AI Disclosure — YourKendra

Our position

Every AI employee on YourKendra (Kendra, Marcus, Aria, Devon, Jordan, Riley) is AI. Not a human. Not a human-in-the-loop for routine interactions. This page documents how we disclose that fact and how our customers stay compliant.

Default behavior: If a caller or recipient asks directly — "Are you a real person?", "Am I talking to AI?" — every AI employee is configured to answer honestly and say yes. Customers can adjust the exact phrasing in their Configure modal, but the underlying disclosure cannot be disabled.

California SB 243 (effective 2026)

California requires AI chatbots to disclose when a consumer knowingly asks if they are interacting with AI. We comply by:

Every AI employee's system prompt includes a rule: if the user explicitly asks whether you are AI, a bot, or a human, you must truthfully say you are an AI assistant.
This rule cannot be overridden by customer configuration.
Logs of these disclosures are retained for audit.

Colorado AI Act (SB 205, effective Feb 2026)

Colorado regulates "high-risk" AI systems making consequential decisions (employment, housing, credit, insurance, healthcare). Most YourKendra uses cases — appointment booking, prospect outreach, invoice reminders — are not consequential decisions. However:

Riley (HR screening) can be used for employment decisions and therefore is a high-risk AI system under Colorado law. Customers using Riley in Colorado must: (1) publish a public statement disclosing their use of AI in hiring, (2) provide affected applicants with the opportunity to appeal decisions, and (3) complete an impact assessment before production use.
We provide a Colorado-specific Riley configuration checklist inside the dashboard. Enterprise customers operating in Colorado should email hello@yourkendra.com for our AI impact-assessment template.

EU AI Act (phased enforcement 2024–2027)

The EU AI Act classifies systems into Unacceptable Risk (banned), High Risk (strict controls), Limited Risk (transparency), and Minimal Risk. YourKendra customer use cases map as follows:

Limited Risk — Kendra voice agents, Aria content, Jordan EA: customers must clearly inform end-users they are interacting with an AI system (Art. 50). Our default system prompts meet this requirement.
High Risk — Riley used for employment decisions (Annex III(4)). Customers using Riley in the EU must register with the EU AI database, complete a conformity assessment, and maintain human oversight of hiring decisions. We are not currently enabling Riley for EU customers pending our own ISO 27001 + AI-specific conformity work (targeted 2027).

Other state laws we track

Illinois (HB 3773, effective 2026) — employer AI hiring tool disclosure. Riley users covered.
New York City Local Law 144 — bias audits for automated employment decision tools. Riley users must complete an annual bias audit before use on NYC candidates.
Utah SB 149 (2024) — generative AI disclosure for regulated occupations. Customers in medical, legal, or financial services using Kendra must include a one-time AI disclosure in first-touch communications.
Texas HB 149 (Responsible AI Governance Act, 2025) — state-employee and government-contact AI use restrictions. Does not apply to private-sector customers but we track it for our own procurement readiness.

What YourKendra commits to

Every AI employee will always truthfully answer "are you AI" — this is a floor, not a toggle.
We will never train foundation models on customer data.
Customers can export every AI interaction's full transcript at any time.
We maintain a changelog of every prompt or policy change that affects disclosures (/changelog).
Enterprise customers in regulated industries get human-in-the-loop configurations on request.

Questions, audits, or legal requests

Legal and compliance counsel — reach out to hello@yourkendra.com for our latest compliance pack (sub-processor list, DPA, BAA template, AI impact assessment template, conformity documentation). We respond within 2 business days.