Revision note: v1.2 changes Section 4.3 ONLY - removes named references to specific AI foundation model providers (previously “Anthropic Claude and OpenAI models”) from this public document, consistent with a decision to keep vendor-stack detail off the public website. No substantive change: Client's rights and Provider's obligations are identical. No other change from v1.1.
This AI Service Addendum (“Addendum”) is entered into between NexTech Advisors, LLC, doing business as Rivvet AI (“Provider”), and the Client identified in the applicable Order, and is incorporated into the Master Services Agreement between the parties as of that Order's Effective Date.
1. NATURE OF AI SERVICES
1.1 AI Agent Behavior. Rivvet’s AI Agents are large language model-based systems that generate responses dynamically based on context, training, and configured knowledge bases. AI Agents do not follow deterministic scripts. Each interaction is generated in real time and may vary based on the specific language, context, and circumstances of the conversation.
1.2 Inherent Limitations. Client acknowledges that AI systems have inherent limitations, including: AI Agents may occasionally produce responses that are inaccurate, incomplete, off-topic, or contextually inappropriate (“Hallucinations” or AI errors); AI Agents may not correctly interpret ambiguous, highly technical, or dialect-specific language; AI speech recognition (for voice AI) may produce transcription errors, particularly with names, addresses, and technical terms; AI Agents operate based on the knowledge base and instructions configured at the time of interaction and do not have access to real-time information beyond their configured integrations; and AI Agent performance may vary based on call volume, network conditions, and third-party service availability.
1.3 Not a Human Employee. AI Agents are software systems, not employees, agents, or contractors of Provider or Client. AI Agents do not have independent judgment, discretion, or authority beyond their configured parameters. Client is responsible for monitoring AI Agent performance and ensuring that AI Agent interactions meet Client’s standards.
2. AI DISCLOSURE OBLIGATIONS
2.1 Provider Disclosure Measures. Provider implements the following AI disclosure measures in all AI Agent interactions: outbound cold call AI Agents are configured to disclose their AI nature within the first sixty (60) seconds of any conversation; AI Receptionist Agents are configured to disclose their AI nature upon direct inquiry by the caller; and email communications facilitated by AI Agents do not impersonate human identities in a manner prohibited by applicable law.
2.2 Client Disclosure Responsibility. Client is solely responsible for ensuring that AI Agent disclosures comply with all applicable federal, state, and local laws in every jurisdiction where AI Agents operate on Client’s behalf, including Federal Trade Commission guidelines on AI-generated content and endorsements, state AI disclosure laws (including California A.B. 2602 and similar legislation), state telemarketing laws requiring caller identification and AI disclosure, and any industry-specific regulations applicable to Client’s business. Provider’s default disclosure measures are not a guarantee of compliance in any specific jurisdiction. Client bears ultimate legal responsibility for jurisdiction-specific compliance.
3. ACCURACY & QUALITY STANDARDS
3.1 Provider Commitments. Provider shall: configure AI Agents with accurate business information provided by Client during onboarding; implement quality gates and monitoring to detect and address systematic AI errors; provide Client with access to call recordings, transcripts, and performance analytics to enable Client’s own monitoring; respond to Client-reported AI errors within a commercially reasonable timeframe based on severity; and update AI Agent knowledge bases as reasonably requested by Client within agreed response timeframes.
3.2 No Accuracy Guarantee. PROVIDER DOES NOT WARRANT THAT AI AGENTS WILL: (a) PRODUCE ACCURATE OR ERROR-FREE RESPONSES IN ALL CIRCUMSTANCES; (b) CORRECTLY INTERPRET ALL CALLER LANGUAGE OR INTENT; (c) SUCCESSFULLY BOOK APPOINTMENTS IN ALL CASES; OR (d) PERFORM AT ANY SPECIFIC ACCURACY RATE BEYOND THE COMMERCIALLY REASONABLE TARGETS SPECIFIED IN THE ORDER. AI ACCURACY LIMITATIONS ARE AN INHERENT CHARACTERISTIC OF CURRENT AI TECHNOLOGY, NOT A DEFECT IN PROVIDER’S SERVICES.
