# Proxara

> Proxara gives approved AI one governed way to retrieve information and complete actions across a regulated firm's internal systems, built for accounting and tax firms and for independent wealth management firms. Proxara Connect links Claude, ChatGPT, Gemini and AI agents to Microsoft 365, SharePoint, Karbon, Salesforce and the firm's tax systems with nothing installed. For each request Proxara identifies the employee, checks their permissions and the firm's policy, retrieves only the records needed, replaces protected client identities with consistent stand-ins before the work reaches the model, resolves proposed actions to real records only at a controlled write boundary, and records what was retrieved, approved and changed. A separate device deployment, Endpoint Protection, extends the same policy to every AI on a managed laptop and recognizes a thousand-plus AI sites and apps.

Tagline: Connect your firm to AI.

## Core Pages
- [Homepage](https://proxara.ai/): What approved AI can complete across the firm's systems: one cross-system job end to end, the firm's own context, model neutrality, the boundary holding under attack, and how a firm gets started.
- [Proxara Connect](https://proxara.ai/connect): The governed retrieval and execution path into the firm's own systems. One Microsoft approval, one connector address, and every employee signs in with the account they already use. Nothing is installed on a device.
- [Connect Actions](https://proxara.ai/connect/actions): How AI completes work rather than only reading. The model proposes a typed action; Proxara resolves the real client, checks the employee's authority and the firm's rules, executes inside the firm's environment, and verifies what changed. Four lanes per action: automatic, approval required, dual approval, prohibited.
- [Work Proxara completes](https://proxara.ai/work): The missing-document chase, the notice response, the meeting that becomes tasks. Named jobs finished across Karbon, Microsoft 365, SharePoint, CRM and tax systems, and verified in the systems of record.
- [Agents](https://proxara.ai/agents): Every agent the firm builds or buys arrives with nothing, appears on the firm's register the first time it acts, and works under authority that is decided again before anything leaves.
- [The Proxara Engine](https://proxara.ai/engine): Customer-local AI processing. Exact work runs locally, sensitive reasoning runs on a contained model, and the frontier model receives only the package built for that job.
- [Section 7216](https://proxara.ai/section-7216): IRC section 7216 and AI in a tax practice. Proxara compiles context inside the firm so the external model receives no client names, no identifiers, and no return figures, and the mapping never leaves the building.
- [Identity & Context](https://proxara.ai/identity-and-context): How Proxara works out who asked, whose permissions apply, and which client the work concerns, across every connected system. The model receives consistent stand-ins; the authorized employee sees the real result.
- [Policy & Authority](https://proxara.ai/policy-and-authority): The firm decides what AI may see, join, infer, say, and do. Proxara applies that policy to each request.
- [Policy & Identity](https://proxara.ai/policy-and-identity): Five permissions, four ways two records may be connected, and one request run under three different policies.
- [Customer-Local Processing](https://proxara.ai/customer-local-processing): Proxara does not send documents with the sensitive parts removed. It builds the payload from claims the firm's policy approved, inside the firm's own environment.
- [Evidence & Audit](https://proxara.ai/evidence-and-audit): The difference between a transcript and a record: who requested, what was retrieved, what the model received, which policy applied, who approved, what changed, and what the source system confirmed. Assembled into governance reports by period, person, client, or system.
- [Exfiltration Defense](https://proxara.ai/exfiltration-defense): What happens when somebody attacks an AI that reaches the firm's systems. A poisoned document, a hijacked prompt, or a rogue agent arrives as a request, and none of it creates authority.
- [Model and System Neutrality](https://proxara.ai/model-and-system-neutrality): Proxara is not a model vendor and not a system vendor. The firm can change models or systems without rebuilding the policy, the connections, or the record.
- [Continuous AI Analytics](https://proxara.ai/continuous-ai-analytics): One structured record of the firm's AI, read through three lenses: spend and value, coverage and proof, reach and exposure.
- [Endpoint Protection](https://proxara.ai/universal-ai-supervision): The device deployment, separate from Proxara Connect: one signed agent pushed through the device management IT already runs, covering every AI on a managed laptop. It reads AI traffic and nothing else, reports unsupported or bypassed surfaces rather than hiding them, and comes off cleanly.
- [The Console](https://proxara.ai/console): Where leaders operate Proxara: define approved kinds of work, manage identities, agents, and delegations, review the work that needs a decision, and see confirmed outcomes.
- [Rox](https://proxara.ai/rox): Proxara's assistant in Microsoft Teams and Slack. Ask what happened, get the weekly picture, and run the controls from the place the firm already works.
- [Security](https://proxara.ai/security): Dedicated single-tenant AWS account per customer, customer-held KMS keys, AES-256 at rest, TLS 1.2+ in transit, signed audit chain verifiable offline.
- [Why Proxara](https://proxara.ai/why-proxara): Why an accounting or tax firm needs a control point of its own, and why now.
- [The Evaluation](https://proxara.ai/pilot): A free twenty-one-day Proxara Connect evaluation on the firm's own tenant: a live demonstration on a practice tenant first, one setup call, everyone connects from inside the assistant they already use, and a decision in writing. Nothing installed.
- [See it live](https://proxara.ai/request-demo): Two ways to see Proxara live: the Claude connector demonstration on a stocked practice tenant, or the sandbox app beside real ChatGPT and Claude.
- [Talk to us](https://proxara.ai/contact): Book a call with the Proxara team to scope a deployment.
- [About](https://proxara.ai/about): The team building Proxara, and why it builds for adoption rather than blocking.
- [Results](https://proxara.ai/results): Semantic redaction test results across 206 prompts and 8 industries.
- [Trust Center](https://proxara.ai/docs/trust-center): What each deployment reaches and where it stops, what Proxara sees and never sees, the record, framework mapping, subprocessors, incident response, and every governing document.
- [FAQ](https://proxara.ai/docs/faqs): Common compliance-team questions on deployment, data sovereignty, employee experience, and AI governance.

