Start with one voice provider, telephony, a business system of record and the tools needed to complete one client outcome. Production delivery also needs testing, monitoring, access controls, reporting, billing and a human fallback. Add infrastructure when an observed capacity or reliability problem requires it, rather than assembling every available tool first.

Agree ownership before launch. A useful starting point is client ownership of its main business number, with agency-managed orchestration where the service requires it. Record who pays each bill, controls credentials, responds to incidents and transfers assets at offboarding. Check the carrier's transfer rules before promising portability.

The current pricing page separates launch services, managed delivery and additional customization. Read the current package and written scope alongside the required platform subscription. Confirm training, website work, support boundaries and ongoing changes explicitly; do not assume a customer-facing website is included just because an agent build is included.

Reselling voice AI to your own clients and earning commission for referring agencies are different arrangements. VoiceAIWrapper supports white-label client resale. Confirm any referral-program availability, eligibility, attribution, payout and support terms directly with the team before promoting a commission offer.

The workspace guide documents direct client accounts. It does not establish a multi-tier reseller hierarchy in which resellers create their own downstream clients. Request written confirmation of hierarchy depth, permissions, asset ownership, billing and reporting if that is required for your business model.

Use the public terms for the standard service and ask the team about your procurement requirements. Current pricing lists a resolution SLA on Pro, but the exact commitment must come from the applicable agreement. Confirm measurement, exclusions, response and resolution targets, incident communication, data export and termination before relying on a contractual promise.

Public product pages do not provide a complete signup eligibility matrix. Confirm your country, business type, payment method and provider availability with support before planning a client rollout. Do not infer account eligibility from portal language settings or from another agency's successful signup.

Start with one bounded business problem: answering missed calls, handling routine inbound questions, or following up with people who have requested contact. Define the action, the human fallback and how you will check the outcome. Expand after the first workflow works in the client's real process. This is our recommended starting approach, not a promise of sales.

Separate the initial implementation, ongoing management and usage. Estimate the time spent on discovery, integrations, testing, support and changes as well as provider and telephony costs. Describe the included scope and overage rules in the proposal. A small pilot can establish the actual workload before you commit to a fixed ongoing service.

Yes, if ownership and handoff are explicit. The client still needs someone responsible for provider changes, testing, integrations and failures after delivery. Offer a separate maintenance agreement or document what the client will operate. Recurring service should pay for defined work, not conceal costs left out of the initial quote.

Look for a repeated call task, accessible business information, a reachable decision-maker and a result the client can check. Ask for real examples of missed enquiries or scheduling work. Avoid choosing solely because a niche appears in a revenue video. Complex exceptions, emergency decisions and inaccessible booking systems can make a first deployment a poor fit.

Concurrency means simultaneous calls. Set the Pod limit to match the connected provider organization, then allocate that capacity between inbound and outbound traffic. Daily platform fair-use limits are separate. Check the provider, carrier and downstream human handoff capacity before committing to a call-volume target.

Sometimes, depending on the provider, phone-number configuration and routing. Test incoming and outgoing calls, caller ID, simultaneous calls, voicemail and forwarding with the exact number. Use a separate number when the flows need different routing. A shared number does not establish that every provider supports the same configuration.

VoiceAIWrapper supports Vapi, Retell AI, Ultravox, ElevenLabs Agents Platform and Bolna. Growth connects Vapi, Retell AI and Ultravox; Scale and Pro add ElevenLabs and Bolna. Ultravox is outbound-only in VoiceAIWrapper. Choose against the workflow you need and check current pricing for plan access. Feature availability differs by provider.

The subscription covers the white-label platform, client portals, campaign management, reporting, and plan-specific billing and onboarding tools. The connected provider runs the voice or chat agent and bills its own usage directly.

VoiceAIWrapper works with phone-number assets connected through the underlying provider, and its current product pages describe existing VoIP numbers and SIP options. Exact compatibility, number ownership, and regional availability depend on the provider and carrier.

The client portal supports configurable localization. Voice, chat, phone-number, and regional signup availability depend on the selected provider, model, telephony carrier, country, and plan.

Use a dedicated provider key for the VoiceAIWrapper Pod and avoid reusing it elsewhere. The exact permission scope depends on the provider and must allow the assets and actions required by the selected campaign.

Voice activity detection (VAD) identifies speech and silence; semantic turn detection considers whether a thought appears complete. Test false cutoffs, response delay, transcript accuracy and interruption recovery using the same phone audio. Change one setting at a time. The right configuration depends on the provider, model, language and noise conditions.

