AI Workforce
AI Workforce is a set of AI agents that do the work, rather than reporting on it afterwards. Each agent owns one job β answering the phone, writing up a treatment note, watching stock levels, contesting a chargeback β and keeps doing that job after hours, at every center, without being asked.
This page explains what an agent is, which agents make up the AI Workforce, what each one is for, and where to find its help.
On this page
- An agent is not a report
- Where the work goes missing
- Serve guests and manage their care
- Run the business
- Document and visualize treatments
- Attract guests and coordinate your team
- The predictive agents
- You stay in control
- What is available for your business
An agent is not a report
A report tells you what happened. A dashboard tells you what is happening. An agent decides what to do about it and then does it.
That difference is what the AI Workforce is built on, and it works because everything already sits in one system:
- Everything is data. Appointments and guest history, every phone call, ratings and reviews, stock movements and payments β all recorded in the same place, as a by-product of running the business.
- Data becomes prediction. Machine-learning models read those patterns and answer forward-looking questions: how many providers next week will need, when a product will run out, which guests are about to lapse.
- Prediction drives a workflow. A prediction on its own is a number on a screen. Zenoti already holds the workflow for the thing that needs doing β booking the slot, raising the purchase order, submitting the dispute.
- An agent runs the workflow. The agent joins the two: it acts on the prediction, follows your existing rules, and records what it did.
Because the agents share the same data, they also hand work to one another. A call AI Receptionist could not complete becomes a text message follow-up. A treatment AI Scribe documented becomes part of the history AI Retention Manager reads. Nothing has to be re-entered.
Where the work goes missing
Most lost revenue is not a performance problem. It is a capacity problem. The same gaps appear in almost every business:
- A call that rings after closing, or during the pre-appointment rush
- Treatment notes written up late at night, from memory
- A consultation where the guest could not picture the result
- A slow Tuesday staffed for a Saturday, and a Saturday staffed for a Tuesday
- A product that runs out halfway through a service
- A guest who quietly stopped coming back, unnoticed
- A card payment disputed, and never contested
- A question that needed answering today, and a report nobody had time to build
None of these need a bigger team. They need something that is always on.
Serve guests and manage their care
Agents that handle calls, clinical documentation, chargeback defense, and guest retention.
AI Receptionist
Answers the phone around the clock and in multiple languages, and books, reschedules, or cancels on the call. It picks up what the front desk cannot: calls after closing, calls during the rush, and second and third calls arriving at once. If a caller hangs up before finishing, it follows up by text message.
It also calls out β appointment confirmations, with identity verification first, then confirming, rescheduling, cancelling, taking a note, or handing over to a person. Every call is logged and recorded.
What makes it usable in a real business is that it works inside your existing rules rather than around them: cancellation policies, deposits, card-on-file requirements, prerequisite and finishing services, provider-specific pricing, mandatory add-ons, and benefit balances. It offers relevant add-ons on the call, and does so on every call.
Use it when
- Calls go to voicemail outside opening hours, or ring out while the desk is with a guest.
- One center answers the phone for several, and the queue builds at peak.
- You want add-ons offered on every call rather than whenever someone remembers.
AI Scribe
For Medspas. Switch it on at the start of a treatment and the provider talks to the patient normally. The agent listens in real time, understands aesthetics terminology, and produces structured, consistent treatment notes, then fills in the chart and the forms that follow. Documentation stops competing with the consultation for the provider's attention.
Use it when
- Providers are finishing charts after hours, or writing them from memory the next day.
- Notes vary in depth and wording between providers, and you need them consistent.
- A provider wants to hold eye contact through a consultation instead of typing.
AI Dispute Manager
Contesting one chargeback by hand means pulling the booking record, finding the signed agreement, downloading receipts and payment history, writing a cover letter to the bank, and submitting all of it before a deadline. Most disputes are lost not on the argument but because that work never got done in time.
The agent alerts you the moment a dispute is raised, assembles the evidence β bookings, receipts, signed agreements, service records β and writes the cover letter in the form banks expect. It submits within the window and tracks every dispute across every center in one dashboard.
Use it when
- Disputes pass their deadline because nobody had time to assemble the evidence.
- Evidence for one dispute sits across bookings, forms, and payment records.
- You run several centers and have no single view of what is open and what was won.
AI Retention Manager
Identifies guests who are drifting away while there is still time to act, rather than reporting the loss afterwards. Each guest gets a risk score with the factors behind it spelled out, so the reason is visible and not a black box. At-risk guests collect into a watchlist your team can work through, alongside a Movement tracker that shows how guests moved between risk bands over the last 30 days.
Use it when
- You want to reach a lapsing guest before they book somewhere else.
- Your team has time for a handful of personal calls a day and needs to know which guests deserve them.
- You want to see whether a change to service, pricing, or staffing moved retention.
Run the business
Agents that handle the decisions behind the front desk.
AI Employee Scheduler
Most schedules are built from availability, coverage, and fairness. Demand rarely gets a say, which is why the same business is overstaffed on a slow Tuesday and short at Saturday peak β a cost already paid before the week begins.
This agent forecasts demand by day and time block, and by service, up to a fortnight ahead. It learns from more than a year of booking history plus the signals no one has time to watch: seasonality, local events, weather. It shows where you are over- and under-staffed before the week starts, then suggests specific shifts for specific employees. A manager reviews and publishes; the forecast improves each cycle.
Use it when
- Overtime and idle time both show up in the same month.
- A manager spends hours each week building a schedule by hand.
AI Inventory Manager
Reads your appointment book and answers three questions: what to order, when to order it, and how much. You tell it how many days of cover you want to hold; it handles the arithmetic.
