Socrates Crayon — Compliance & Risk Controls

Risk Controls

Six Layers of Protection.
Zero Client Data Exposure.

1. Full Audit Trail (Attestation Chain)

Every stage of the analysis produces a unique digital fingerprint of its inputs and outputs. If any data changes between stages, the analysis halts. The result is a tamper-proof audit trail that shows exactly what documents were received, how they were processed, and what findings were generated.

What this means for compliance: You can verify that the report your advisor presents was generated from the exact documents the client provided, with no modifications, insertions, or omissions at any stage. The hash chain is independently verifiable.

2. Data Isolation

Client documents are processed in a closed system. No data is transmitted to OpenAI, Google, Meta, or any third-party AI provider. No client documents are used to train any model. No personally identifiable information leaves the processing environment.

What this means for compliance: Unlike advisors using ChatGPT or similar consumer AI tools, Socrates Crayon does not expose client data to third-party services. There is no risk of client information appearing in another user's AI output or being used to improve a third-party model.

3. Separation of Judgment and Computation

AI agents handle document interpretation and issue identification. A deterministic projection engine — written in code, not AI — handles all numerical outputs: tax projections, exemption utilization, discount modeling, and exposure calculations. The AI never produces dollar amounts.

What this means for compliance: The most common AI failure mode — generating plausible but incorrect numbers — is architecturally eliminated. Every dollar amount in a Socrates Crayon report was produced by tested, deterministic code covering federal and all 50 state tax jurisdictions, verified against 225 regression assertions.

4. Source Citation on Every Finding

Every finding in a Socrates Crayon report includes: the specific statute or regulation cited, the page and paragraph in the source document where the issue was identified, and a confidence score indicating the system's certainty.

What this means for compliance: Findings are verifiable. A compliance officer can trace any claim in the report back to the specific document provision and legal authority. No "black box" conclusions. No unsupported assertions.

5. Mandatory Uncertainty Disclosure

The system is architecturally required to disclose what it does not know. In our benchmark case ($26.9M estate), 47% of findings were flagged as low confidence with specific explanations of what additional information would be needed to increase certainty.

What this means for compliance: The system will never present a confident conclusion where insufficient data exists. Low-confidence findings are clearly labeled with the reason (missing documents, ambiguous provisions, jurisdiction uncertainty). This is the opposite of how consumer AI tools behave — they present uncertain information with false confidence.

6. Scope Limitation

Socrates Crayon produces analysis, not advice. Every report includes prominent disclosure that the output is for professional review only, does not constitute legal or financial advice, and should be verified by qualified counsel before client action. The system does not generate recommendations — it identifies issues and cites authority.

What this means for compliance: The tool functions as a research and analysis aid for licensed professionals, not as a client-facing advice engine. It occupies the same compliance category as Westlaw, Bloomberg Law, or a research memo from a junior associate — professional input that requires professional judgment before action.

Risk Comparison

Socrates Crayon vs. What Your Advisors Are Already Using

Your advisors are using AI for planning ideas right now. The question isn't whether they use AI — it's whether they use AI with controls or without them.

Control	Socrates Crayon	ChatGPT / Consumer AI
Client data stays in closed system	Yes	No — sent to OpenAI servers
Data used to train models	Never	Default yes unless opted out
Audit trail of inputs/outputs	Cryptographic attestation	No audit trail
Math verified by deterministic code	225 regression assertions	AI generates numbers
Statute citations verified	Cited to specific section	Known to hallucinate citations
Uncertainty disclosed	Confidence scoring on every finding	Presents uncertain info confidently
Scope limitation / disclaimers	Built into every report	None
Compliance review available	Architecture walkthrough on request	Consumer product

The Standard of Care Question

When Does Unreviewed Become Unreasonable?

$2M+

The Complexity Line

Above $2M, estate plans stop being simple. Multiple states. Multiple entities. Trust provisions that interact in ways nobody reviewed since the documents were drafted. An outdated beneficiary designation on a $2M IRA can cost a family more than the advisor earned on the relationship.

The Emerging Standard

As AI-powered analysis tools become available, the standard of care for estate planning review is shifting. When a tool exists that can identify $7.2M in value at risk from a $26.9M estate — with full citations and an audit trail — the question becomes: can you justify not using it?

Consider: If a client's estate plan contains a beneficiary designation conflict that costs the family $1.85M, and a review tool could have flagged it with a specific citation to the trust document and the applicable IRC section — what does the post-loss conversation with the client look like? What does the E&O claim look like?

Risk Reduction, Not Risk Creation

Socrates Crayon doesn't introduce AI risk — it mitigates existing review gaps. Every finding is verifiable. Every number is deterministic. Every uncertainty is disclosed. The tool creates a documented record that a thorough, multi-domain analysis was conducted — with an audit trail that survives regulatory scrutiny.

The compliance argument: Approving Socrates Crayon for your advisors isn't adding a technology risk. It's removing a documentation gap. The audit trail creates exactly the kind of auditable, verifiable record that compliance teams want to see when reviewing advisor activities around complex estates.

Show This to Your Compliance Team.