Privacy Policy

1. Website visits and technical information

When you visit, our infrastructure receives information needed to deliver and secure the request, such as your IP address, browser and device information, requested page, and request timing. Hosting and network providers may process and retain operational logs. We use this information to serve the site, maintain security, diagnose failures, and prevent abuse.

The site can request fonts from Google Fonts, which exposes the browser request and associated technical information to Google. Necessary browser storage and other website technologies are described below. Technical information and pseudonymous identifiers may still be personal information; hashing an identifier does not necessarily make it anonymous.

2. Contact forms, inquiries, and support

When you submit a contact form, we collect the name, email address, optional phone number, business type, and message you provide. When you specifically authorize William to send your contact details to our team, those details become an inquiry record. We use the information to respond, follow up about the request, and maintain the associated business correspondence.

Inquiry records are stored in our application database. An internal email can contain the details you submitted. When an internal text alert is configured, it can also contain inquiry contact information. The recipients and their email or telephone providers may retain delivered copies separately from the application record.

We contact you about your inquiry using the channel and permission you provide, subject to applicable law. Submitting information is not blanket permission for every automated or marketing call or text. You can ask us to stop optional follow-up by replying or emailing hello@helloanswered.com. Necessary responses to a request or security and legal notices may still be sent where appropriate and permitted.

Support and privacy correspondence may include your contact details, the issue you report, proportionate verification information, and the steps taken to resolve it. Do not send passwords, full payment details, or unnecessary sensitive records. Ordinary inquiry and support records do not currently have an automatic deletion deadline; we retain them for handling the relationship, resolving issues, and necessary legal or business records, subject to applicable rights and deletion requests.

3. William on the public website

William uses OpenAI to generate replies from your messages and active conversation context. The request also uses a keyed session identifier. It is not your raw name, email, or network address. Information you type into a message is nevertheless part of the message sent to the provider, so do not type sensitive information.

hello, answered. does not save the general chat transcript in its application database. Active conversation context is held temporarily in server memory, with an inactivity expiry of up to 12 hours after the last message. A restart can remove it sooner. This local window does not determine OpenAI's separate retention.

If you explicitly ask William to send your contact details to our team and agree to be contacted, those details are saved as an inquiry. To prevent duplicate chat lead submissions, the application can retain a one-way hash of the submission identifier for approximately 24 hours, without message content. This is a hash of the submission identifier, not a hash of your contact details, and can be associated with the corresponding submission.

4. Browser and phone voice demos

Browser voice demos use OpenAI or xAI for the selected demo route. Audio can travel directly between the browser and OpenAI or through our application to the selected AI provider. Public phone demos use Twilio for telephone transport and OpenAI or xAI for the selected voice route. The selected AI provider processes the conversation to generate replies; our application processes the session and routing information needed to operate and secure the demo. These providers also handle information under their applicable terms and settings. Naming both AI providers does not mean every demo sends its content to both.

Our application does not save voice-demo recordings or transcripts in its application database and does not enable application-level content tracing for the demos. This does not mean every provider immediately discards the conversation or that the call is unprocessed or unrecorded everywhere. Provider conversation history, request and response content, and operational records may have different retention, described in section 9. Please use sample information in demos.

To enforce public phone-demo limits, our application stores one-way keyed hashes of the caller number and provider call ID for approximately 24 hours. Raw caller and call identifiers are processed transiently for routing, verification, and call control. These demo-limit records do not persist the raw identifiers. Infrastructure and telephone-provider records are separate.

Browser voice-demo limits use a one-way keyed hash of the visitor network identifier in process memory for up to approximately 24 hours. On an OpenAI-controlled phone route, separate one-way keyed call and token hashes and limited receipt timestamps can be retained for approximately 31 days to prevent replay. An OpenAI browser route can transmit a pseudonymous network-derived safety identifier to OpenAI. Changes to the selected processing route must be reflected in the notice for that feature.

5. Private setup and optional AI assistance

Private setup uses your verified email, the business information and instructions you enter or approve, selected plan and coverage, voice choice, approvals, and operational records needed to save and protect the setup. The current short William path asks for your business name and type and uses scripted prompts to propose a profile for review. That proposal does not itself send an interview to an AI model. The manual alternative collects the business facts and instructions you enter and approve. Either path can offer the separate voice practice feature described below.

