Legal research has always involved a familiar pattern: open a docket, scroll through dozens of entries, open the filings that look relevant, read them, and try to hold the whole picture in your head. For complex litigation with hundreds of entries, that pattern doesn't scale.

Today we're introducing Talk to a Docket, a context-aware AI assistant built directly into the docket view in Docket Alarm. Click an icon on any docket page, and a chat drawer opens alongside the docket. Ask a question in plain English and get answers with citations linking back to the exact filings and documents it used.

What it actually does

Talk to a Docket is an AI chat drawer embedded inside every docket page. You'll see a small icon on the docket. Click it, and a right-side panel slides over roughly a third of the screen. The docket stays visible on the left, the conversation lives on the right, and you can work between the two without losing your place.

You can ask the assistant things like:

• "Summarize the latest filing from the Plaintiff."

• "What is the current status of the Motion to Dismiss?"

• "Are there any upcoming deadlines mentioned in the recent orders?"

Prompts will appear as starter suggestions in the empty state to help new users get oriented. From there, the conversation is yours to drive. Follow-up questions are encouraged, and the assistant remembers what you've already discussed in the thread.

Verifiable citations, not just answers

Anyone working in law has the same first question about an AI tool: can I trust the answer? Talk to a Docket is built around that concern.

Every response includes inline source markers and a Sources footer listing the exact filings the answer was drawn from. Click a source badge and you're taken directly to the document — the same PDF or filing in the docket — so you can verify the claim yourself. Nothing in the answer stands alone; everything points back to the underlying record.

This is the part of the design we spent the most time on. Lawyers don't want a chatbot that sounds confident; they want one that shows its work.

Designed for the way legal work actually happens

A few details from the design that matter in daily use:

• The docket stays visible. The chat opens as a drawer, not a takeover. When the chat is open, the Party Details pane collapses to make room, so you keep the docket entries and the conversation on screen simultaneously.

• Feedback is built into every response. A thumbs up / thumbs down sits under each answer. Mark something unhelpful and you'll see a quick field asking how the answer could be better and that feedback comes straight back to the team.

• Conversation context carries forward. Ask a follow-up question and the assistant remembers what you've discussed. Reset the thread anytime with the New Chat icon in the header.

Talk to a Docket is a step toward a future where litigators don't have to choose between deep court data and conversational AI. You can have both, side by side, with the citations to back every answer.