Skip to content
Legal · ABA Formal Op. 512

Use ChatGPT and Claude on Your Firm's Email — Without the Privilege Risk

PortEden replaces client identifiers, settlement terms, and privileged content with placeholders before your prompt reaches OpenAI or Anthropic. Use any AI you want; the model never sees the underlying client information.

See pricing

Free for solo practitioners · No credit card · Op. 512-aligned audit log built in

Maps to
ABA Op. 512
GDPR
SOC 2
CCPA
HIPAA
The Risk

What Goes Wrong Without PortEden

You Paste a Client Email Into ChatGPT to Draft a Reply

That thread holds client identifiers, matter details, sometimes settlement figures. The moment OpenAI's API receives it, you've sent confidential client information to a third-party processor — without the safeguards Model Rule 1.6 and Op. 512 expect.

Your Assistant Uploads Case Documents to Claude to Summarize

Discovery PDFs, deposition transcripts, expert reports — all confidential, often privileged. Claude's context window now holds opposing counsel's best discovery target. Privilege challenges in litigation can turn on this exact moment.

You Connect Outlook Calendar to an AI Assistant

Meeting titles like "Settlement w/ ACME — final number" and attendee lists with client GCs leak more than you think. Most AI assistants log calendar metadata for as long as their retention policy allows.

How PortEden Protects Your Firm

Privileged Client Data, Redacted Before It Reaches the Model.

PortEden inspects every field your AI is about to see. Client identifiers, matter numbers, settlement terms, and privileged phrases are replaced with placeholders at the boundary — never sent to OpenAI or Anthropic.

Your data
PortEdenRedact
Your AI
Claude
ChatGPT
Copilot
Gemini
Grok
Safe
Sensitive
Redacted
Compliance Reality

What ABA Formal Opinion 512 Actually Requires When Your Firm Uses ChatGPT or Claude

  • Maintain technological competence — understand what your AI does with client data, where it lives, and who can read it.
  • Protect confidential client information. Sending it to an AI vendor whose contract doesn't restrict storage, training, or human review can breach Model Rule 1.6 — and may waive privilege.
  • Communicate with clients about your AI use when it could affect their representation, and obtain consent where the facts call for it.
  • Supervise AI work product the way you'd supervise a junior associate's draft, and don't bill clients for the time AI saved you.
ABA Formal Opinion 512 (PDF)
The Solution

Built For Lawyers

Privilege-Aware Redaction in Gmail and Outlook

Replaces 50+ identifier types — client names, matter numbers, settlement amounts, SSNs, EINs, and privileged phrases — with placeholders in under 200 ms before the prompt leaves your perimeter. The AI receives only the redacted version; the original text never reaches OpenAI or Anthropic.

Calendar, Drive, and SharePoint Metadata Stripping

Meeting titles, attendee lists, document names, and Drive/SharePoint contents are redacted in transit. Triage your week with AI without leaking your client list — same redaction policy across Google Workspace and Microsoft 365.

Engagement-Letter-Driven Consent Log

Configure per-client AI rules at intake. Every prompt that touches a client's data is logged, timestamped, and exportable as a CSV or to your SIEM — the kind of record bar opinions like Op. 512 are starting to expect.

Works With Claude, ChatGPT, and Copilot — No Plugin Install

Your team keeps using whatever AI they already prefer, including the Claude desktop app and ChatGPT Enterprise. PortEden sits in front of any model with no per-lawyer browser extensions to roll out — partner approval to associate adoption in days, not months.

