AI Policy

Dale AI, Inc. ("Dale AI," "we," "us," or "our") builds an AI-powered platform that helps real estate agents manage their business through voice, text, and automated workflows. This policy explains how we use artificial intelligence in our product, what it can and cannot do, and how we protect you and the people you serve.

This policy applies to all users of the Dale AI platform ("you" or "Agent") and to the clients, leads, and contacts ("Contacts") who interact with Dale AI on your behalf. It supplements our Terms of Service and Privacy Policy, which govern your overall use of the platform.

How Dale AI Uses Artificial Intelligence

Dale AI uses third-party large language models (LLMs) to power three distinct AI agents that work together to support your real estate practice:

Receptionist

When a caller reaches your dedicated phone number and you are unavailable, the Receptionist answers on your behalf. It can take a message or schedule a callback request. It does not access your CRM data, does not provide information about you beyond your name, and does not make commitments about your availability or response time.

Voice and SMS Assistant

When you call your Dale AI number or text with Dale AI, the Assistant helps you manage your CRM hands-free. It can look up contacts, log interactions, schedule follow-ups, search your email and calendar, send text messages on your behalf (with your explicit confirmation), initiate outbound calls, create opportunities, and browse property listings. Every action that modifies your data or sends a communication to a Contact requires your direction.

Summarizer

After a call ends, the Summarizer reviews the call transcript and automatically extracts CRM data: call summaries, follow-up items, new contact information, life events, relationship connections, opportunity updates, and behavioral observations. The Summarizer works from the transcript — it does not require or use audio recordings. This processing happens asynchronously and does not involve a live conversation.

AI Disclosure and Transparency

Dale AI identifies itself as an AI

Dale AI never pretends to be human. When the Receptionist answers a call on your behalf, it identifies itself as your assistant — not as you, and not as a human being. Your Contacts interact with a clearly synthetic voice.

No voice cloning

Dale AI does not clone, replicate, or simulate your voice. The AI uses its own synthetic voice for all spoken interactions. There is no feature, current or planned, that would allow Dale AI to impersonate your voice or any other person's voice.

Automated communications are identified

When Dale AI sends a text message on your behalf, it does so from your assigned Dale AI phone number. You must review and confirm every outgoing message before it is sent. Dale AI will never send a message to one of your Contacts without your explicit approval during the conversation.

How AI Decisions Are Made

Dale AI does not make autonomous decisions about your business. It operates as a tool that executes your instructions and surfaces information to help you make better decisions.

Call summaries and interaction logging

After each call, the Summarizer generates a written summary and classifies the conversation by sentiment, urgency, and topic. These summaries are based solely on what was said during the call. The AI may miss nuance, misattribute a statement to the wrong speaker, or misinterpret context. Summaries are a convenience — they are not a substitute for your own judgment about what happened on a call.

Contact enrichment

The Summarizer extracts contact details (phone numbers, email addresses, employment, life events, relationships) from call transcripts. It only records information that was explicitly stated during the conversation. It does not infer or guess at details that were not discussed. When the AI lacks sufficient information to complete a record, it creates a pending action for you to resolve rather than filling in placeholder data.

Behavioral observations (DISC scoring)

Dale AI observes communication patterns during calls and scores Contacts on the DISC behavioral model (Dominance, Influence, Steadiness, Conscientiousness). These scores are based on how a Contact communicates — their pace, directness, warmth, and attention to detail — not on demographic characteristics. DISC scores are observational estimates, not psychological assessments. They carry a confidence rating that reflects how much conversational data informed the score.

Lead qualification and prioritization

Dale AI may prioritize contacts for follow-up based on interaction history, recency, and behavioral signals. These are suggestions to help you manage your time. They do not determine who receives your services, and you always have full control over which Contacts you engage with and when.

Follow-up scheduling

When the AI detects commitments or next steps in a conversation, it creates follow-up items with due dates and confidence scores. High-confidence items (exact date and time specified) may be automatically added to your connected calendar. You can modify or delete any follow-up at any time.

Human Oversight

Dale AI is designed to keep you in control:

• No unsupervised outreach. Dale AI will never contact your clients on its own initiative without your explicit instruction and confirmation during an active conversation.
• Confirmation before sending. Every outgoing text message is read back to you for approval before it is sent. Every outbound call is initiated at your request.
• You review AI output. Call summaries, contact updates, and behavioral scores are saved to your CRM for your review. You can edit, correct, or delete any AI-generated data.
• Receptionist boundaries. The Receptionist can only take messages and schedule callback requests. It cannot share your personal information, make promises on your behalf, discuss pricing, or provide real estate advice.
• Pending actions for uncertainty. When the AI is unsure about a piece of information, it flags it as a pending action for you to resolve rather than guessing.

Data Usage and AI Training

How we use your data for AI training

Dale AI uses data generated through the Services — including transcripts, conversation summaries, interaction patterns, and call recordings (when enabled by the User) — to train, fine-tune, and improve our AI models. This training is what allows Dale AI to get better at transcription, conversation quality, lead qualification, and understanding real estate workflows over time. By using the Services, you consent to this use of your data for model improvement.

We implement safeguards in our training process: training data is de-identified where feasible, we do not sell or license training data or trained models to third parties, and access to training datasets is restricted to authorized personnel.

How your data flows through AI

When you interact with Dale AI, the relevant context (your CRM data, call transcript, contact profiles) is sent to the AI model as part of the conversation prompt. The model processes this information to generate a response. Separately, conversation data may be used in our training pipelines to improve model performance over time.