A verification code expires after 10 minutes. Unverified email challenges are scheduled to expire within 24 hours. A secure, HTTP-only browser cookie for up to 30 days lets this browser return to the same setup. The cookie's expiry is different from the saved setup's retention period.

If you continue an older guided William setup interview, we ask for a separate confirmation before his first AI-generated reply. Creating a new short setup uses the scripted profile flow described above. On that guided path, OpenAI receives your current message, up to four recent private-setup messages held temporarily in server memory, answers you already approved, a published question and safety guide, and a one-way keyed safety identifier. The optional voice practice feature is separate from this interview. We do not add your raw email address, setup reference, or raw network address to those interview requests. If you put personal information into your message, it is included as part of that message.

hello, answered. does not durably save raw William interview turns or unapproved suggestions in its application database. The active page and server process can hold context temporarily. We store your assistant-use confirmation, explicitly approved answers, approval receipts, coverage choice, applicable guide version, and bounded content-free usage and active-request records. The guide does not grant William access to your CRM, calendars, or contacts, or permission to activate service. The secure manual form makes no William provider request.

If you start an offered voice practice conversation, OpenAI receives your microphone audio, conversation content and the draft business profile, selected voice and answering settings needed to rehearse the agent's responses. This can include draft information you have not yet finally approved. Our application processes the authenticated preview session and its bounded control and usage records. The practice route is configured not to store its AI session and has no tools to save callbacks, book or send. We do not save a practice recording or full transcript in our application database. This does not determine OpenAI's separate retention under its applicable terms and settings.

A short text-to-speech sample is a different preview. When that feature is available and you request it, only the approved preview script is sent to OpenAI's text-to-speech service. The application stores that generated sample encrypted and schedules it to expire within 24 hours. A setup-only review generates no such sample. Availability of a voice practice conversation does not mean a generated sample has been saved or approved.

The current separately activated Sophie Phone route uses OpenAI for voice AI processing and Twilio for telephone transport, as described in section 7.2. Neither a setup sample nor a practice conversation activates that service.

Submitting a setup stores the selected service and a reserved seven-day free-service entitlement, including its version and reservation time. The start and end remain unset while reserved. For a new self-service offer, separate readiness-gated activation establishes the service state and free-period deadline shown in the account. Earlier reserved offers keep any separately accepted managed-QA and live-confirmation-email conditions. Submission can also produce a receipt email and internal review alert. Authorized administrators can review the contact and business information needed to handle the setup. If you continue to an account flow offering card saving, Stripe handles that separate step under section 7.1; saving a card does not itself start service or the free period.

6. Setup retention and Forget

If you do not become a customer, we schedule the setup's contact information, approved business facts, approval records, and remaining preview material to expire within 90 days after you first verify your email. The 90-day period does not restart when you return or make an edit. Generated audio is scheduled to expire within 24 hours, and unverified email challenges within 24 hours.

For these scheduled setup expiries, the application sets the database deadline one hourly sweep early, runs cleanup at startup and hourly while running, and uses database expiry as an independent backstop. An unexpected application or database interruption can delay physical removal until cleanup resumes. These periods describe application cleanup, not a promise that every provider or delivered copy is erased at the same moment.

You can use the setup's Forget option to delete that setup's content earlier. If you cannot access the setup or its browser session, email hello@helloanswered.com for an appropriately verified request. Forget applies to the setup you can access; it is not a search of every device, prior setup, inquiry, or separate account associated with you.

To enforce preview limits after Forget, we retain a one-way keyed abuse-prevention identifier derived from the verified email with a scheduled expiry within 48 hours. It is not the raw email, but the system can match it when the same email is entered again. Forget deletes the setup content but does not delete this short-lived identifier before it expires. William limits similarly use keyed email and network identifiers and content-free usage and active-request records scheduled to expire within 48 hours; some other rate-limit records expire after one hour. These records contain no raw interview text, approved business facts, raw email, or raw network address.

A setup-code email requested before or during Forget, or a submitted-request receipt already accepted for delivery, may still arrive afterward. Neither can reopen or authenticate the forgotten setup. Forget does not recall an email or text already delivered or directly erase a provider's independent records. We may keep content-free deletion, security, and aggregate measurement records for the associated protection and accountability purposes. Any later customer account has a separately explained lifecycle; this policy does not promise that setup expiry cancels or deletes a future activated service.

