# Protect client information. Keep control of the work.

Professional work needs clear lines around who can see information and who can act on it. Here is how Rutledge protects client information, keeps records and keeps people in charge of decisions. Ask us for the supporting detail your practice needs.

## Know who is signing in
- Clients verify their identity with a photo ID, a selfie and a liveness check before sensitive access.
- Staff and clients sign in without passwords, using passkeys, through one single sign-on.
- Our own staff use separate sign-in for sales and administration, apart from how visitors are admitted.
- When our support team helps your firm, they use a read-only support session. They can see what you see, but they cannot change your records.

## The client decides who sees what
A client chooses what each professional may see. The client can revoke that consent in one action, and access ends right away. A connection between two professionals does not open their private files to each other. Accepting a referral is separate from getting access. Access is separate from authority to make a professional decision.

## One archive as the system of record
Every captured channel lands in one encrypted, searchable compliance archive: email, calendar items, Teams calls and chats, Zoom recordings, phone calls, texts, voicemail, faxes, signed documents and web snapshots. The archive also writes the CRM activity record, so the history your team works from and the record you keep are the same. Search across channels when you need to find or produce a record.

## Stored safely, checked on arrival
- Connections are encrypted in transit, and stored records and files are encrypted at rest.
- Every uploaded file is scanned before anyone can open it.
- Generated documents are private to the people allowed to see them, and browsers are told not to keep copies.
- Saved work stays until you delete it. Records under retention or preservation rules stay for as long as those rules require.

## Agents act only with approval
AI agents work through approved tools only. Actions that matter, such as sending a message, changing a record or publishing content, wait for a person to approve them. Every agent step is written to a tamper-evident log, so you can see what was proposed, who approved it and what happened. Some actions are never available to an agent.

## Client data goes to zero-data-retention AI only
Client information is sent only to AI providers under zero-data-retention terms. Other information is routed separately. A provider's non-retention terms do not erase the records Rutledge keeps for your work, such as saved conversations, professional records and delivered documents. [See how AI is used](/ai).

## Calculations and sources you can check
Planning and tax work use dedicated calculation tools. Research answers quote the exact source text. Review the inputs, assumptions, coverage and sources behind a result. A generated explanation does not replace that review.

## Every fact and result shows where it came from
Every fact shows who supplied it: you, the assistant, an example or a placeholder. When the assistant fills a gap, the fact shows what it assumed. When a fact is read from a document, it names the document and the line it came from. Every result shows the facts it used and the version of the software that worked it out. When a tax return cannot be worked out, it says which fact is missing instead of guessing. Research quotes the exact text it relies on.

## Charges you can see
Usage is metered and shown on a monthly statement by service and by user. Spending limits pause new paid work until an authorized person raises them. [See pricing](/pricing).

## Providers and processing
The services you use decide which providers process information. This table shows the main processing roles. Ask for the list and terms that apply to your setup.

| Provider | Processing role | Information involved |
|---|---|---|
| Microsoft Azure | Hosting, storage and document processing | Service data and documents for the selected function |
| AI providers under zero-data-retention terms | AI tasks, with client data and other information routed separately | Inputs and outputs for the task |
| Twilio SendGrid | Business and consent-based email | Recipient addresses and message contents |
| Twilio | Text messages and calls | Phone numbers, message contents and call details |
| Stripe | Billing and payments | Billing details and payment information |
| Cloudflare | Website edge services | Requests, connection information and related security data |
| Persona | Identity verification | Photo ID, selfie and liveness check results |
| BoldSign | E-signature | Documents for signature and signer details |
| DocuSign | E-signature through your firm's own account | Documents for signature and signer details |
| Microsoft 365 | Mail, calendar and Teams capture through your firm's account | Messages, calendar items, calls and chats |
| Zoom | Meeting scheduling and recordings through your firm's account | Meeting details and recordings |
| RingCentral | Calls, texts, voicemail and fax through your firm's account | Call records, recordings, messages and faxes |

## Ask for the evidence
[Talk with us](/contact) about access controls, encryption, provider processing, incident response, continuity and independent assessments. Tell us the services and data involved so the answer fits your use. Send vulnerability reports to **security@rutledge.ai**.