The Supervised Intake Standard™
Is this AI safe with the bar?
No bar certifies AI. So we show our work. All of it.
Every answer a managing partner needs before go-live, in writing, on one page.
Nobody answers this.
Bar guidance covers drafting and filings.
Vendors publish features.
Nobody publishes what the AI does on the intake call.
As of August 2026. Claim disclosures.
Seven tests. Run them on anyone.
The Supervised Intake Standard™ is ours. It’s written so you can score any vendor — Clio Grow AI, Smith.ai, or a human answering service — against the same seven questions.
- 01
Disclosure
Does the caller know it’s AI?
Rule 1.4, 7.1, 8.4(c); FL 24-1, CA 2026
Yes. First sentence. Every call, text, and chat.
Ask your vendor: Does your AI say it’s AI in the first sentence of every call, text, and chat?
- 02
No advice
What does it say when asked for legal advice?
Rule 5.5 (UPL), 5.3
It redirects to an attorney. Scripted. Logged.
Ask your vendor: When a caller asks for legal advice, does it redirect to an attorney — scripted and logged?
- 03
No promises
Can it quote fees or outcomes?
Rule 7.1, 1.5
No. It cannot state a fee, a result, or that the firm will take the case.
Ask your vendor: Can it state a fee, predict a result, or say the firm will take the case?
- 04
Prospective client
What happens to what a declined caller told you?
Rule 1.18
Captured minimally, flagged, retained per firm rule, deletable.
Ask your vendor: What happens to a declined prospect’s information — and will you confirm it in writing?
- 05
Conflicts
Who runs the conflict check?
Rule 1.7, 1.9, 1.18
The AI captures parties. A person at the firm decides.
Ask your vendor: Who runs the conflict check — the AI, or a person at our firm?
- 06
Data
Where are recordings and transcripts, who reads them, are they used for training?
Rule 1.6, 1.15 analog; state privacy
Firm-controlled storage. Named roles. Never used to train models.
Ask your vendor: Where are recordings stored, who can read them, and are they ever used to train models?
- 07
Supervision
Did a lawyer approve the script, and can you prove it?
Rule 5.1, 5.3; ABA 512
Yes. Sign-off log, version history, export on request.
Ask your vendor: Can you show a lawyer approved the script, with a sign-off log and version history?
What ShiFt says when a caller asks for legal advice.
Caller
“Can I sue?”
ShiFt
“An attorney will answer that. I’ll note it down and get you booked with one.”
Caller
“Will you take my case?”
ShiFt
“An attorney will answer that. I’ll note it down and get you booked with one.”
Caller
“How much will I get?”
ShiFt
“An attorney will answer that. I’ll note it down and get you booked with one.”
Same words every time. Scripted, logged, and never a legal opinion. This is the page’s proof.
Recordings. Transcripts. Who can see them.
- Storage
- Firm-controlled storage. Not a shared vendor pool.
- Retention default
- Kept per the firm’s retention rule. The firm sets the clock.
- No training
- Never used to train models. In the agreement, not just on this page.
- Access roles
- Named roles only. Every read is logged.
- Deletion on request
- Deleted on request.
- Declined prospect (Rule 1.18)
- Minimal capture, flagged prospective, retained per firm rule, deletable.
We capture. Attorneys decide.
What it captures
Parties, adverse parties, matter, jurisdiction.
How it hands off
Routed straight into the firm’s existing conflict process.
Why the AI never decides
A conflict call is a legal judgment. A person at the firm makes it.
No leading AI intake vendor published its conflict handling as of August 2026. ShiFt publishes this one.
Callers know it’s AI. Every time.
- Disclosure line
- The caller hears it’s AI in the first sentence.
- Recording consent
- Disclosed per one-party and two-party consent states.
- Texting consent
- Captured before the conversation continues.
- State variations
- Wording adjusts by state — CA, FL, TX, GA baseline.
- Human takeover
- Anytime.
A lawyer signs off. Then it goes live.
- Firm-approved scripts
- A lawyer signs off before anything goes live.
- Supervision log (Rule 5.3)
- Sampling and oversight are recorded.
- Change control
- Every script version is tracked.
- Export on request
- Sign-off log and version history, exportable.
The eight questions partners ask first.
Send this to any vendor.
Seven questions, copy-ready. Send them to your current or proposed intake vendor. ShiFt’s answers are already on this page.
Copy-ready
We are reviewing AI intake vendors against the Supervised Intake Standard. Please answer these seven questions in writing.
- 1.Does your AI tell the caller it is AI in the first sentence of every call, text, and chat?
- 2.When a caller asks for legal advice, does it redirect to an attorney — and is that redirect scripted and logged?
- 3.Can it state a fee, predict a result, or say the firm will take the case? If so, how is that turned off?
- 4.What happens to a declined prospect’s information — is it captured minimally, flagged, retained per our rule, and deletable? Confirm in writing.
- 5.Who runs the conflict check — your AI, or a person at our firm?
- 6.Where are recordings and transcripts stored, who can read them, and are they ever used to train models?
- 7.Can you show a lawyer approved the script, with a sign-off log, version history, and an export on request?
The Supervised Intake Standard™ — a working paper.
The seven tests written out in full: the rule each maps to, ShiFt's answer, and the question to put to any vendor. One idea, two formats — the page you're reading and a document you can circulate to partners.
The working paper is in counsel review. Leave nothing to guess — we publish it here the moment it clears.
Run the check. Two minutes.
Summarizes public guidance for planning. Not legal advice. Claim disclosures · Methodology