3.3 Client Monitoring Obligation. Client is responsible for regularly reviewing AI Agent performance, including call recordings, transcripts, and booking outcomes. Client must promptly report to Provider in writing any AI errors, inaccurate responses, or compliance concerns. Provider is not liable for errors or compliance failures that Client failed to report within a reasonable time after they became reasonably discoverable.
4. INTELLECTUAL PROPERTY IN AI OUTPUTS
4.1 AI-Generated Content. Content generated by AI Agents during service delivery (email copy, call responses, chat messages) is generated on behalf of Client and is considered work product delivered under the Services. Provider retains no ownership interest in AI-generated content specific to Client’s campaigns and interactions.
4.2 Underlying Technology. The AI models, algorithms, prompts, workflows, training methodologies, and platform infrastructure used to generate content remain Rivvet IP as defined in the MSA. Client’s receipt of AI-generated outputs does not transfer any ownership in the underlying technology.
4.3 Third-Party AI Model Terms. Provider’s AI services are built in part on one or more third-party AI foundation models. Use of these services is also subject to the applicable terms of service of those model providers. Client acknowledges that limitations imposed by model providers may affect AI Agent behavior and that Provider cannot guarantee outcomes that conflict with model provider policies.
5. VOICE AI SPECIFIC PROVISIONS
5.1 Call Recording Consent. Client is solely responsible for obtaining all legally required consents for recording calls handled by AI Receptionist Agents, including compliance with federal and state wiretapping and recording consent laws. In states requiring two-party or all-party consent (including California, Florida, Illinois, Pennsylvania, and others), Client must ensure that callers are informed that their call may be recorded. Provider’s standard AI Receptionist configuration includes a general recording disclosure, but Client must verify that this disclosure satisfies the requirements of all jurisdictions in which Client operates.
5.2 Emergency Calls. AI Receptionist Agents are configured with emergency transfer protocols. However, Client acknowledges that AI systems may not correctly identify all emergency situations. Client is responsible for configuring appropriate escalation and transfer procedures and for maintaining human availability for genuine emergencies. Provider is not liable for AI failures to correctly identify or escalate emergency situations.
5.3 TCPA Compliance for Voice AI. Outbound calls made by AI Agents on Client’s behalf are subject to TCPA and applicable state telemarketing laws. Client is solely responsible for ensuring that: (a) all called parties have provided legally sufficient consent; (b) calling hour restrictions are observed; (c) Do-Not-Call list requirements are met; and (d) required disclosures are made. Provider implements calling-hour restrictions and DNC checking as technical safeguards, but Client bears ultimate legal responsibility for TCPA compliance.
6. RISK ALLOCATION FOR AI FAILURES
6.1 Provider Responsibility. Provider is responsible for maintaining platform infrastructure at the uptime levels specified in the MSA; promptly addressing systematic AI errors reported by Client; implementing and maintaining the quality gates and monitoring described in Section 3.1; and updating AI knowledge bases as requested by Client within agreed timeframes.
6.2 Client Responsibility. Client is responsible for providing accurate and complete business information for AI Agent configuration; monitoring AI Agent performance and promptly reporting errors; obtaining all required consents for AI interactions with Client’s customers; ensuring compliance with all applicable AI disclosure laws; configuring appropriate escalation procedures for complex situations; and not directing AI Agents to make representations that Client is not authorized to make.
6.3 Shared Responsibility. Outcomes that result from the combination of AI system limitations and Client’s business circumstances (including missed appointments, customer dissatisfaction from AI errors, or booking failures) are the shared responsibility of both parties. Provider’s liability for such outcomes is limited by the liability cap and exclusions in the MSA.
7. UPDATES TO AI SERVICES
Provider may update, modify, or improve AI Agent capabilities, underlying models, and platform features during the term of the MSA. Provider will provide Client with reasonable notice of material changes to AI Agent behavior or capabilities that may affect Client’s operations. Minor updates, bug fixes, and model improvements that do not materially affect AI Agent behavior do not require advance notice.
8. GENERAL
This Addendum is governed by the same governing law and dispute resolution provisions as the MSA. In the event of any conflict between this Addendum and the MSA, this Addendum controls with respect to AI-specific provisions. All other MSA terms apply in full.
IN WITNESS WHEREOF, the parties execute this Addendum via the signature exhibit attached to the applicable Order.