7. Business orders and separately enabled customer services

7.1 Orders, account administration and payments

When you enter into a business service order, we process the business and representative identities, contact details, approved instructions, selected services, recipients, pricing and billing choices, and records of the order and its acceptance. Agreement evidence includes the documents and versions supplied, the order, who accepted for the business, when and how acceptance was given, and the permissions relevant to the transaction. This is separate from an unconverted setup’s temporary records. A label or business-profile approval is not treated as permission for an undisclosed charge or unrelated use of information.

Where an account feature is offered, we process the account email and access-verification records, approved profile, notification destinations, configuration, service and subscription state, support correspondence and related operational records. Signing an order or storing a payment method does not by itself activate phone or texting service. The existing public pre-line setup still uses the verified email destination described in section 5.

If payment collection is separately offered and authorized, Stripe collects payment-card details through its secure payment form, which may appear within our account page or on a Stripe-hosted page. We keep the Stripe customer, checkout, payment-method and subscription references and billing or transaction state needed to administer the account. Our application does not receive full card numbers or card security codes from that form. Stripe also handles information for its own payment, fraud-prevention and legal purposes. The accepted Order must explain any recurring charge before it is authorized; this policy is not charge authorization.

7.2 Customer phone calls

For the current separately activated phone route, Twilio processes caller and routing information and carries the call media to OpenAI. OpenAI receives the conversation and relevant approved business instructions to generate replies and perform the supported message-taking function. A different enabled route, including an xAI route, must be identified in the applicable notice and customer service record before use.

A captured message can contain the caller's name, callback number, reason for calling and other details the caller provides. The application also processes call and service identifiers, admission and call times, duration, call state, capture time and operational or usage records. These records can exist even if a call does not produce a captured message. Where enabled, captured messages can be viewed in the authenticated business account and delivered to its approved account email. The recent-call view shows call details and captured callback messages. Optional recordings, full transcripts and AI summaries are described in section 7.10. A different destination or integration applies only if expressly offered, authorized and enabled.

Unless the business enables the optional call-saving feature in section 7.10, our application does not intentionally save an audio recording or full transcript from this phone route. The AI and telephone providers process call content and can have separate history, security or other retention under their applicable terms and settings. A provider’s recording or retention cannot be inferred from whether our database stores an audio file. The business must provide an accurate notice for its actual call path and any required permissions, and we remain responsible for notices and obligations that apply directly to us. Contact us or the business if you need help identifying the relevant handling.

The local encrypted call-message content is scheduled to clear 30 days after call admission. That clears the content envelope, not the entire call record: call, account and service references, timing, duration, usage and replay-protection metadata can remain under the purposes and deletion process below. Copies already emailed to the business are separate. Scheduled cleanup can be delayed by an interruption until it resumes; provider records do not inherit this 30-day period.

7.3 Operational texting with the business owner

If owner texting is separately enabled and the owner pairs a number through the supported permission process, Twilio handles the mobile number, texts and delivery information. The configured AI service can receive the owner’s message, approved business profile, recent saved call-message information and conversation context to generate a reply. This is a distinct communication program from texts that outside customers send to a service number.

The described owner-texting feature clears older message content and specified contact or delivery fields after seven days. Account/number pairing, permission, opt-out, usage, delivery receipts and replay-protection records have separate purposes and can remain. Seven days is not a promise to erase every number, delivery identifier or suppression record, and it does not govern Twilio’s copies.

Enrollment must identify the sender, purpose and available support. You can revoke permission by a reasonable method that clearly communicates the request, including STOP or contacting hello@helloanswered.com. We apply the request to the communications it covers under applicable law and respect applicable provider suppression controls. Restarting messages uses the supported permission process; a generic setup or privacy acknowledgment does not authorize restarting a suppressed program. We do not sell mobile numbers or text-message permissions or provide them for another party’s marketing.

7.4 Texts sent to a business service number

If outside-customer automated texting is separately enabled, Twilio and the configured AI provider process the sender’s number, message, approved business context, recent conversation and generated replies. The described automated conversation feature replies to the sender; it does not deliver that conversation to the business. An automated reply is not evidence that a person received or read a message. Any additional delivery feature must be separately offered and accurately described.

This feature schedules conversation content to clear after seven days and retains separate delivery, replay, service and suppression records where needed for those purposes. It may retain a number or a linkable identifier for permission or delivery protection after content clears. Provider handling is separate. STOP and other reasonable revocation requests must be honored for the applicable program; starting a conversation is not blanket permission for unrelated marketing, automated calls, or a different sender.

