AI Inbox & Meetings Agent
Reads every overnight email and classifies it. Researches every meeting attendee and drafts a briefing. Transcribes every Zoom and writes the follow-up. The two workflows that consume the most attorney time, automated end to end with you in the pilot seat.
- Overnight inbox triage with classification, labeling, and a prioritized digest
- Pre-meeting briefings on every external attendee, in Slack by 7am
- Post-meeting transcripts and follow-up email drafts within an hour
- Action items extracted and routed to the responsible attorney
Blueprint refunded if we don't leave you with a clear path forward.
A morning that runs itself
It's 6:45am. The Inbox & Meetings agent has been working since 6. It read the 47 emails that came in overnight, classified them (12 from clients, 8 from opposing counsel, 4 from the court, 23 newsletters and outreach), archived the noise, applied labels, and drafted a 4-line digest of what needs attention.
It also looked at your calendar. You have a 9am call with a prospective client, a 10:30 deposition prep with co-counsel, and a 2pm settlement conference. For each meeting, the agent has researched the attendees (their firm, their recent matters, their published work), pulled the matter file, and drafted a 1-page briefing that lands in your Slack with a link to the source documents.
By the time you finish breakfast, the inbox is clean, the day's meetings are pre-briefed, and you've sent 4 emails by tapping approve on the drafts. The work that used to consume the first 90 minutes of your morning is done before you start your morning.
What the Inbox & Meetings agent does
The agent runs in two coordinated tracks. The inbox track triages every incoming email, classifies it (client, opposing counsel, court, vendor, marketing, internal), applies the right labels in your email system, archives the noise, and surfaces a prioritized digest of what needs attorney attention. For routine replies (scheduling confirmations, document acknowledgments), it drafts a response and leaves it in your drafts folder.
The meeting track watches your calendar for external meetings, researches each attendee, pulls the relevant matter file, and assembles a pre-meeting briefing. After the meeting, it pulls the recording from Zoom or Teams, transcribes the conversation, extracts action items, drafts the follow-up email, and updates the matter file in your practice management system. Both tracks share the same audit trail.
Inputs and outputs
Inputs
- Gmail or Outlook inbox with the appropriate scope of access
- Google Calendar or Outlook Calendar feed
- Zoom, Microsoft Teams, or Google Meet recordings
- Practice management system matter records and contact lists
Outputs
- Morning inbox digest with prioritized items and routine replies drafted
- Pre-meeting briefing in Slack 60 minutes before each external meeting
- Meeting transcript filed to the matter within an hour of the meeting ending
- Follow-up email draft with action items extracted
- Matter timeline updated automatically with the meeting summary
- Weekly review of inbox response time and meeting throughput
Integrations
- Gmail or Outlook (with appropriate OAuth scopes)
- Google Calendar or Outlook Calendar
- Zoom, Microsoft Teams, or Google Meet for the recording feed
- Slack or Microsoft Teams for digest delivery
- Clio, MyCase, Practice Panther, or your practice management system for matter context
Practice-area fit
Every engagement tunes the agent to your firm's specific matter mix. The areas below are where we've seen this agent fit most naturally.
Function
Safety posture
The agent reads aggressively, writes conservatively. It can apply labels, file emails, archive noise, and draft replies. It cannot send an email, accept a meeting on your behalf, delete a message, or share a transcript externally. Every outbound communication waits in your drafts folder for review and send.
Privileged content stays inside your account. The transcription layer runs through partners with zero data retention so call audio isn't stored after processing. For matters that contain privileged or work-product content, the briefings and follow-ups respect the same access boundaries that govern the underlying file.
Frequently Asked Questions
The threshold is conservative. Anything the agent isn't certain belongs in a low-priority bucket gets surfaced in the digest for human review. Over the first 30 days the threshold gets tuned based on which calls the agent got right and which it got wrong, with the partner reviewing the misses each week. After tuning, the false-archive rate is typically under 1%.
Start saving time and money on Day 1 with Dodonai
Learn how Dodonai can help take your law practice to the next level.