Clear answers before discovery.

OnixDesk offers a standardized AI Front Desk, custom AI business employees, connected business automations, lead-generation systems, and Google Ads support. Separately from the voice products, we implement AI into business processes end-to-end and train the team that runs it. The Front Desk has fixed public pricing; every other layer is scoped as a custom plan.

CAD $1,500 setup plus $500 per month, including 500 minutes. Additional usage is CAD $0.40 per minute. This price applies only to the standardized AI Front Desk.

Five genuine service appointments booked in the first 30 days after go-live, or the second month is free. Scope: AI Front Desk only. Qualifying appointment: A real non-test customer service appointment written to the configured calendar during the first 30 days after go-live. Spam, tests, and duplicates do not qualify; a later cancellation or no-show does not erase a genuine completed booking. Eligibility: The business must provide accurate hours, services, prices, and availability; keep service and phone routing active for the full 30 days; maintain a working calendar integration; and provide at least 20 bookable hours per week. Evaluation and claim window: Evaluated once on day 30. Claims must be made within the following 14 days. Exclusions: Setup fees, usage overages, third-party charges, spam, test calls, and duplicate bookings are excluded. A later cancellation or no-show does not erase a genuine completed booking. Sole remedy: Waiver of the second month’s CAD $500 service fee.

They are custom plans. Pricing depends on workflow depth, integrations, permissions, volumes, controls, testing, reporting, and human handoffs. They do not inherit the fixed Front Desk price or guarantee.

Yes. An implementation covers the processes agreed during discovery: Claude Cowork for knowledge work such as research, analysis, and document production; Claude Code for engineering and operational repositories; custom agents built on the Claude Agent SDK and connected to your systems over the Model Context Protocol; and training for the people who will run it. We state which processes suit an agent and which do not before anything is built.

No. OnixDesk is an independent implementation firm. Claude, Claude Code, Claude Cowork, and the Claude Agent SDK are products of Anthropic, and the Model Context Protocol is an open standard created by Anthropic and donated to the Agentic AI Foundation under the Linux Foundation. OnixDesk is not affiliated with, endorsed by, or sponsored by Anthropic, and does not speak on their behalf.

The business does. Plan level and seat count are confirmed during discovery. Custom agents built on the Claude Agent SDK are a second, usage-based cost on top of the plan, because Anthropic requires those agents to run on API credentials rather than a Claude login. OnixDesk does not resell Anthropic plans or API usage, and all Anthropic charges are billed separately from OnixDesk implementation, training, or service fees.

The Anthropic account belongs to the business, so Anthropic’s terms apply to the business directly rather than through OnixDesk. Anthropic states that inputs and outputs from its commercial products — the API and Claude for Work Team and Enterprise plans — are not used to train its models by default, and that the customer is the controller of that data. Consumer plans work the other way: Anthropic’s privacy policy says inputs and outputs may be used for training unless the account holder opts out in settings. That difference is one reason plan level is decided during discovery rather than left to whatever someone already signed up for, and Anthropic’s current terms should be reviewed directly before anything sensitive goes through the system.

No. A reservation is a non-binding discovery request. No payment is collected and no credit card is required at this stage.

No. Ad and media spend, platform charges, and other third-party charges are billed separately from OnixDesk setup, management, automation, or service fees.

No. Insurance, dental, and medical examples are administrative only. They do not provide licensed, clinical, or medical advice, make coverage decisions, or perform patient intake. Regulated questions and decisions go to an authorized human.

No. These are intended capabilities only when the relevant system, permissions, consent rules, fields, failure handling, and human workflow are configured and verified for the implementation.