## Product Surfaces
- [Solutions](https://proxara.ai/solutions): The two deployments side by side, and how to choose: Proxara Connect for governed access to the firm's own systems with nothing installed, Endpoint Protection for every AI on a managed laptop.
- [Compare Deployments](https://proxara.ai/compare-deployments): One policy engine, two ways a firm can run it. What each deployment reaches, and what it does not.
- [Deployment](https://proxara.ai/deployment): How Proxara becomes operational: a dedicated cloud environment under the firm's own keys, connectors authorized, identities mapped, approved kinds of work configured, then a pilot group before wider rollout.
- [Device Rollout](https://proxara.ai/solutions/deployment): Silent MDM deployment of the device agent across the workforce.
- [Device Deployment Guide](https://proxara.ai/solutions/deployment/device): MDM push or single installer. AI traffic on the device passes one checkpoint.
- [Integrations](https://proxara.ai/integrations): The systems an accounting or tax firm already runs, joined for one piece of work: Karbon, Microsoft 365, SharePoint, Outlook, Salesforce and the tax systems, reached from Claude, ChatGPT or a firm-built agent.
- [Walkthrough](https://proxara.ai/walkthrough): Step-by-step of how the device proxy intercepts, redacts, and restores sensitive data in real time.
- [Due Diligence Guides](https://proxara.ai/due-diligence): The documentation a security or IT team needs to assess Proxara, for Proxara Connect or the device agent.

## Industry Briefs
- [Financial Institutions](https://proxara.ai/financial-services): AI coordinates work across wealth, banking, trust, and other business units without dissolving the boundaries between them. The hub for the wealth management and accounting pages.
- [Accounting and Tax Firms](https://proxara.ai/industries/accounting): AI through busy season with tax return information held inside the firm. Section 7216 consent captured per purpose where a task needs it, the FTC Safeguards Rule WISP answered from the record, and every use attributed to a named preparer.
- [Wealth Management](https://proxara.ai/industries/wealth-management): Meeting preparation, CRM updates, client follow-up and service requests completed across the advisor's own systems, for the regulators the firm answers to.
- [Regulated SMEs](https://proxara.ai/for-regulated-firms): AI governance for regulated small and mid-size firms.

## Flagship Resource
- [AI Governance for RIAs: What You Actually Need to Do in 2026](https://proxara.ai/resources/finra-ai-governance-playbook): A practical playbook for CCOs and compliance officers. What FINRA and the SEC expect, what most firms are missing, and a 9-step checklist to close the gap before an examiner finds it.

## Core Explainers
- [Docs Guides](https://proxara.ai/docs/guides): Task-first guides: set up Proxara in Claude, ChatGPT, Copilot, or Gemini, connect Microsoft 365, Karbon, Salesforce and tax software, put autonomous agents to work, and run the firm from the console.
- [Proxara in Claude](https://proxara.ai/docs/guides/claude): An Owner adds the connector once; everyone signs in with their Microsoft work account.
- [Proxara in ChatGPT](https://proxara.ai/docs/guides/chatgpt): Publish it to the workspace once; sign in with Microsoft.
- [Proxara in Microsoft Copilot](https://proxara.ai/docs/guides/copilot): IT deploys the agent once; it appears in Copilot on its own.
- [Connect Microsoft 365](https://proxara.ai/docs/guides/microsoft-365): One delegated, read-only admin consent; every read carries the signed-in employee's own access.
- [How Proxara works](https://proxara.ai/docs/guides/how-proxara-works): What the model sees, what it never sees, and where the real names live.

## Trust Center
- [Trust Center](https://proxara.ai/docs/trust-center): Public security, privacy, and compliance documents.
- [Security Overview](https://proxara.ai/docs/trust-center/security-overview): Architecture, encryption, isolation, access controls, audit logging.
- [Product Privacy Policy](https://proxara.ai/docs/trust-center/product-privacy-policy): How Proxara processes data in customer-managed and managed deployments.
- [Data Processing Addendum](https://proxara.ai/docs/trust-center/data-processing-addendum): Data-processing roles, rights, and contractual controls.
- [IRC § 7216 and the disclosure boundary](https://proxara.ai/docs/trust-center/section-7216): What the rule says, which technical facts Proxara enforces, which deployment assumptions the reading depends on, and which questions belong to the firm's own tax counsel.
- [HIPAA BAA](https://proxara.ai/docs/trust-center/hipaa-baa): Healthcare-specific contracting support.
- [Subprocessor List](https://proxara.ai/docs/trust-center/subprocessor-list): Public list of subprocessors used in managed deployments.