Artificial Intelligence Policy
How Vesperlight Real Estate Services uses AI responsibly, securely, and transparently in service of our clients.
Vesperlight Real Estate Services, Inc. employs artificial intelligence as a tool to enhance the quality, speed, and depth of service we deliver to our clients. We believe in being transparent about how we use these technologies, what safeguards are in place, and where the boundaries lie between machine assistance and human judgment.
This policy applies to all AI and machine learning systems used in our operations, including but not limited to: document analysis, market research, data processing, email communication assistance, and property intelligence gathering. It is reviewed and updated as our practices evolve.
How We Use AI
Our AI systems serve in supporting roles under human oversight. The following principles govern every deployment:
Human Oversight
Every AI-assisted output that reaches a client, partner, or external party is reviewed and approved by a human member of our team before it is sent. Our AI does not act autonomously in client-facing communications.
Data Security
Sensitive client documents, including lease agreements, financial statements, and proprietary materials, are processed using local, on-premises models. They are never transmitted to public cloud-based AI services.
Transparency
We disclose our use of AI to parties when it is relevant to the engagement. We do not represent AI-generated analysis as exclusively human work, nor do we conceal AI involvement when it materially shapes deliverables.
Boundaries
Our AI systems do not make binding decisions, provide legal or financial advice, negotiate terms, or substitute for the professional judgment of our licensed real estate professionals. AI assists; it does not decide.
Privacy Protection
We do not share client data with third-party AI platforms without explicit consent. Vendor relationships involving data access are governed by written agreements specifying how data is handled, stored, and protected.
Continuous Review
Our AI workflows are subject to weekly internal audits. Outputs are checked for accuracy, tone, and compliance with our operational standards. Issues are flagged, documented, and corrected through iterative improvement.
Safeguards and Verification Gates
Our AI operations are governed by a multi-tier verification architecture. No AI-assisted communication, data output, or client deliverable proceeds without passing through each of the following gates:
- Draft and Design: AI-assisted content is drafted for its specific purpose, following established templates and operational requirements.
- Authenticity Review: Every output is checked for accuracy, tone, and authenticity. AI-generated content that does not meet our human-quality standards is revised before proceeding.
- Operational Truth Verification: Hard-gate checks verify that email recipients are confirmed contacts, no duplicate communications are sent, signatures are correct, content is accurate, and business-hour rules are respected. Any gate failure halts the process.
- Human Approval: A human team member reviews and approves the final output before it is externalized to any client, broker, or third party.
- Audit Trail: All AI-assisted communications and data operations are logged with timestamps, recipient provenance, and verification results for retrospective review.
What this means in practice: If you receive an email from our office, a human reviewed it before it was sent. If you receive market analysis from our team, a licensed professional stands behind it. Our AI is a tool that helps us work faster and more thoroughly, but it does not replace the judgment, accountability, and personal relationships that define our practice.
What We Process and How
Local processing for sensitive documents: Lease agreements, proprietary financial documents, and client-identifying materials are processed exclusively on our local infrastructure using private models. These documents are never submitted to public cloud AI services.
Cloud AI for non-sensitive tasks: Public AI models may be used for general research, market intelligence, document formatting, and similar tasks that do not involve confidential client information.
Vendor data agreements: Where third-party platforms have access to our data (CRM, email, website), we obtain written confirmation of how AI features within those platforms interact with our data. We disable AI features that route our data to external models without our explicit authorization.
No unauthorized data sharing: We do not share client contact information, property details, or transaction terms with AI platforms or third parties unless explicitly authorized by the client or required by the engagement.
Compliance Framework
Our AI usage complies with all applicable California real estate regulations, including the California Business and Professions Code governing licensed real estate activity. AI systems do not hold real estate licenses and do not perform acts requiring licensure.
We adhere to the California Consumer Privacy Act (CCPA) and applicable data protection regulations. Clients may request information about what data has been processed, how it was used, and request deletion of their data from our systems.
Our AI-assisted communications are subject to the same professional standards as all correspondence from our office. We do not use AI to generate misleading, deceptive, or unverified claims about properties, market conditions, or transactions.
Questions about our AI practices? We welcome them. Contact Robert Brown, President, at rbrown@vesperlightre.com or (310) 601-7358.