CHAPTERS
A time-pressed legal request lands in Slack
Mark, an in-house product lawyer at Anthropic, receives a quick question from a product manager about a feature launched months ago. He has only minutes before his next meeting and lacks the original context he had when writing the initial memo.
The old workflow: re-reading memos and reconstructing context
Mark describes the previous approach: spending a large chunk of the morning locating documents, rereading prior work, and figuring out what changed before answering. The bottleneck is the overhead of finding and synthesizing information rather than the legal analysis itself.
Daily scheduled support in Claude Cowork: a “chief of staff” memo
He sets up Claude Cowork to run a scheduled task first thing in the morning. The result is a daily memo that surfaces what’s on his plate, what’s new, and what’s urgent so he can start the day oriented.
Inbox integration and task triage by what’s due today
With Gmail connected, Claude Cowork presents a sorted list of key items due today. Mark sees a concise set of categorized tasks (routine, new, and a follow-up tied to the product question).
Using /brief to get up to speed on a product change fast
Mark’s most-used skill is “/brief,” designed to rapidly summarize context for a specific product. He invokes it with the product name and the specific change, relying on the skill’s awareness of where review materials and templates live.
Open protocol plugin built from Anthropic Legal’s playbooks
Mark explains that he helped build the /brief plugin based on how Anthropic’s legal team operates. Because it’s an open protocol, other legal teams can tailor it to match their own internal playbooks and processes.
Automated retrieval and synthesis across Slack, Gmail, and the review folder
Claude pulls relevant materials: the prior review document and the latest Slack/Gmail threads. It returns a structured brief matching Mark’s mental model—what was concluded previously, what’s changing now, and what parts of the old analysis are impacted.
Precision over volume: jumping to the exact paragraphs that matter
Instead of reading a 40-page memo, Mark uses the brief to locate the few paragraphs that answer the question. He clicks into the earlier review and verifies the quoted section directly.
Human-in-the-loop legal judgment: “trust but verify”
Mark emphasizes that his name goes on the response, so he checks the underlying sources. The workflow is designed to keep a human in the loop, using Claude to reduce busywork while preserving accountability.
Drafting the reply and closing the loop with Jira tickets
Once comfortable, Mark asks Claude to draft a response to the product team, then approves it. He can also have Claude close out the Jira ticket so the team retains context for future questions.
Building shared legal knowledge and reducing information silos
By closing tickets with context and consolidating learnings, the team builds a growing corpus of knowledge. This makes it easier for the legal department—and potentially the broader company—to access prior analyses without recreating work.
Outcome: faster, confident responsiveness—doing more with less
Mark ends with the practical impact: he can walk into the next meeting confident he responded quickly to a fast-moving product team. He frames Claude Cowork as a way for lawyers to meet rising expectations to “do more with less” while embracing AI responsibly.