7.5 Business and messaging registration

If the selected registration process requires it, we collect legal business name, business type, industry, address, website, the applicable business tax identifier, representative name, title and contact details, messaging purpose, opt-in description, policy links and sample messages. Use the designated fields for requested registration information; do not place tax identifiers or identity documents in general AI instructions. The fields and eligibility depend on the registration route actually offered.

Registration information is submitted through Twilio to the relevant registration authorities and carriers to verify and administer the program. They may contact a representative for confirmation and retain information under their own requirements. Our chosen phone configuration may depend on this registration even if a customer does not purchase texting; that is not a statement that all voice calls legally require A2P registration.

Registration content and provider references can be retained separately from a setup, including while an application is pending, active, rejected, being corrected or being retired. Local deletion does not itself cancel a carrier registration, release a number or erase a provider’s independent records. We handle correction or withdrawal requests through the applicable local and provider processes and explain any required retained records.

7.6 Roles, retention and requests for customer-service records

The business customer directs the purpose of its communication program and supplies the approved information and recipients. We process Customer Data on its documented instructions under the accepted order and processing addendum, including the selected communication, capture, delivery, support and security functions. Each party retains its direct legal obligations. Neither the customer nor a contract can grant ownership of another person’s privacy rights.

We do not use customer or caller content to train our own general-purpose AI models. We do not opt Customer Data into a provider’s optional general-purpose training program without a separate lawful agreement and any required permissions. A provider’s independently permitted activity is a separate question described in section 9; it must not be used to circumvent our processing restrictions or applicable law.

Approved account/profile information, orders and acceptance evidence, registration, call or message metadata, usage, billing, permission/suppression and security records can have different lifecycles from the content periods above. We keep our independently controlled records only as reasonably needed for their described purpose and applicable obligations. For data processed on the business’s behalf, we follow its lawful instructions and the addendum, including return or deletion at its direction or when processing ends, unless law requires retention. We do not apply a blanket account-lifetime or seven-year rule. We identify the category and basis for a required hold rather than keeping every record indefinitely.

An unconverted setup’s 90-day expiry and Forget control do not search for or cancel a separate customer account, subscription, registration or provider resource. Conversely, ending service does not restart the original expiry of a message. Records in earlier account, business-profile, trial, email-log or other storage paths are included in a relevant request if we still hold them; an older storage path is not an exemption from privacy obligations.

You can request access, correction, a usable copy or deletion through hello@helloanswered.com. We can fulfill a verified request manually; we do not promise an account download button. We identify still-held records, coordinate with the business and relevant providers, and explain completion or a permitted exception. Already expired information cannot be recreated for a copy, and already delivered emails or texts cannot be recalled from a recipient’s account. We pass required instructions to processors and assist with records held for a provider’s independent purpose as applicable. Required backup handling must protect retained copies, restrict their use and prevent a restoration from silently undoing a deletion request.

7.7 Separate Website Chat service

A Website Chat service supplied for a business customer is distinct from William on our public website in section 3 and from the phone and texting features above. This policy does not represent that a paid Website Chat feature is enabled. Before such a feature collects information, we will supply its actual processing notice and identify the selected provider, information and conversation context processed, recipients, storage and retention schedule, and request options in the matching service/data and provider/storage schedules. The public William provider and 12-hour context window do not automatically describe a customer's Website Chat service. Those public-site details must not be used as a substitute for the separate feature's notice.

7.8 Outbound and other additional programs

Outbound processing is outside the phone and texting scope described above unless the specific accepted Order and its matching service/data schedule expressly authorize the program. Before its collection or processing begins, the actual program must have a notice describing its purpose, recipient categories, sources of contact information, actions, providers, retention and required permissions. It must also be technically available and separately enabled. An order or schedule alone does not supply another person's consent or replace required notices. No outbound calling, marketing, new message channel, recording program or integration is activated merely because possible data handling is mentioned here.

7.9 Optional Google Calendar and HubSpot connections

These functions apply only where offered and enabled for your account. You choose whether to connect Google Calendar or HubSpot and authorize the provider's requested access. We retain the provider account identifier, granted permissions, encrypted access and refresh tokens, connection status and limited operational records to maintain that connection. Our Google sign-in verification uses the provider's account identifier; this connection does not request access to Gmail or require Google email-profile access.

Google Calendar access lets you select a calendar you can write to and approve booking rules. We read calendar identifiers, names, access permissions and time zones, and availability and event information needed to check conflicts and manage authorized appointments. Enabled booking can create, move and cancel its appointments on the selected calendar. Events created by this integration are private, generically titled Appointment, and contain appointment times and operation references rather than caller names, numbers or conversation notes. The integration adds no attendees or reminders and requests no update emails. This is not a guarantee about every notification Google or another calendar user may produce. Booking results and relevant times can be returned to the voice assistant; existing calendar event titles and descriptions are not supplied to that assistant by this integration.

HubSpot lookup, where enabled, searches contact phone fields and checks the name and number supplied by the caller. It can tell the assistant that those supplied details match an existing contact. That is not identity authentication or permission to reveal private records. This lookup does not return CRM notes, email addresses, financial information or contact history to the assistant and does not keep a local contact cache.

Separately enabling callback-note delivery authorizes future captured callbacks to be matched to a HubSpot contact or, where supported, to create a contact using the caller-provided name and number. The captured callback details and capture time can then be written as a note on that contact. Existing contact fields are not overwritten by this function. Delivery can require attention when a match or provider result is uncertain. Connecting HubSpot alone does not enable delivery, and enabling delivery does not import earlier callbacks. Pausing it and enabling it again begins a new period for future captures.

Our local callback-delivery content keeps the original call's 30-day content deadline described in section 7.2; copying it into a delivery queue does not restart that period. Local booking details expire according to the retention period in the approved booking rule, measured from appointment end and adjusted for a pending or confirmed move. Cancellation does not restart that period. Links from an appointment to captured caller details expire no later than the original call-message deadline and are not extended by moving the appointment. Limited noncontent references, operation state and duplicate-prevention records can remain after content expires. Stored booking rules and connection records support configuration, pending operations and account administration; these are separate from caller-content retention.

You can pause an offered function or disconnect its provider connection in the account controls, and can remove access through the provider's own controls. Disconnecting locally removes the usable tokens and stops new requests under that connection; we also attempt provider revocation. The account may report that provider revocation could not be confirmed. Requests already dispatched can still finish. Disconnecting, canceling service or expiring our local records does not automatically delete contacts, notes or events already created in the provider account. Manage those records with the provider, or contact hello@helloanswered.com for a verified access or deletion request. Provider retention, backups and legal records have their own handling.

Information obtained through Google APIs is handled according to the Google API Services User Data Policy and its Limited Use requirements: https://developers.google.com/terms/api-services-user-data-policy. We use it for the connected scheduling functions you choose and their necessary support and security. We do not sell it, use it for advertising or credit decisions, or use it to train general-purpose AI models. Transfers are limited to providing the disclosed function with your consent, security, legal requirements, or a business transfer with your prior consent. Human access is limited to your affirmative permission for the specific data, necessary security or legal work, or permitted aggregated and anonymized internal operations. These restrictions also apply to derived Google data.

7.10 Optional saved call recordings, transcripts and summaries

Call saving starts off. Where available, the authenticated business account owner chooses whether to save recordings, transcripts and AI summaries of calls handled by the AI receptionist. Summaries require a saved transcript. Before enabling saving, the owner must choose whether the service plays a spoken notice describing the selected saving. Neither notification option is preselected. Choosing no spoken notice requires acknowledgment of the business's responsibility for applicable recording, notice and consent requirements, then allows the owner to continue. That acknowledgment records the owner's instruction; it is not caller consent or a determination that notice is unnecessary.

When recording and notification are both on, the call archive includes the spoken notice recording alongside the conversation recording. No notice-only audio is saved when audio recording is off. Transcripts can contain omissions or errors; summaries use the saved transcript and are labeled as AI-generated. An archive may be incomplete and is not proof that an appointment, payment or other action occurred. Saving ends before a transfer to a human.

A full transcript may be made from the call recording. If transcripts are on and saved audio is off, Twilio temporarily records the AI-handled conversation and OpenAI processes that audio to produce text. We do not place the temporary audio in your owner-accessible archive or make it available for playback or download. Only the selected transcript and summary follow your chosen archive period. Turning saved audio off does not mean that no temporary recording occurs. When spoken notification is on, the notice describes this recording and transcription. Pressing 1 at the opening excludes the conversation from temporary recording and saved transcripts; ordinary answering can continue. A later withdrawal stops further capture and triggers removal of that call’s saved content and temporary audio. Saving stops before a transfer to a human.

Twilio handles selected saved recordings and temporary recordings used for full transcripts on the current phone route; OpenAI handles voice processing, transcription of the selected call audio and selected summaries generated from saved transcripts. Our application stores selected archive content encrypted and linked to its account and call. We also keep the owner's settings, notification choice, required acknowledgment, relevant revisions and limited capture/deletion evidence. Owner instructions and notice-delivery records are kept separate from any actual caller response and are not labeled as explicit caller consent. Provider handling remains subject to applicable terms and settings; our local expiry period does not establish provider-wide deletion or zero retention.

Archive-feature access is limited to the authenticated business account owner, who can download available saved content before it expires or is deleted. The feature does not automatically email archives, publish playback links, share them with a team or export them to a CRM. The owner can delete an archive or turn off new saving directly. When spoken notification is on, the short notice tells callers they can press 1 to continue without saving. A caller can also ask the AI receptionist to stop saving during the call. An opt-out stops further archive capture and removes that call's archive. Ordinary answering and a separately requested callback message can continue. We do not claim a recording has stopped or a provider copy has been deleted when that result is unconfirmed.

When the owner enables saving, the default retention period for new calls is 90 days. The owner can choose 30 days, 60 days, 90 days, 6 months (180 days) or 1 year (365 days). Each archive expires at the end of the period selected when the original call started, or earlier if deleted or withdrawn. A later settings or policy change does not extend an existing archive's expiry. Calls saved under the earlier 30-day edition keep their original deadlines. Retries, summaries and exports do not restart retention. Saving remains subject to the service's capacity and per-call limits; reaching a limit can prevent new content from being saved. Downloaded copies are controlled by the business, which is responsible for their storage, sharing and deletion; HA's later expiry or deletion does not remove those copies.

Temporary transcription processing ends no later than 30 minutes after the call ends; retries do not extend that deadline. When saved audio is off, we hold temporary audio only in memory during an attempt, for at most two minutes and never beyond the original processing deadline. It is discarded when that attempt finishes or is stopped. We request deletion of the Twilio recording when transcription completes or permanently fails, the owner deletes the archive or turns saving off, the caller withdraws, or the processing deadline ends the work, whichever happens first. If provider deletion cannot be confirmed, we show that status and retry within the separate cleanup period described below. We do not represent an unconfirmed provider deletion as completed. Separate provider logs and other independent records remain subject to section 9.

We request deletion of the Twilio recording after verified local import and when deletion, expiry or withdrawal requires it. Only encrypted provider-resource identifiers and limited cleanup state may remain for at most 24 hours after provider deletion first becomes due. That cleanup record contains no transcript, audio, caller name or caller number; it is removed by its deadline even if provider deletion remains unconfirmed. Other call, usage, account and provider records follow their separately described handling.

Contact the business or hello@helloanswered.com about access, deletion or a recording concern. Turning off spoken notification does not remove the business's duties or ours. This policy does not itself enable saving or establish that a particular call may lawfully be saved.

8. Why we use and disclose information

We use the information described above to deliver and secure the requested features, verify access, save and review setup, operate William's optional private-setup interview, respond to inquiries and requests, send requested verification and setup messages, prevent fraud and abuse, troubleshoot failures, comply with applicable law, and establish or defend legal claims. A stated purpose does not authorize a use prohibited by law or a separate agreement.

We limit internal access to people who need information for those tasks. Our hosting, database, AI, email, and telephone providers receive information needed for their roles. Some process on our instructions; some also have independent purposes under the applicable terms, as explained below. We do not claim that every provider is limited to acting solely on our instructions.

We may disclose information to advisers under appropriate duties, in response to a lawful requirement, or as reasonably necessary and legally permitted to protect rights, investigate fraud or abuse, or address a security threat. A sale, financing, reorganization, or other business transfer may include information, subject to applicable law and protections for existing information. A transfer does not itself authorize materially incompatible new uses without required notice and consent.

9. Providers and their separate handling

Where a selected feature uses OpenAI, it processes the messages, conversation audio and text, draft or approved business context, or requested preview script and audio needed for that feature, including the routes described in sections 4, 5 and 7.2. Scripted short-profile creation is separate from an optional OpenAI practice conversation, AI interview or generated sample. Its API documentation says API data is not used to train models by default unless the customer opts in. This is a description of the API default, not a claim that every OpenAI product or account setting is identical. Default abuse-monitoring logs can include prompts and responses for up to 30 days, with longer retention when required by law or reasonably necessary to protect the service or a third party from harm. Endpoint-specific application state and approved retention controls are separate. See https://developers.openai.com/api/docs/guides/your-data and https://openai.com/policies/services-agreement/.

Where a selected feature uses xAI, it processes the audio or text, generated replies and approved context needed for that feature. Its enterprise terms restrict specified training and new-product uses subject to their scope, customer disclosures, and settings. Retention depends on the product, agreement, and controls and can include conversation history and safety, security, compliance, or abuse-investigation records. We do not promise that every hosted-agent record is deleted after 30 days or that Zero Data Retention applies to these demos. See https://x.ai/legal/terms-of-service-enterprise and https://docs.x.ai/developers/faq/security.

If you connect Google Calendar or HubSpot, the selected provider receives the requests and data described in section 7.9 and maintains records under your relationship with that provider. Google Calendar and HubSpot are optional connected destinations; naming them does not mean they are connected for every account or that all their functions are enabled. Our use of Google API data remains subject to the Limited Use commitments in section 7.9.

Twilio handles telephone routing, caller and call metadata, call media, and any configured internal text delivery. For separately enabled customer features, it also handles the selected telephone/SMS transport, delivery and registration information described in section 7. Its data-protection terms allow certain independent processing, including security and abuse-prevention tools and related model training, legal compliance, service improvement, and analytics. This is not a claim that Twilio trains a general-purpose model on every call. Telephone and delivery records can have retention separate from message content and from our local records. See https://www.twilio.com/en-us/legal/data-protection-addendum.

Railway hosts the application, and TiDB Cloud/PingCAP provides application database infrastructure. Resend or the configured SMTP email provider processes recipient addresses, email content, and delivery information for messages we send. Contact email may also pass through domain forwarding and the mailbox provider. These providers may handle operational, security, and delivery records separately from the content stored in our application.

Stripe handles payment information when authorized payment collection is offered. Relevant registration authorities and carriers receive the information needed for an offered registration process. These activities can include independent payment, fraud, legal, identity-verification and carrier purposes. They are not all described as processing solely on our instructions. The customer order’s provider/storage record identifies the selected services and relevant roles, locations and handling.

A provider's product documentation is not a promise that every optional control is enabled for this service. Provider retention is not governed by our local 24-hour, 48-hour, or 90-day cleanup periods. Deletion of an application record, cancellation of a future customer account, or disappearance from a provider console does not necessarily delete provider logs, backups, or legally required records. Contact us if you need help with a request involving information we supplied to a provider.

10. Cookies, measurement, and advertising

We use necessary cookies and browser storage to maintain setup access, preserve interface preferences, and protect the features. You may clear browser storage or restrict cookies through your browser, but doing so can sign you out or prevent you from returning to a saved setup. Clearing browser storage does not itself delete the corresponding server record; use Forget or contact us for that.

We do not sell your personal information. We do not use information about your mobile number or text-message permission for another party's marketing. Optional advertising technology must not be enabled in a way that contradicts these commitments or applicable privacy choices.

The website does not change its behavior in response to the older browser Do Not Track setting. Global Privacy Control is a different signal for privacy opt-outs. With sale and advertising sharing disabled, the current website has no such advertising processing to turn off; we must honor preference signals where applicable law requires them if that processing changes.

We use Umami Cloud to measure visits to public website pages and counts of selected actions, such as starting a browser demo or successfully submitting an inquiry. Measurements include the public page, referring website when available, approved campaign labels, browser language and screen size. Umami processes network and browser information to produce visit statistics and approximate location and device reports. Our analytics payloads exclude form answers, names, email addresses, phone numbers, call audio, transcripts, account identifiers, URL fragments and unapproved query parameters. We do not track private setup, account, dashboard, support or administrator pages with this analytics connection.

This analytics connection does not set tracking cookies or enable session recordings, advertising pixels or cross-site advertising. Umami Cloud hosts the reports in its U.S. region; the selected free plan retains analytics data for six months. See https://umami.is/privacy for Umami's handling of information. Browser restrictions and blockers can prevent measurements, and visitor counts are estimates rather than a complete record of every person. We also review content-free aggregate operational measurements to understand and improve the features.

Meta Pixel and Meta Conversions API remain disabled. We do not use this analytics connection to share personal information for cross-context behavioral advertising or targeted advertising. Enabling advertising or materially different analytics would require an updated, accurate notice and any applicable choices or consent.

11. Security, other records, and location

We use measures designed to protect information, including secure website connections, restricted administrative access, session controls, abuse limits, and encryption for selected stored setup content and generated samples. Not every record is stored in the same form, and these measures do not make every field anonymous or every communication end-to-end encrypted. No internet service or safeguard can guarantee absolute security.

Different records serve different purposes. The specific demo and setup periods above do not apply to every inquiry, support message, security log, delivered email, provider record, or necessary legal record. We keep these records only as long as reasonably needed for the purposes described in this policy, including handling the relationship, protecting the service, resolving issues, and meeting applicable legal and recordkeeping needs. We consider the sensitivity of the information, the ongoing relationship, and applicable deletion rights when deciding what is still needed. We do not claim a uniform seven-year or account-lifetime retention rule.

Information may be processed in the United States and other places where our providers operate. The selected customer service’s provider/storage record supplies its verified locations and relevant backup/deletion arrangements; a U.S. business order is not a promise that all processing occurs only in the United States. Privacy laws and legal-access rules can differ from those where you live. We apply safeguards and transfer requirements applicable to the activity; using the website does not waive mandatory transfer protections. If a security incident requires notice to affected people or authorities, we will provide notice as required by applicable law.

12. Your choices and privacy requests

You can ask to access, correct, or delete personal information we hold about you by emailing hello@helloanswered.com. Include enough information to identify the interaction, such as the email used for setup and whether your request concerns a demo, inquiry, or saved setup. Do not send identification documents or sensitive records unless we explain why a specific item is reasonably necessary. We may verify your identity and authority proportionately to protect someone else's information.

Depending on your location, the information, and which laws apply to the activity, you may also have rights to confirmation of processing, a portable copy, restriction or objection, withdrawal of consent, an authorized representative, appeal of a denial, or a complaint to a privacy regulator. Some laws provide opt-outs for sale, sharing for cross-context behavioral advertising, targeted advertising, or certain consequential profiling. Where those requirements apply, we honor the applicable rights and required preference signals. We do not claim that every regional law applies to every interaction or that a small business is automatically exempt from all privacy law.

For a request about a business customer’s communications, identify the business and approximate interaction date if possible. We may coordinate with that business as the controller and provide assistance under our processing agreement, while handling obligations that apply directly to us. If you cannot identify the business or which provider holds the information, give us the details you have and we will help identify the appropriate route. Do not send a full conversation containing other people’s sensitive information merely to ask a question.

If we cannot fulfill a request in whole or part, we will explain the applicable reason and available appeal or complaint route where required. An appeal can be sent to the same email address with a reference to the original request. We respond within the deadlines required by the applicable law and do not unlawfully discriminate or retaliate for exercising privacy rights. A request may affect a feature that needs the information to function.

You can decline the optional William setup interview and use the manual form. You can avoid voice demos, stop a conversation, or revoke browser microphone permission. These choices stop further participation in that feature; they do not automatically erase information already processed. The Forget control and retention limits are described in section 6.

13. Children and sensitive information

The business setup is intended for adults, and the website and demos are not directed to children under 13. We do not knowingly seek personal information from children under 13. If you believe a child has supplied information, contact us so we can investigate and take the steps required by applicable law, including deletion where required. Do not submit children's information as sample demo content.

Do not enter medical records, financial-account or card details, passwords, government identity documents, or other unnecessary sensitive information into these public features. We do not offer the current demos or setup as a protected-health-information service. If sensitive information is submitted accidentally, contact us for assistance. That restriction does not remove any obligations that applicable law imposes on information actually received.

14. Changes and contact

We will post changes to this policy with a new effective date. For a material change, we will provide additional notice and obtain consent where required. Posting a new policy does not retroactively authorize an incompatible use of previously collected information. We will describe additional collection and processing before introducing a feature that requires a different notice.

hello, answered. is operated by Alexander Killian, an individual operating as hello, answered., at 720 S Gaines St, Portland, OR 97239.

For questions, privacy requests, or an appeal, contact hello@helloanswered.com. English is the original version of this policy, subject to any mandatory legal language requirements.