Works With What You Already Use
Gmail
Gmail
Inbox-side redaction for AI replies and summaries
Outlook
Outlook
Microsoft 365 inbox-side redaction
Google Calendar
Google Calendar
Strip meeting titles and attendee lists before sync
Teams
Teams
Microsoft 365 chat and channel redaction
With and Without PortEden

The Same Workflow, Two Very Different Outcomes

Drafting a Client Email Reply With ChatGPT
Without
Client name, matter, and privileged content sent to OpenAI in plain text — confidentiality breach and potential privilege waiver.
With
Client identifiers and privileged phrases replaced with placeholders before the request leaves your network.
Searching Your Inbox With Copilot or Gemini
Without
Every matching email — sender, subject, body, attachments — is sent to the AI in plain text. Includes any privileged content the search touches, even results you ultimately discard.
With
Email content reaches the model with client identifiers and privileged phrases replaced with placeholders. The AI ranks results without seeing the underlying content.
Calendar-Aware AI Assistant on Outlook or Google Calendar
Without
Meeting titles like "Settlement w/ ACME" and attendee GCs logged by the AI vendor for the length of their retention policy.
With
Meeting titles and attendee identifiers sanitized before any sync.
Op. 512 Audit or Bar Inquiry Into Your AI Use
Without
No record of which client data went to which model. You're reconstructing it from screenshots.
With
Per-client, per-matter audit log of every prompt, exportable on demand.
Multi-Lawyer Rollout to Associates and Paralegals
Without
Each user follows AI policy by hand; one paste-and-prompt is a firm-wide problem.
With
Firm-wide redaction defaults; per-matter overrides flow from your engagement letter.
Try It on Your Firm

Five-Minute Setup. Free for Solo Practitioners.

Connect Gmail or Outlook via OAuth. Pick a redaction profile. Keep using ChatGPT or Claude exactly the way you do today — with privilege protected by default.

See pricing

Frequently Asked Questions

Does using ChatGPT or Claude with PortEden count as disclosing client information to a third party?
PortEden replaces client identifiers and privileged content with placeholders before the prompt leaves your perimeter. The third-party AI receives only the redacted version, so the underlying confidential information isn't disclosed in the clear. Your firm still owes the usual Op. 512 obligations — competence, supervision, and client communication — but the confidentiality leak is closed at the boundary.
Do I still need to talk to clients about AI use under ABA Op. 512?
Op. 512 says you should communicate with clients about AI when it could materially affect their representation. PortEden makes that easier: it ships with a sample engagement-letter addendum and logs every AI interaction by client and matter, so you can produce records on demand if a client or a regulator asks.
Will PortEden change my workflow? I don't want yet another AI tool to learn.
PortEden sits in front of ChatGPT, Claude, and Copilot — you keep using whatever AI you already prefer. There's nothing new to learn; redaction happens in the background between your inbox/calendar and the model.
If opposing counsel learns we used ChatGPT on a matter, can they pierce privilege?
Privilege turns on whether you took reasonable steps to keep the communication confidential. PortEden's audit log demonstrates the kind of technical and procedural safeguards bar opinions like Op. 512 are starting to expect — far stronger than "we trusted OpenAI's privacy policy." We're not aware of binding case law on AI yet; that's part of why a defensible record matters.
Does PortEden work with Microsoft 365 and Outlook?
Yes. PortEden integrates with Outlook, Microsoft Calendar, OneDrive, SharePoint, and Teams alongside the Google Workspace stack — same redaction policies across both.
Can my paralegals and associates use it without per-matter reconfiguration?
Yes. Set firm-wide redaction defaults once; per-matter and per-client overrides flow from your intake form or matter management system, so the right rules apply without anyone touching a config.
What does it cost and how long does setup take?
There's a free tier for solo practitioners. Firm pricing scales by lawyer — full pricing is on the pricing page. Setup is under 5 minutes for a solo lawyer on Gmail + ChatGPT or Claude. Larger firms typically take a half-day for SSO and matter-management integration.

Keep Exploring

Ready to Use AI Without the Privilege Risk?

Five-minute setup. Free for solo practitioners. Op. 512-aligned audit log from day one — exportable per client, per matter, on demand.

See pricing

Multi-office firm or AmLaw 200? Talk to sales →