# Unistaff.ai — full text for language models

> Unistaff.ai is an AI employee platform for small teams. You hire a ready-to-work digital employee for one role — sales, customer support, HR and recruiting, accounting, or operations — brief it in plain language, connect it to your material, and it works a queue around the clock. It drafts; a person sends. Money, outbound messages and any action its rulebook has never seen wait for a human decision.

This file expands https://unistaff.ai/llms.txt with quotable facts for each page. Every statement here is taken from the live page it sits under. Blocks marked "FAQ (verbatim)" are the site's own question-and-answer text, copied without edits. Figures come from three places only: the published price list, the open model behind the ROI calculator, or an external source cited with its year. Where a control or a connector is missing, this file says so.

Site: https://unistaff.ai
Index file: https://unistaff.ai/llms.txt
Sitemap: https://unistaff.ai/sitemap.xml
Compiled: 2026-08-17. Language: English, with one French page.

---

## Canonical facts

### Identity

Unistaff.ai is an AI employee platform — the category is also called digital employees. It is built by Innova Consulting Group Inc., based in Ottawa, Ontario, Canada. Contact: hello@unistaff.ai, +1 613 867 7767, reply within one business day. The product grew out of the marketing operating system the group built to run its own work, and it still shares a codebase with it: a demo request sent from the site lands in the same lead table the group's own operators work from.

### Pricing

One flat monthly price per AI employee. Three plans:

| Plan | Price | AI employees | Support | Included |
|---|---|---|---|---|
| Starter | $149 / month | 1 | Email | 14-day free pilot, cancel anytime |
| Scale | $599 / month | Up to 5 ($120 each) | Priority | Custom briefs, shared workspace |
| Enterprise | Custom | Unlimited | Dedicated success manager | Custom roles and workflows, SLA |

The billing unit is the digital employee. Human teammates in the workspace cost nothing extra, and usage such as messages or tokens is unmetered. Plans can be mixed: one Starter employee alongside Scale employees, billed per employee on the plan each is assigned. Upgrade, downgrade and cancellation are available at any time; annual Enterprise contracts carry SLA terms. The Meta connection is included on every plan. The Starter pilot runs 14 days with full feature access and no credit card.

### Roles

Unistaff.ai staffs five desks: sales, customer support, HR and recruiting, accounting, and operations. Five roles exist as code inside the platform today — market analyst, creative writer, campaign proposer, lead responder, brief composer. A role is a prompt and a validator kept together in code; the validator throws out a result that breaks the role's rules, such as a number without a named source or a checklist item marked green with no reason written next to it. One employee covers one role: each ships with its own scripts, KPIs and memory, and mixing two jobs into one employee makes results impossible to measure.

### What runs without a human, and what waits

Reading, analysing, drafting and journaling run unattended. Money, outbound messages and anything the rulebook has never seen stop and wait for a person. Six gates sit in the code, each labelled with the string it throws:

| Gate | Behaviour | Code string |
|---|---|---|
| Send | The first reply to a lead is a draft until a person sends it; the role has no send path of its own | drafts only |
| Approve | An action in the approval class waits for a human decision; an unrecognised action counts as requiring approval | approval_required |
| Cap | Spending is compared against daily and monthly caps before the money moves | cap_exceeded |
| Halt | The kill switch is checked before every outbound write; engaging it takes a written reason | kill_switch_engaged |
| Decide | A pending approval that runs out of time stays undecided; expiry is never read as consent | expired ≠ approved |
| Fail shut | When the governance database cannot be read, the outbound write is refused | enforcement_unavailable |

The kill switch has three scopes: everything, ads, AI. Objects created through the write path — campaigns, ad sets, ads — come into existence paused. One limit stated plainly on the security page: a spend cap you have not set gives the caps gate nothing to compare against, so the approval gate is what holds the spend.

### Connections and their status

Running today:

| Connection | What it does | Cadence |
|---|---|---|
| Meta Ad Library | Scheduled capture of the public Ad Library for a query and country; advertiser, primary text, call to action and placements parsed into Postgres, raw HTML kept for replay | Daily |
| Web forms and quizzes | A form posts to an HTTPS endpoint and the lead lands in Postgres; body capped at 10 KB before parsing, a hidden field absorbs bots, a submission with no contact channel comes back with the reason | On submit |
| Lead notification email | One notification attempt per saved lead with a five-second timeout, outcome written to the log as a short code; the lead reaches the database first | On submit |

Built and waiting on your credentials: Meta Ads import (Graph API v25.0, ads_read scope — ad account, campaigns, ad sets, ads and daily metrics), Meta Ads changes (separate write token; pause, daily budget and campaign creation, everything created paused), Meta account link (token verified with a live call to Meta, expiry read back from Meta, sealed with AES-256-GCM under a key held in the environment).