Signals arrive days before a shelf empties, not after a service has been cancelled. Products are ranked urgent, stable, or slow-moving, and stock that is going to sit until it expires is flagged in advance so capital is not tied up in it. Purchase orders can be raised in bulk by vendor. Every recommendation shows its reasoning and its confidence, and nothing is ordered without your approval.
Use it when
- A service has been cancelled or substituted because a product ran out.
- Capital is tied up in stock that moves slowly or expires on the shelf.
- Ordering depends on one person's memory of what each center gets through.
AI Business Advisor
Ask a question about the business in plain English and get the answer, rather than building the report that contains it. Fast mode returns daily KPIs and trends in seconds. Think mode takes the harder questions β cross-center, cohort, and the "why did this change" kind β and works through them in a couple of minutes. It explains the answer, suggests what to do next, and can pin the result as a dashboard that refreshes on its own.
Use it when
- You need an answer before a meeting and there is no existing report that gives it.
- A number has moved and nobody can say why.
- You compare centers often and want the comparison kept current rather than rebuilt.
Document and visualize treatments
For Medspas.
AI Photo Manager
Guides photo capture with on-screen anatomical overlays so every image is taken at the same angle and distance, which is what makes a before-and-after comparison trustworthy. During a consultation it surfaces relevant transformations from your own gallery, so the evidence you show a guest is your own work.
Use it when
- Your before-and-after photos are inconsistent enough that comparisons are hard to defend.
- Providers need past results to hand during a consultation, filtered to the treatment being discussed.
- You want treatment imagery held against the guest record rather than on someone's phone.
Treatment Visualizer
Part of AI Photo Manager. During a consultation, the provider picks a guest photo, chooses one or more treatment concerns, and shows the guest a simulated view of the likely outcome for dermatology, aesthetics, and lip treatments β from inside the guest's own gallery. The conversation moves from describing a result to showing one. Treatment Visualizer requires beta access; contact your Zenoti representative to request it.
Use it when
- A guest is hesitant because they cannot picture what a treatment will do.
- A consultation covers several concerns and you want to show them addressed together.
- You are discussing a treatment plan over multiple visits rather than a single appointment.
Attract guests and coordinate your team
AI Digital Marketer
Audits everything that decides whether a guest searching nearby finds you: your Google Business Profile, the consistency of your name, address and phone number across listings, website SEO basics, and your review profile. It turns the audit into a score, then into a ranked task list β each task showing the time it takes and the improvement it earns. Every location is audited and scored on its own, re-checked on a schedule, and set against the businesses competing for the same searches.
Use it when
- You run several locations and have no clear picture of which are hard to find online.
- Your listings have drifted out of date after a move, a rename, or a change of phone number.
- You want a ranked list of fixes rather than a general audit to interpret yourself.
ZenChat
An agent surfaces something a person then has to act on, and that handover usually needs a conversation.
ZenChat is real-time messaging between your own employees, on the Zenoti web platform and in the MyZen app. It is not an agent: nothing here acts on its own. It is where the work an agent raises gets picked up β direct messages, group chats, broadcast announcements to a team, and Zenoti records such as a guest profile, appointment, or invoice attached to a message as an interactive card.
Use it when
- An at-risk guest or an urgent reorder needs one named person to own it.
- A manager has an announcement for a shift or a center.
- A question about a guest needs the guest record attached, not described.
The predictive agents
Some of the agents are predictive: they act on what is about to happen rather than what already has. AI Retention Manager flags a guest before they stop booking. AI Employee Scheduler forecasts demand before the week is published. AI Inventory Manager reorders before the shelf is empty.
They are the clearest illustration of why the AI Workforce sits inside Zenoti rather than beside it. A prediction needs history to learn from and a workflow to act through, and both are already here.
You stay in control
An agent working unsupervised is only useful if you can see what it did and why.
- Recommendations show their reasoning and a confidence level, rather than an unexplained answer.
- The decisions that carry cost β publishing a schedule, raising a purchase order β wait for a person to approve them.
- Agents work within the policies you have already configured, not a separate set of rules.
- Every call, message, and action is logged against the guest record.
What is available for your business
The agents are licensed together as an add-on. When your organization subscribes, a starting set is selected automatically to suit your business type and location, and those agents are switched on in the background with their default settings β you do not have to request them one at a time, and there is nothing to configure before they start work.
Which agents make sense depends on the kind of business you run:
- Salons, spas, and fitness β AI Receptionist, AI Dispute Manager, AI Retention Manager, AI Employee Scheduler, AI Inventory Manager, AI Business Advisor, AI Digital Marketer, and ZenChat.
- Medspas β the same set, plus AI Scribe for clinical documentation and AI Photo Manager with Treatment Visualizer for treatment imagery.
From your subscription in the Zenoti Marketplace you can see which agents are on, turn an individual agent on or off, and open one that is running with Launch. An agent that still needs its one-time setup is marked Setup required and cannot be switched on until that setup is finished. Contact your Zenoti account manager if an agent you expect is not listed.
Next steps
- Pick the gap you most want closed, and read the overview for the agent that covers it.
- Check whether that agent is already on. Agents included in your subscription are switched on for you with their default settings, ready to open with Launch; an agent marked Setup required needs its one-time setup finished first.
- Read that agent's getting-started article for anything it needs before it can do useful work.
Once an agent is on, it acts on the next relevant event β the next call, the next treatment, the next reorder point β and records what it did against the guest, the schedule, or the order. There is nothing to trigger by hand, and you can see everything it has done from its own reports.
The predictive agents are the one exception to "straight away". AI Employee Scheduler and AI Inventory Manager learn from your trading history. With little history behind it, AI Inventory Manager shows Insufficient data in place of a prediction, and AI Employee Scheduler shows a provisional forecast and holds scheduling back for that week. The agent is on and collecting; its recommendations start once it has something to learn from.