AI Disclosure & Responsible AI Statement

BOS uses AI in:

• the voice receptionist that answers, qualifies, and routes phone calls;
• dispatching, with manual, AI-assisted, and fully-automated modes;
• customer-relationship workflows: drafting messages, summarizing call notes, surfacing follow-ups, scoring leads;
• sales coaching, providing real-time feedback during live calls;
• back-office tasks: drafting quotes and invoices, suggesting catalog matches, summarizing job history;
• B2B lead enrichment through scraping and analysis pipelines;
• the agent orchestration system that coordinates multi-step workflows.

AI outputs (voice responses, drafted messages, dispatch suggestions, summaries, recommendations, quotes, transcripts) are assistive. They can be:

• factually inaccurate,
• incomplete,
• ambiguous,
• inconsistent across runs, or
• mistaken about which person, property, or job they refer to.

Customer is responsible for reviewing AI outputs before relying on them for material business actions, including:

• dispatching a technician to a property,
• agreeing to a price or scope of work,
• modifying or canceling a customer’s appointment,
• sending invoices or processing payments,
• making employment decisions,
• making representations to regulators, insurers, or counterparties.

Where the consequence of being wrong is high, Customer should put a human in the loop before acting.

The voice receptionist included with BOS announces AI involvement at the start of voice interactions. Customer must keep this disclosure enabled and unaltered. Customer must not configure the voice agent to claim to be a human, to misrepresent its capabilities, or to mimic a real person’s voice or vocal style without that person’s documented consent.

Several US states (including California, Utah, and Colorado) have adopted, or are adopting, laws requiring disclosure of AI involvement in commercial communications and content. Customer is responsible for compliance with the disclosure rules of any state where it operates.

This is the central commitment.

• All distillation, refinement, fine-tuning, vector memory, and skill content evolves on Customer’s Appliance.
• Customer Data is not aggregated, anonymized, pseudonymized, or otherwise extracted from one Appliance to train models that benefit any other Appliance.
• Software updates that we ship contain only Dino-authored code, configuration, and prompt content. They never contain Customer Data.
• If we ever propose to change this commitment, the change will be opt-in only, after at least thirty (30) days’ written notice to Customer.

This is also reflected in our Privacy Policy and DPA.

Customers may opt in to a cloud language-model fallback for selected workloads. When enabled:

• the routed prompt and completion are processed by the third-party provider under that provider’s terms;
• the cloud provider is identified on the Subprocessor page;
• Customer pays for cloud usage either directly or via metered billing;
• Customer’s per-customer commitment in Section 4 still applies on the Appliance, but is subject to the cloud provider’s data-use policies on the cloud side. Customer should review those policies before opting in.

Cloud LLM fallback is off by default.

We design BOS to support fair, respectful interactions. We acknowledge that no AI system is bias-free. Customer must not configure or use the Service to:

• make automated decisions about individuals on the basis of protected characteristics (race, color, religion, sex, gender identity, sexual orientation, national origin, disability, age, or veteran status);
• make automated employment, housing, lending, or credit-related decisions, where AI use is regulated;
• generate content that demeans, degrades, or depicts violence against specific persons or groups.

If Customer’s use case involves automated decision-making about individuals, Customer is responsible for performing applicable impact assessments, providing required disclosures and opt-outs, and complying with state laws (Colorado AI Act and similar).

The Service must not be used to:

• generate audio, video, or text that purports to be from a real person without that person’s documented consent;
• clone, mimic, or evoke the voice or likeness of a real person without that person’s documented consent;
• impersonate any business, government agency, or official.

We treat violations as a material breach of the Acceptable Use Policy.

Customer owns Customer Data and the AI outputs generated from it. See Terms of Service Section 7.

If you are an end-user, customer, employee, or third party and you believe an AI output from a Dino-powered system has caused or is about to cause harm, contact us at [email protected] with the subject line “AI Concern” and we will investigate and coordinate with the relevant business operator.

We will update this Statement as the AI features in BOS evolve and as the legal landscape develops. Material changes will be flagged on dinoaihost.com.

Dino RD&C LLC

5830 E 2nd St

Casper, WY 82601, USA

[email protected]