On the roadmap with no connector code today: Telegram briefings and approvals, Google Ads, Google Analytics 4, Google Search Console, Google Business Profile, call tracking, n8n.

Everything else — CRM, help desk, ATS, accounting software, calendar, SMS, telephony — is scoped and built as part of a deployment. The integrations page lists every connection with its status, and a row moves up a tier when the code behind it ships.

### Security posture

Running today: TLS 1.2/1.3 with forward-secrecy ciphers and HSTS at a two-year max-age; AES-256 at rest on Neon with 30 days of encrypted backups; platform credentials sealed with AES-256-GCM before storage; a signed operator session cookie that fails shut without its key; bearer secrets on machine triggers; lead records masked outside an operator session to an initial plus the names of the channels offered; logs holding ids and outcome codes rather than submitted values; parameterized SQL; a 10 KB body cap on public form endpoints (4 KB on operator login, 413 on oversize); list page sizes clamped to 200; platform DDoS mitigation from Vercel. The access, masking, encryption and gating paths are pinned by an automated suite: 605 tests, 0 failures on the last run.

Absent today, stated plainly on the security page: a SOC 2 report for Unistaff (no certification of its own, none in progress), SSO and MFA and per-user operator accounts, a customer-facing audit-log export, a signed DPA, a subprocessor list, standard contractual clauses, and self-serve data export and deletion. Separation between customers runs on the deployment boundary — its own database, its own operator key, its own credential key — and the schema holds no tenant, organisation or workspace column, so row-level separation inside one shared database is absent.

### The ROI model and its sources

The ROI calculator prices the human side of a desk and sets it against the published Unistaff price. Formula: weekly hours × number of employees × routine share × loaded hourly rate × 4.33 weeks, minus the platform fee. Defaults: 40 hours a week on a desk, 70% of it routine, $35 an hour loaded (salary, benefits, taxes, tools and management time). At those defaults, 28 hours a week moves across, which models at $4,243 a month against $149 — a difference of $4,094 a month, $49,133 over twelve months, with the modelled human cost 28.5 times the Unistaff fee.

Three worked presets ship with the calculator:

| Preset | Role | Employees | Loaded rate | Routine share | Hours moved | Human cost, model | Unistaff, published | Difference |
|---|---|---|---|---|---|---|---|---|
| Small team, one desk | Sales / SDR | 1 | $35/hr | 70% | 28 h/week | $4,243/mo | $149/mo | $4,094/mo |
| Support queue, two shifts | Customer support | 2 | $30/hr | 75% | 60 h/week | $7,794/mo | $298/mo | $7,496/mo |
| Bookkeeping desk | Accounting assistant | 1 | $40/hr | 60% | 15 h/week | $2,598/mo | $149/mo | $2,449/mo |

External reference points cited on the site, with year and source:

- US private-industry employers paid $46.60 per hour worked in total compensation in March 2026, and $34.78 at the median wage percentile. Benefits take 30.1% of what a private-sector employer pays per hour worked, on top of wages. Paid leave costs $2.24 per hour worked at the median wage percentile. Source: U.S. Bureau of Labor Statistics, Employer Costs for Employee Compensation, March 2026 — https://www.bls.gov/news.release/ecec.nr0.htm
- Replacing one employee costs one-half to two times that employee's annual salary. Source: Gallup, 2019 — https://www.gallup.com/workplace/247391/fixable-problem-costs-businesses-trillion.aspx

The human side of every comparison on the site is a model built from those inputs. The Unistaff side is the published price. Customer results stay off the site until they can be published with permission.

### How a deployment runs

Configuration from a role template takes under an hour and needs no code. The work loop runs on a timer — every fifteen minutes in the reference deployment — and each pass picks up at most five items that have no draft yet. A draft reply runs two to four sentences, 100 to 500 characters, ready to send as written; any price the role was not given carries a pending-approval marker. Every run is written to a journal with its inputs, its outcome and its error text, and a run that failed stays visible as a run that failed.

