Scope and shared responsibilities
This policy applies to conversational agents, voice agents, chat experiences, automations, generated content, recommendations, and other AI-enabled features designed, hosted, or managed by HG Digital Lab.
HG is responsible for building and configuring the technology within the agreed scope. The client is the operator of the deployed experience and remains responsible for its business policies, customer relationships, content, contact lists, lawful basis for processing, sector obligations, and decisions made from AI output.
Transparency with people
An agent must accurately identify the business it represents and must not pretend to be a specific real person. The client must provide a clear AI or automated-assistant disclosure whenever required by law or when omission could materially mislead a reasonable person.
Material limitations, fees, eligibility conditions, and the commercial purpose of a conversation must not be hidden. A person must be able to ask for a human or another reasonable support channel when the context calls for one.
Calls, texts, consent, and opt-outs
The client must obtain and document every consent required for artificial or prerecorded voice calls, automated dialing, marketing texts, call recording, transcription, and follow-up. Consent must match the specific business, communication type, purpose, and destination required by applicable law.
The client must maintain required suppression and do-not-call lists, honor revocation and opt-out requests promptly, use permitted calling hours, provide required caller identification, and keep any required compliance records. An existing business relationship is not a substitute for consent where the law requires specific permission for an artificial voice, prerecorded message, recording, or marketing communication.
Human oversight and escalation
Clients must designate people who can review agent behavior, correct business information, receive escalations, and pause a workflow. Agents should transfer or defer when a request is outside scope, the user asks for a person, confidence is low, identity cannot be verified, or the situation presents safety, legal, financial, or reputational risk.
AI output must be reviewed before it is relied on for consequential action. Automation may support staff, but it does not replace the client’s responsibility to apply professional judgment and appropriate supervision.
Restricted and prohibited uses
HG agents may not be used to deceive, defraud, harass, discriminate, manipulate vulnerable people, impersonate a real person, conceal a sales purpose, evade platform safeguards, collect data unlawfully, or create content that violates law or another person’s rights.
Unless a separate scope has been reviewed and approved in writing, agents may not diagnose or treat medical conditions, provide legal or financial advice as a licensed professional, contact emergency services, make binding promises outside approved rules, or autonomously determine employment, housing, education, credit, insurance, health care, legal status, or access to essential services.
Data, recordings, and confidential information
The client must tell HG what categories of data an agent will handle and may provide only data it is authorized to use. Sensitive or regulated information may not be collected or processed unless the agreement expressly supports that category and the required safeguards are in place.
Where configured, an interaction may generate audio, transcripts, summaries, contact details, tool calls, and operational logs. The client is responsible for customer-facing privacy and recording notices. HG and its providers process that information only as described in the applicable agreement, configuration, and privacy terms.
Synthetic voices, likeness, and identity
No client may clone or imitate a real person’s voice, name, image, or persona without documented authorization sufficient for the intended use. A synthetic voice may not be used to create a false endorsement, impersonate an employee or public figure, disguise the source of a call, or confuse people about who is speaking.
HG may request proof of consent and may reject or remove a voice, persona, or use case that creates impersonation, fraud, safety, or rights concerns.
Models and third-party providers
AI services can depend on model providers, communications carriers, transcription services, hosting platforms, and other vendors. Their technical limits, acceptable-use policies, data terms, geographic availability, and pricing may apply and can change.
HG does not transfer ownership of third-party models or services. A handoff may require the client to open its own provider accounts, accept separate terms, replace unavailable tools, and pay provider charges directly.
Testing, monitoring, and enforcement
Before launch, the client must participate in testing material workflows and confirm that business information, transfer routes, permissions, and prohibited actions are correct. After launch, HG and the client may review samples, logs, incidents, and feedback to improve performance and controls, subject to the applicable privacy terms.
HG may pause or restrict an agent when it presents a credible legal, safety, security, provider, or abuse risk. Material policy violations may result in suspension or termination under the service agreement. Questions or concerns may be sent to sales@hgdigitallab.com.