Third-party AI providers

We use third-party AI services (currently from Anthropic and potentially other providers) to power certain features. Our agreements with these providers govern data handling, including whether data processed through their APIs may be used for their own model training. We select providers based on their data protection commitments and evaluate these commitments regularly. We will update this policy if our provider relationships change in ways that materially affect data handling.

Data residency

All Dale AI infrastructure runs on Amazon Web Services (AWS) in the United States. Your data is stored and processed within US data centers.

AI Limitations and Accuracy

AI is a powerful tool, but it is not infallible. You should be aware of the following limitations:

Transcription accuracy

Dale AI uses speech-to-text technology to transcribe calls. Transcriptions may contain errors, especially with names, addresses, technical terms, heavy accents, background noise, or poor call quality. The AI may occasionally misidentify which speaker said what.

Summary and extraction accuracy

Call summaries are generated interpretations, not verbatim records. The AI may miss important details, misunderstand context, or incorrectly attribute statements. Contact information extracted from calls (phone numbers, addresses, employment details) should be verified before relying on it for important communications.

Behavioral scoring limitations

DISC behavioral scores are estimates based on limited conversational data. A single short call provides very little signal. Scores improve in accuracy over multiple interactions but should never be treated as definitive personality assessments.

AI-generated text

When Dale AI composes text messages or email drafts, the content is generated by AI based on your instructions. Always review AI-generated text before sending it to ensure it accurately reflects your intent, tone, and the specific details of your business relationship with the recipient.

Not a substitute for professional judgment

Dale AI does not provide legal, financial, or real estate advice. AI-generated suggestions about lead prioritization, follow-up timing, or opportunity management are informational aids, not professional recommendations. You remain responsible for all business decisions and client communications.

AI and Bias

We take the risk of AI bias seriously and address it through the following practices:

What we do

• Behavioral, not demographic. DISC scoring and communication style adaptation are based on observed conversational behavior — word choice, pace, directness, expressiveness — not on a Contact's name, voice characteristics, accent, gender, age, race, or any demographic attribute.
• No discriminatory screening. Dale AI does not make decisions about who deserves your attention based on protected characteristics. Lead prioritization is based on interaction history, recency, and explicitly stated intent — not on who the person is.
• Conservative defaults. When the AI cannot determine information with confidence, it defaults to neutral options or flags the item for your review rather than making assumptions.

What we monitor

We regularly review AI outputs for patterns that could indicate bias in how contacts are classified, scored, or prioritized. If you observe AI behavior that appears biased or unfair, we encourage you to report it so we can investigate and address it.

Security

Protecting AI interactions

• All communications between Dale AI and third-party AI providers are encrypted in transit.
• Call transcripts and CRM data are encrypted at rest.
• AI model providers do not have persistent access to your data — context is provided per-request and not stored by the provider.
• Access to your CRM data is governed by role-based access controls with row-level security enforced at the database layer.
• Session authentication is required before the AI Assistant can access any CRM data. The Receptionist operates without CRM access by design.

Voice security

• Inbound callers are not automatically granted access to your CRM data. Unknown callers interact with the Receptionist, which has no CRM access.
• Authenticated access (the full Assistant experience) requires verification through your account.
• Audio recording is optional and controlled by the User. When enabled, retention periods are governed by our Privacy Policy.

Your Control Over AI Features

You have control over how AI operates in your Dale AI account:

• Review and edit AI-generated data. All summaries, contact updates, behavioral scores, and follow-up items can be reviewed, edited, or deleted from your CRM.
• Pending action management. You can resolve, skip, or dismiss any pending action the AI creates.
• Custom instructions. You can provide custom instructions that shape how the Receptionist interacts with your callers.
• Communication confirmation. You must explicitly approve every outgoing message and outbound call.

How We Improve AI

Product improvement

We analyze aggregated, anonymized usage patterns to improve the Dale AI product — for example, understanding which types of follow-ups are most commonly created or which AI suggestions are most frequently accepted. This analysis does not involve sharing your identifiable data with AI model providers.

Your feedback

When you edit an AI-generated summary, correct an extracted contact detail, or dismiss a suggestion, that feedback helps us understand where the AI falls short. We use these patterns to improve our processing logic. Your specific data corrections remain private to your account.

Responsible AI Principles

Our approach to AI is guided by these principles:

1. Transparency. We tell you and your Contacts when they are interacting with AI. We explain what the AI does and does not do.
2. Human authority. AI assists; you decide. Dale AI never takes consequential action without your knowledge and approval.
3. Data stewardship. We are transparent about how we use your data, including for AI training. We de-identify training data where feasible and never sell or license it to third parties.
4. Accuracy over speed. When the AI is uncertain, it asks rather than guesses. We prefer a pending action over bad data.
5. Fairness. We design AI features to evaluate behavior and intent, not identity. We actively monitor for and work to eliminate bias in AI outputs.
6. Security by design. AI access to your data is scoped, authenticated, and encrypted. We follow the principle of least privilege — each AI agent has access only to the data it needs to perform its function.
7. Continuous improvement. We regularly evaluate our AI systems, update our practices based on evolving best practices and regulations, and communicate material changes to our users.

Contact Us

If you have questions about this AI Policy, want to report a concern about AI behavior, or want to understand more about how AI is used in a specific feature, contact us at:

Email: support@dale.app

This policy should be read alongside our Terms of Service and Privacy Policy. In the event of a conflict between this policy and the Terms of Service, the Terms of Service govern.

Dale AI, Inc. reserves the right to update this policy as our product, technology, and the regulatory landscape evolve. We will notify users of material changes through the platform or by email.