---

## Pages

### https://unistaff.ai/ — Hire an AI employee, ready to work today
Answers: what is an AI employee, what can it do, how fast does it launch, and what does it cost?
An AI employee ships with a role, behavioural scripts, KPIs and memory. Unistaff.ai employees handle routine sales, support, HR and accounting work, connect to your material, and launch in a day. The home page states the difference from a plain LLM agent in one line: a plain agent starts empty and you write the prompt and wire the tools, while a Unistaff employee starts with a job — a role, a validator that rejects work breaking its rules, and the connections that have actually been built. The example scenario on the page is the ROI calculator's default preset: 28 routine hours a week handed off, $35 an hour loaded human cost, $4.1k net monthly saving, $149 starting price per employee.
FAQ (verbatim)
Q: What is an AI employee?
A: An AI employee is a digital employee that comes with a role, behavioral scripts, KPIs and memory. It handles routine work in a specific job, connects to your tools, and keeps history so it knows your business.

### https://unistaff.ai/product — The AI employee platform: pick a role, brief it, keep the last word
Answers: how does an AI employee work, what does the platform do without asking, and which connections exist today?
Four steps run a deployment: brief (describe the job, tone, rules and outcome in plain language), configure (pick a role, point it at the source of work, set which actions need a human yes and what the spend cap is), run (the role runs on a schedule and its output lands as a draft with the reasoning attached), review (every run is written to a journal with inputs, outcome and error text). Three layers sit under that: a general frontier model does the reasoning and answers into a single JSON object, so prose that fails to parse never leaves the runner; the role is a prompt plus a validator kept in code; your brand pack — services, positioning, prices, tone — is read into the prompt on every run, and a missing input is named in the output rather than filled with an invention. Five roles ship as code: market analyst, creative writer, campaign proposer, lead responder, brief composer. The page prints the six oversight gates with the exact string each one throws.
FAQ (verbatim)
Q: Can a non-technical owner set this up?
A: Yes. You pick a role and write the instructions in plain language — the job, the tone, the rules, what a good outcome looks like. That brief is what every run reads from.

### https://unistaff.ai/use-cases — AI employees, by role
Answers: which AI employee role should a small business hire first, and what does each role cover?
Ten role cards grouped into revenue, operations, support and custom: AI sales assistant (meetings booked), AI e-commerce agent (cart recovery), AI real estate assistant (lead response time), AI operations assistant (tasks closed), AI accounting assistant (invoices out), AI HR and recruiting agent (candidates screened), AI customer-support agent (tickets deflected), AI healthcare front desk (appointments booked), plus custom roles. The page's own advice: start where the backlog costs revenue now — usually inbound sales or support — run one digital employee against one KPI for one week, then decide whether to add the next role. Keep one employee per role.
FAQ (verbatim)
Q: Which role should a small business start with?
A: Start where the backlog costs you revenue now — usually inbound sales or support. One digital employee, one KPI, one week of measurement, then decide whether to add the next role.

### https://unistaff.ai/pricing — Pricing: $149, $599, custom
Answers: what does an AI employee cost, what is the billing unit, and what is in each plan?
An AI employee costs $149 a month on Starter, $599 a month for up to five on Scale ($120 each at that size), and custom pricing on Enterprise. Every plan carries the 14-day pilot and the same Meta connection. Billing runs per digital employee — per seat, per token and per message billing are all absent, and human teammates in the workspace are free. Plan comparison: Starter gives 1 AI employee and email support; Scale gives up to 5 AI employees, priority support, custom briefs and a shared workspace; Enterprise gives unlimited employees, custom roles and workflows, SLA-backed support and a dedicated success manager. Every plan runs on the same platform: TLS 1.2/1.3 in transit, AES-256 at rest on Neon, and a human approval ahead of any outbound spend.
FAQ (verbatim)
Q: What does an AI employee cost?
A: $149 a month for one AI employee on Starter, $599 a month for up to 5 on Scale — $120 per employee at that size — and custom pricing on Enterprise. The fee covers the role itself: its brief, memory, workflows, analytics and logs. Human teammates in the workspace are free. To weigh that against a human hire, the ROI calculator models the loaded cost of the same desk and shows the monthly difference.

### https://unistaff.ai/integrations — Every connection, with its real status
Answers: what does Unistaff.ai connect to today, what is waiting on credentials, and what is planned?
Meta is the platform this system talks to today. The connection ledger runs in three tiers under one rule: a claim earns its place after the code that performs it exists in the repository. Three entries run today (Meta Ad Library capture, web form and quiz intake, lead notification email). Three are built and waiting on your credentials (Meta Ads import over Graph API v25.0 with ads_read, Meta Ads changes on a separate write token, Meta account link sealed with AES-256-GCM). Seven sit on the roadmap with no connector code (Telegram, Google Ads, Google Analytics 4, Google Search Console, Google Business Profile, call tracking, n8n). Five rules apply to every source: a missing key closes the path, every attempt is journaled, unconnected reads as unconnected, changes wait for a person, and one switch halts outbound writes.
FAQ (verbatim)
Q: How many systems can Unistaff connect to today?
A: Meta. The Ad Library capture runs daily, and the ads import and change paths are built and waiting on credentials. Web form intake and lead notifications run on their own path. Everything else listed on this page carries a roadmap tag.

