked Claude to build a workflow based on what legal cares about, and it produced a workable setup.

    • Marketing often requests review a day before launch
    • Legal needs a consistent way to spot common risk areas fast
    • Claude can generate a workflow from stated priorities
    • Early standardization reduces fire-drills and missed issues
  4. 1:31 – 2:01

    Using the marketing material self-review tool: paste, analyze, review

    He walks through the self-review tool experience: a marketer pastes content and clicks “Analyze Content.” Claude then performs an initial pass to identify potential legal issues.

    • Self-review tool is shared/pinned for easy access
    • User copies/pastes blog or marketing content into the tool
    • Single action (“Analyze Content”) triggers Claude review
    • Claude produces structured review results for the submitter
  5. 2:01 – 2:32

    Issue spotting + positive feedback: what Claude flags and what’s strong

    Claude returns a set of specific concerns to address and also notes what looks good. The flagged categories include accuracy, security claims, third-party rights, and partnership considerations.

    • Outputs a list of issues to address (example: five items)
    • Flags accuracy concerns and security-related claims
    • Checks for publicity rights/third-party content usage
    • Highlights partnership considerations
    • Also reports strengths (clear descriptions, use cases, integration detail)
  6. 2:32 – 3:02

    Turning analysis into action: auto-generated Slack message and legal tickets

    Mark shows how the workflow packages Claude’s findings into a Slack message and then routes the work into a ticketing flow for legal review. This bridges the gap between “insights” and “execution.”

    • One-click generation of a Slack summary for the legal team
    • Copy-to-clipboard handoff to speed communication
    • Direct navigation into legal ticket filing
    • Standardized summaries reduce back-and-forth and missing context
  7. 3:02 – 3:32

    Risk triage and guardrails: prioritization with humans in the loop

    The system prioritizes issues using a low/medium/high risk signal based on a framework Mark provided. He emphasizes that legal still reviews the output because hallucinations and errors remain possible.

    • Claude acts as a first-pass “eyes and ears” for review
    • Risk levels (low/medium/high) are driven by a provided framework
    • Helps prioritize what needs legal attention most
    • Human-in-the-loop is required due to hallucination risk
    • Speed improves without removing accountability
  8. 3:32 – 3:39

    Other legal department use cases and how to get started

    Mark lists additional ways the legal team uses Claude, including contract redlining and compliance-style reviews. His advice for adoption is to start with the most routine work and experiment.

    • Commercial contract redlining support
    • Conflict of interest policy assistance
    • Reviewing outside business activity requests
    • Practical adoption tip: start with routine, repeatable tasks
    • Discover capabilities by trying real workflows in Claude

Get more out of YouTube videos.

High quality summaries for YouTube videos. Accurate transcripts to search & find moments. Powered by ChatGPT & Claude AI.