Identify which system makes the deprecated API request. Assign the change to the owner of that system, whether it is the platform, your automation or another integration. Test response parsing, pagination, webhooks and downstream updates against the replacement. Keep a rollback plan. Confirm any platform migration commitment directly with support.

Connect a number supported by the chosen provider or carrier, then assign it to the intended campaign. Check country inventory, verification, porting, ongoing fees and inbound or outbound compatibility first. Number pools distribute configured numbers; they do not prove that clients can buy local numbers themselves. Confirm any self-service purchasing requirement separately.

Yes. Keep the CRM as the contact and workflow system, and use VoiceAIWrapper for the branded voice AI, client portal and campaign layer. Move contacts and outcomes through configured APIs and webhooks. Test field mapping, calendar tools and webhook delivery in one sub-account before repeating the setup. Pro adds SaaS Creator onboarding automation.

Treat voicemail, no answer and rejected calls as distinct outcomes. Set retry limits and gaps for each supported state, preserve Do Not Call status and prevent duplicate callbacks. Answering-machine detection can be wrong. Check the detected outcome against the call record before changing retry behavior.

Messaging can run through supported provider tools or an external automation. SMS, WhatsApp text, WhatsApp calling and voice messages are separate capabilities. Confirm the channel, country and provider account before promising a workflow, and keep consent and opt-out records in the source system.

Check the calendar record before retrying. The spoken response, tool response and saved appointment can disagree. Inspect the booking payload, time zone, event identifier and the prompt that interprets success. Retrying without checking can create a duplicate appointment. This is a diagnostic method, not a confirmed single cause for every provider.

Yes. Current pricing and campaign documentation describe white-label chat alongside voice, inbound chat campaigns, and embeddable website widgets. Chat and widget access is plan-dependent.

Use a forwarding webhook to send call results into your CRM or automation. Configure the receiving workflow to send the intended notification, handle failures and avoid duplicate sends. Provider-native tools may support additional actions. White-label account emails do not, by themselves, prove that an agent can send a call recap.

Use a stable identifier that is actually present in a captured, tested webhook payload, then map it to the correct client in your receiving workflow. Reject unknown identifiers and duplicate events. The public guide confirms forwarding webhooks, but does not specify every available tenant field. Confirm the payload before building the mapping.

Outbound campaigns support a schedule and dial frequency. The current pricing page lists timezone-aware dialing on Pro. Agencies should store the lead's local time zone, define allowed calling windows, and test daylight-saving transitions where relevant.

Use manual entry for a test or one-off lead, CSV for a prepared batch, and the API for automatic ongoing flow from a CRM or another system. Use the current CSV template and keep consent and Do Not Call status in the source-of-truth system.

Eligible plans include white-label website widgets, and Pro lists headless mode. A custom interface still needs an explicit design for authentication, provider tool calls, state and error handling. Test the supported API flow before selling the integration. Embedding a workflow interface does not automatically make its underlying service white-label or properly licensed.

A campaign can accept new leads manually, by CSV, or through the API, so an ongoing campaign can receive later batches. Use a stable deduplication key, preserve consent, and keep the schedule and retry rules suitable for every batch.

Define a stable field name, populate it before the call, map it to the provider's dynamic-variable or metadata format, and reference the exact same name in the prompt and tools. Confirm the resolved value in a test call and webhook payload.

Keep a session identifier until the visitor provides a reliable contact identifier. Do not assign every anonymous visitor the same dummy phone number or email. If you create temporary records, define how they are merged and removed. Check the current CRM API and test the actual VoiceAIWrapper mapping before using this in production.

Use the provider's calendar or booking tool for availability and appointment creation. Send confirmations and downstream updates through provider-native actions or an authenticated post-call webhook. Confirm provider-specific calendar fields and the exact scope of VoiceAIWrapper's email configuration.

Test a successful call, silence, background noise, interruptions, incorrect details, an unanswered question, a failed tool call and an unavailable human transfer. Then check the calendar or CRM record, client permissions and reporting. Repeat the relevant tests after prompt, model, tool or provider changes. Keep a dated checklist and a rollback owner.

Use approved business facts: services, hours, service areas, pricing boundaries, booking rules and escalation contacts. Give each source an owner and review date. Remove conflicting versions and test questions the material cannot answer. The agent should acknowledge missing information or hand off instead of inventing a price, policy or availability.

Define the destination, business hours and what happens if nobody answers before adding a transfer action. Test that routing does not send the call back into the same agent. Decide what context reaches the human and how the failed transfer is recorded. Confirm provider and phone-system behavior with a real call.

Start with the outcome, duration, summary and any action the client must take. Add recordings, transcripts and captured fields only where permitted and available. Separate raw activity from confirmed bookings or sales. Preview the client role and agree which records need to be retained before launch.