### https://unistaff.ai/security — Every gate here fails shut
Answers: how is data protected, what stops an AI employee from spending money, and which controls do not exist yet?
Twenty-four controls in four groups: access (5), customer data (7), public surface (4), autonomy limits (8). Five fail-shut conditions are printed on the page: operator key unset means every session cookie is rejected; trigger secret unset means scheduled jobs answer unauthorized; an action missing from the matrix requires human approval; an expired approval is refused; an unreachable governance database stops the outbound write. Live posture as published: TLS 1.2/1.3, HSTS max-age 63072000, AES-256 at rest on Neon, operator API answering 401 without a session, a 10 KB public body cap, and closed as the default. The automated suite behind the access, masking, encryption and gating paths is 605 tests with 0 failures on the last run. The page also prints, in full, the list of controls this platform lacks — SOC 2, SSO and MFA, audit-log export, DPA, subprocessor list, standard contractual clauses, self-serve export and deletion — so a procurement checklist can be answered in one pass.
FAQ (verbatim)
Q: Who can read the data submitted through the site?
A: A full lead record has two doors: an operator session or a bearer secret. The API verifies one of them itself before it touches the database, independent of the page gate. Every other caller receives a masked record — an initial and the names of the contact channels, with the name, phone, email and free-text answers left in storage.

### https://unistaff.ai/roi-calculator — ROI calculator: AI employee vs. human hire
Answers: what does an AI employee save against the human cost of the same desk, on my own numbers?
The calculator prices the human side of a desk and sets it against the published Unistaff price. At the defaults — 40 hours a week, 70% routine, $35 an hour loaded — 28 hours a week move across, modelling at $4,243 a month against $149, a difference of $4,094 a month and $49,133 over twelve months. Six role presets are selectable: sales / SDR, customer support, HR / recruiting, accounting assistant, operations, healthcare front desk. Three worked scenarios ship as starting points (see the ROI model table above). The page states its boundary in its own words: the human side is a model whose sources sit on the cost comparison page, the Unistaff side is the published price, and customer results stay off the page until they can be published with permission.
FAQ (verbatim)
Q: What does an AI employee cost?
A: $149 a month for one AI employee on Starter, $599 a month for up to 5 on Scale — $120 per employee at that size — and custom pricing on Enterprise. Billing runs per digital employee: human teammates in the workspace are free, and messages and tokens stay unmetered.

### https://unistaff.ai/solutions/sales — AI sales assistant
Answers: how does an AI employee qualify inbound leads and write the first reply?
Every form fill lands as one record and is scored against the three questions that decide fit. The first reply is written in your approved language before a rep opens the lead, and it waits for the rep to send it. The rep receives the draft, the score and the reason in one place. A price or a promise the assistant was never given comes back marked as pending you, with the gap named — a hard rule in the role, and a draft that states a number without the marker is rejected before anyone sees it. Published counters on the page: 0 messages leave without a person on the send, inbound form leads are accepted 24/7 including weekends, and the pilot target ranges modelled are +20–40% meetings booked and 15–20 SDR hours handed off weekly. Out of the box the platform ships one connector, Meta ads, plus a lead endpoint any form can post to; a CRM connection is scoped and built as part of a deployment.
FAQ (verbatim)
Q: Does it replace my sales team?
A: No. It handles the top of the funnel so human reps spend more time on calls and negotiations.

### https://unistaff.ai/solutions/support — AI customer support agent
Answers: how does an AI employee draft answers to repeat tickets and escalate the rest?
The routine majority of tickets arrives answered, sourced and ready to send. Each draft carries the passage of your material it came from, so an agent can verify it in seconds. Anything outside the material it was given goes to a person with the thread and the reason attached, and the draft says why it stopped. A ticket arriving at 3 a.m. is read and drafted on the same schedule as one at noon. Published counters: 0 answers reach a customer without a person on the send, tickets are accepted 24/7 through your site form, and the pilot targets modelled are 60–70% of routine tickets answered from a draft with CSAT held. Week one runs in shadow mode so the draft accuracy is measured before anything depends on it. One platform connector ships today, Meta ads; Zendesk, