Deep dive · Reception
An AI front desk is one AI employee that covers the intake side of reception: it reads every inbound enquiry, drafts the answer from your own material, carries the details into a record, and journals the run. A person sends the reply. One AI employee costs $149 a month.
It runs on text today: a form or quiz posts to the lead endpoint and the enquiry is stored before anything else happens. Telephony, SMS and calendar connections are scoped per deployment. Below are the split with a person, five venue scenarios, the channel statuses behind them, and the numbers behind the price.
Unistaff · Front desk
Intake
Reception work breaks into four moves, and the role runs all four in one pass. Each move is a step in code with a validator behind it, which is why the same enquiry handled on a Tuesday and on a Sunday comes back with the same shape.
A web form or quiz posts to the lead endpoint and the enquiry lands in Postgres before anything else runs. The body is capped at 10 KB, a hidden field absorbs bots, and a submission carrying no contact channel comes back with the reason.
The role writes the first reply from your brief, service list and house rules: two to four sentences, 100 to 500 characters, ready to send as written. Any price is marked pending your approval until you confirm it.
Form answers travel with the enquiry as structured fields, so the reply answers what was actually asked. The full name, phone number and email address stay out of the prompt — the role works from an initial and the answers.
Every run is written with its inputs, its outcome and its error text. A run that failed stays visible as a run that failed, so the share of drafts sent unedited is a query rather than an impression.
The loop runs on a timer — every fifteen minutes in the reference deployment — and each pass picks up at most five enquiries that have no draft yet. A burst on a Friday night is worked through in batches, and the drafts are waiting before anyone opens the inbox on Saturday.
Read the person column as the job that remains, and staff for that job rather than for the whole desk. Every row is the platform as it ships today.
| Task | AI employee | Person |
|---|---|---|
| The first reply to a new enquiry | AI employeeWritten on the next run — two to four sentences, in your tone, ending on the next step you set | PersonReads it and sends it |
| Repeat questions | AI employeeAnswered from the brief, the service list and the knowledge items, the same way every time | PersonKeeps that material current |
| Qualification | AI employeeCarries the form answers into the record as structured fields | PersonDecides which enquiries jump the queue |
| A price in a first reply | AI employeeMarked pending your approval; a draft that states a price without the marker is rejected in code | PersonConfirms the number before it goes out |
| Booking | AI employeePrepared with the details the booking needs | PersonPuts it in the calendar — no calendar connector ships today |
| The language of the reply | AI employeeDrafted in the language the enquiry came in — a standing instruction in the role's brief | PersonKeeps the material in both languages where the city is bilingual |
| Anything outside the material | AI employeeStops and says why it stopped | PersonTakes the conversation |
| The record of the week | AI employeeJournals every run with its inputs, its outcome and its error text | PersonReads the journal and corrects the brief |
A front desk in a café and a front desk in a dental practice run the same machinery on different material. Each scenario below is a model of a deployment shape — the enquiries that desk receives, what the role drafts, and the line where it stops. Open one for the enquiry mix, the split table and the model behind its cost.
The desk is a phone on the bar and an inbox nobody owns. Catering and hire enquiries arrive during the morning rush and get answered after close.
Bookings and quote questions arrive while every chair is full. The reply that wins the booking is the one that goes out first.
The desk is with a patient, the phone is ringing and the inbox holds new-patient enquiries that decide next quarter's schedule.
The advisor is under a hood, the phone is on the counter, and every quote request needs six facts before anyone can price it.
One person covers phones, check-in and scheduling. Requests that arrive during a consultation wait for a gap that never comes.
A front desk is only as good as the channel it listens on, so here is where each one stands today. A row moves up when its code does.
| Channel | Status | What that means |
|---|---|---|
| Web form and quiz intake | Live | Enquiry stored on submit, first reply drafted on the next run. |
| Lead notification by email | Live | One attempt with a five-second timeout; the outcome is logged either way. |
| Meta Ads import | On request | Built and tested; turns on once you connect your own ad account. |
| Calendar booking | Scoped per deployment | No calendar connector ships today, so live availability is scoped first. |
| SMS reminders | Scoped per deployment | No SMS provider ships today; the provider is chosen with you. |
| Phone answering (voice) | Custom scope · Enterprise | Telephony, recording consent and the handover rule come before any build. |
| Call tracking | Planned | Calls stay outside attribution until the source exists, rather than being estimated. |
Full connector list and its evidence: /integrations
The ROI calculator on this site ships a Healthcare front desk role, and the calculator’s own defaults answer the cost question without any new arithmetic: one role, 40 hours a week, $35 an hour, 70% of those hours handed across.
Read it as a comparison of hours. The 30% left over is the part of the desk that carries judgement — the upset customer, the exception, the decision — and it still needs a person. What changes is that the person spends the shift on those hours instead of retyping opening times.
Source: Healthcare front desk role at the ROI calculator defaults
Write the ten questions your desk answers every week, with the answer you already give. That list is the first version of the material the role reads on every run.
Set the house rules: what may be stated as fact, what must be marked pending your approval, and the point where the role stops and hands over.
Point the form at the lead endpoint and run in shadow. Read every draft before it goes and correct the material where the draft was wrong.
Count the drafts that went out unedited. That share decides how much of the desk moves across next, and it comes out of the journal rather than out of a feeling.
One AI employee that covers the intake side of reception: it reads every inbound enquiry, drafts the reply from your own brief and service list, carries the form answers into a structured record, and journals the run. A person reads the draft and sends it.
On Unistaff, no. Text intake runs today — a web form or quiz posts to the lead endpoint and the first reply is drafted on the next run. A voice front desk is custom scope on an Enterprise plan and begins with telephony, recording consent and the handover rule. Call tracking sits on the integrations roadmap.
One AI employee is $149 a month on Starter, with a 14-day pilot and cancellation at any time. Up to five AI employees with custom briefs and priority support is $599 a month on Scale. Enterprise is custom-priced and carries custom roles, workflows and SLA-backed support.
The machinery is identical; the material differs. A coffee shop briefs the role on catering lead times and hire rules, a dental practice on visit preparation and what must go to a person. Both read that material on every run, so a Tuesday answer matches a Sunday answer. This page carries five venue scenarios, and four of them open into a page of their own.
Coffee shops, salons and barbershops, dental practices, auto repair shops, and a medical clinic modelled end to end in the case study. Each scenario lists the enquiry mix that desk receives, the split between the AI employee and a person, the point where the role stops, and the cost model run on that desk's hours.
It prepares the booking with the details the booking needs, and a person puts it in the calendar. Unistaff ships no calendar connector today, so a deployment that depends on live availability starts by scoping that connection.
The role works from an initial and the answers on the form. The full name, phone number and email address never enter the prompt, and the contact record stays in the database the reply is sent from.
A chatbot answers inside its window and the conversation ends there. The front-desk role stores the enquiry, drafts a reply your team sends, and journals the run with its outcome, so the week's volume and quality are countable.
Blog
Scenario
Scenario
Solution
Tool
Integrations