Connect a provider Pod and sync the required agents first. Create the client account, create its workspace, assign the Pod and agents, choose the client permissions, and assign a billing plan if you will collect payments through the portal. Finally, assign the campaign and preview the portal as that client.

For assigned campaigns, the portal can show campaign metrics, leads, call logs, transcripts, recordings, chat logs, and analytics according to the client's permissions. Structured caller fields must be enabled in the provider agent when those fields should appear as lead data.

Set the portal subdomain, add the supplied CNAME at the DNS provider, verify the domain, upload a square logo, add required portal languages, and paste the support-widget embed code. Preview the result with Login as User before handoff.

Client access follows the assets and permissions you assign. The workspace guide documents view and edit access for assigned agents. Before handoff, sign in as the client and inspect campaigns, recordings and billing. Confirm any specific role restriction you require with support; do not assume every setting has an independent permission switch.

The agency can grant agent editing access, and the outbound campaign guide states that clients can be allowed to build campaigns and add their own leads. The right permission set depends on whether the agency sells a managed service or a self-service product.

Standardize one tested configuration, then repeat the client workspace, asset assignment and billing steps. Replace each client's business facts, knowledge sources, credentials and routing. SaaS Creator adds onboarding automation on Pro. Confirm provider-specific cloning and agent-creation support before promising one-click replication of the whole system.

Each client account receives its own workspace, assigned Pod and agents, billing plan, users, and client-facing campaign access. Agencies should assign assets deliberately and preview each account as the client before launch.

Pro lists SaaS Creator for automated intake and creation of a billed client account. Use it after a manual client setup works reliably. Test the resulting workspace, asset allocation, plan and login. Confirm the exact client-facing agent creation, voice selection and workflow controls for your provider before promising them.

Scale and Pro list custom client-portal pages and forms. Access controls depend on the action and plan. Public documentation reviewed for this page does not establish a general single sign-on feature. If single sign-on is required, get the identity provider, role mapping and supported login flow confirmed before purchase.

For inbound campaigns, caller information appears in Leads. Fields such as email and phone can be populated when structured data is enabled in the provider agent. Campaign analytics is a separate performance view and should not be treated as a closed-sale metric without CRM attribution.

Growth includes a branded client portal but does not list custom pages. Scale and Pro list custom client-portal pages, forms, and access controls. Check the live pricing page before quoting because plan contents can change.

A shared agent reduces duplicated configuration but increases the consequences of a wrong client variable or knowledge source. Prefer separate client workspaces and explicit asset assignments. Before sharing an agent, confirm provider support, isolate tools and knowledge, and test cross-client access. Dynamic variables alone do not prove tenant isolation.

Agree the definitions with the client: enquiries captured, qualified appointments, completed transfers, response time and exceptions needing attention. Use the CRM or booking system to confirm the outcome. Calls, minutes and a successful API response describe system activity; they do not establish a sale. Compare against the client's starting process without claiming causation you cannot measure.

Add the platform subscription, extra client accounts, provider usage, telephony, payment processing and your delivery and support costs. VoiceAIWrapper adds no per-minute markup, but the provider bills usage directly. Model low, expected and high volumes, including testing and unsuccessful calls, before choosing a client fee and overage rate.

The trial lasts seven days and requires a card. Your selected paid plan starts on day seven unless you cancel first. Trial access follows the selected plan. Provider and telephony usage remain separate. Use a test workflow to check provider connection, one completed call, the client view and billing before deciding.

VoiceAIWrapper documents three billing paths: Stripe for automatic recurring payments, external billing handled outside the platform, and API-based billing for another processor. Plans can include subscriptions, usage packs, one-time charges, and subscription plus overage.

Check the subscription terms before canceling or changing plans, then export the records you need. Public documentation does not fully define every post-cancellation retention period or a general pause option. Ask support to confirm the effect on your client configuration and data in writing. Inactive-campaign cleanup is a separate policy.

Yes. External Billing lets you handle payment outside VoiceAIWrapper while defining client plans and usage packs. Pro also lists an API-based billing option. A named payment processor is not automatically a native integration. Confirm payment collection, balance updates, failed-payment handling and reconciliation for the processor you intend to use.

Configure tax in the payment system and client billing setup that actually creates the charge. Use a fresh test checkout to inspect what the customer will pay. A billing setting cannot decide your tax obligations. Have a qualified adviser confirm registration, place-of-supply and invoice requirements for your business and customers.

Keep invoice emails and export invoices while billing access is available. If you can no longer retrieve an invoice, contact support from the account email with the billing period needed. Public documentation reviewed here does not establish a guaranteed post-cancellation portal-access period, so avoid relying on continued access.

The billing guide confirms subscriptions, usage packs, one-time charges and subscription plus overage. That does not establish every invoice export or custom accounting field. Before offering the workflow, confirm the required export columns, labels, tax treatment and reconciliation in the payment system and current product configuration.

Inspect a newly created checkout, not only a dashboard toggle. Automatic tax can be enabled in the checkout request, and a merchant-of-record payment mode can control tax separately. Check the payment system and the platform configuration together. Ask support to confirm the session settings if the charge still differs from your intended configuration.

Current pricing includes 5 client accounts on Growth, 10 on Scale and 20 on Pro. Extra accounts cost $15, $12 and $10 per month respectively. Check the live pricing page before quoting. Confirm whether test, inactive or trial client accounts count toward your allowance. Provider usage is separate.

The current pricing policy protects specified subscription pricing and material plan features while the subscription remains continuously active and in good standing. Taxes, provider charges, optional add-ons and separate services are excluded. Check the current policy before canceling or changing plans, because returning or switching can use the then-current pricing.

Healthcare work needs a deployment-specific review of data flows, access, recordings, providers and contracts. A signed VoiceAIWrapper Business Associate Agreement is available to paid Pro customers, not during a trial. Confirm every organization handling protected health information and the agreements required before using real patient data. A platform badge does not approve your complete workflow.

Compare the provider's call record with the campaign record. Check agent and number assignment, synchronization, webhook configuration and whether a final call event arrived. Preserve timestamps and both call identifiers for a private support request. A missing final event is a possible cause, not a diagnosis established by the status alone.

For assigned agents and numbers, the setup guide shows Sync now in campaign settings. After changing provider assets, check the assignment and sync state, then run a test call. Knowledge-base behavior differs by provider. Verify that the intended content is actually returned during the test rather than assuming every asset syncs identically.

A campaign pause should not be treated as proof that provider-side work has ended. Check active calls and queued requests in the provider as well as the wrapper. If an immediate stop is required, use the provider's documented controls or contact support. Queue cancellation behavior must be checked for the specific provider.

Campaigns with no activity are marked after 30 to 45 days, depending on status and whether they hold leads. A 30-day grace period follows. Editing, a new lead, a call or a chat can rescue the campaign; viewing alone does not. Final deletion removes the campaign and its leads, call logs and chat logs. Export needed records before deletion.

Separate an import problem from a provider rejection. Check the number format, required fields, suppression rules and provider call record. Correct the identified input before resubmitting a small test. Status labels and rejection details differ by provider; if the reason is unclear, ask support to trace the rejected attempt.

Pause new dispatches while you compare campaign attempts with provider call and queue records. Check retry rules, answered events and duplicate lead intake. Save call identifiers and timestamps for support. Do not assume every repeat call comes from the same cause, or remove suppression rules to get the queue moving.

Check carrier and agent-provider status separately, then trace the failed call using timestamps and call records. Use a fallback route only if it has already been configured and tested. Document who updates the client. Public documentation reviewed here does not establish automatic cross-provider failover in VoiceAIWrapper.

Inactive-campaign cleanup is documented, but the public material reviewed here does not establish a configurable timer for deleting individual logs from an active campaign. Confirm that requirement with support. Set retention separately for provider recordings, exported data and downstream systems; deleting one copy does not establish deletion everywhere.

First distinguish a missing call record from a missing recording. Confirm the provider call exists, then check the agent, number, campaign assignment and webhook delivery. Check whether the provider produced a recording artifact. Send call identifiers and timestamps privately to support; do not share client recordings in a public community thread.

Use the connected provider's documented active-call controls or support path to stop the call. Then set and test a hard maximum duration and suitable silence timeout in the agent. Vapi and Retell document maximum-duration controls. Confirm the test end reason and review usage; do not rely on the wrapper status alone to prove termination.

A call-duration limit controls one call. A daily per-caller cap is a different requirement, and a native VoiceAIWrapper control is not established by the public documents reviewed here. Discuss a provider or telephony rule with support and test anonymous callers, reset times and exceptions before using it as an abuse control.

A signed VoiceAIWrapper Business Associate Agreement is available to paid Pro customers. It is not included in any free trial. Confirm the agreement's continuing requirements before downgrading or changing the deployment. Do not assume that a previously signed agreement makes a different plan or provider configuration suitable for protected health information.

Check the current feature matrix and the paid plan on your account, then reload and sign in again. If the feature should be included, send support the feature name and upgrade prompt so they can check the entitlement. A previous support exception does not establish a general plan rule.