Product

From first prompt to the report a carrier asks for.

The workspace runs on your firm's own model and sits inside how a firm already works: no browser babysitting, no intercepting anyone's traffic, no new habit to enforce. Six steps, below.

Candor is the AI

Your team uses AI inside Candor.

Connect your firm's own model once. Every prompt is screened before it is sent, and every use is logged to the matter.

Every use above runs through the same loop. Tap any step to see it.

See a sample carrier renewal packet →

Works with your case management

It files against the matters you already have.

AI use is filed against your real matters. Works with Clio, MyCase, Filevine, PracticePanther and Smokeball.

Your administrator authorises it

Sign in on the platform's own site, the standard way. Candor never sees a password, and your firm can cut off access from your side at any time without asking us.

Matters are read, never kept

Candor reads the matter list when it needs it and lets it go. Your client list is not copied into our storage, and every time it is read the access log records it.

Facts without the client name

Pull the matter number, practice area, and status straight into a prompt so nobody retypes them. The matter title stays out by default, because titles usually carry the client's name.

One thing to check on your side: private integrations need a mid tier plan or above at most platforms, and the person connecting has to be an administrator.

Also included

Beyond the core loop.

Readiness score

How completely your firm can document its AI use, with the gaps named.

Court disclosure certificate

Generate a draft AI use certification for a filing, populated from the matter's record, for the attorney to review and sign.

Client consent tracking

Whether a client agreed to AI use on their matter. Where they declined, Candor says so before anyone logs a use.

Exception log

A skipped screen or an unapproved tool is recorded with a reason and raised for a partner.

Supervision queue

What is waiting on a partner, for how long, and where AI use is outrunning review.

Billing note

Mark when an AI use affects what the client is billed, at the time it happens.

Time on task, measured

Time actually spent with the AI, with a draft billing entry. Any comparison uses your estimate, never ours.

Export it, or take it with you

The whole firm record for your carrier, or one matter for a court with every other client left out. Always yours to keep.

Governance around the work

The firm's AI policy, put into practice.

Your firm’s rules, applied at the moment someone uses AI and recorded with everything else. Candor does not decide what the rules should be.

Rules, pre-flight and review holds

Rules over the tool, the task, how sensitive the material is and who is doing it. Before a use is logged, Candor answers go ahead, review first, or not permitted, with the reasons.

Approved tools and vendor profiles

For each tool: whether it trains on what it is sent, how long it keeps it, whether a data agreement is on file, and a review date.

Risk radar

The last 30 days against the 30 before: flagged uses, unapproved tools, sensitive material, unsigned research and open incidents. Plain counts.

Matter files and receipts

Every AI use on a matter in order, with its sign-off and consent history. Any use prints as a one-page receipt.

Incidents

Record a problem when it is found, then each step to closed. Facts only, never edited afterwards.

Engagement-letter check

Candor points to the sentences about AI in a client’s engagement letter, read in your browser and never uploaded. The attorney decides what to record.

Candor Academy

Short courses on using Candor and on your firm’s own policy, with quizzes. A rule can hold research until a course is passed.

Or work where you already are

Prefer ChatGPT, Claude, or Clio? The panel rides along there too.

The same Ask · Screen · Check panel follows your team into the AI tools and practice systems they already use, so the screening, citation checks, and one-click logging come along. Nobody has to remember to open a separate app.

Screen before you send

A quiet panel on the page redacts client identifying details before anyone pastes them into a chatbot. The text stays on the device.

Check the answer for real cases

On ChatGPT, Claude and Gemini, press Check and Candor reads the answer on screen and flags case citations that may not exist, checked against a public database of real court opinions. A prompt for the reviewer, never a verdict.

Paste to check, anywhere else

On other AI sites, desktop apps and legal AI platforms, paste the answer into the panel and the same citation check runs. Candor never reads the page on Microsoft Copilot or Microsoft 365 Copilot; highlight or paste instead.

Log it in one click

Capture the tool, the task, and the matter without leaving the page. Your firm's rules answer first, and the entry files into the same sealed record as everything else.

It reads the page only when asked and sends out only the bare citations it checks. A match means the case exists, not that it says what the AI claimed. Setup takes about a minute. See every tool Candor works with →

Record of AI Use APPEND-ONLY · HASH-CHAINED · SIGNED a3f9c1 Wilson: demand letter MATTER 24-0417 · CHATGPT · 03/14 7b21e8 Ramirez: research memo MATTER 24-0392 · CLAUDE · 03/15 f04d9a Okafor: deposition prep MATTER 24-0418 · COPILOT · 03/15 A. Park ATTORNEY SIGN-OFF · 03/15/2026 C CANDOR
What comes out

Six steps in. One sealed record out.

Everything above lands in a single append-only, hash-chained record: screened inputs, checked citations, the attorney's sign-off. Export it whole for your carrier, or one matter for a court, with the chain intact and every other client left out.

How the record holds up →
Common questions

Questions firms ask about Candor

Does Candor store what I type into an AI tool?

No. The confidentiality screen runs in your browser, and Candor keeps only the fact that a check ran and which categories fired, never the text. Retaining any underlying content is an explicit choice a firm turns on, off by default.

Whose AI does Candor use, mine or its own?

Your firm's own model, on your own key. Every prompt is screened before it reaches the model, and Candor keeps the fact of each use, not its content.

Does Candor write anything into my case-management system?

Only if your firm turns it on. Candor reads the matter list when it needs it and never copies your client list into its storage. With Clio write-back switched on by an admin (it is off by default), each logged AI use becomes a note on its Clio matter, written under that person's own Clio login. The note records the fact of the use (tool, task, who, when), never a prompt or an answer.

Can Candor apply our firm's AI policy?

Yes, as rules your admin writes: which tools, which tasks, how sensitive the material, and for whom. Before a use is logged Candor gives your rules' answer (go ahead, needs review first, or not permitted) and the entry records it. Candor applies your rules; it does not decide what they should be, and it is not legal advice.

What does Candor record when something goes wrong?

The facts of an incident: what kind, how serious, the matter number, the tool, when it was found, and each step taken with its date. It keeps no written description, so the account of what happened stays in your firm's own files.

Does Candor read our engagement letters?

Only in your browser, and only when you pick one. Candor points to the sentences about AI so the attorney can record consent. Nothing from the letter is sent to Candor or kept.

See the whole loop on your own matters.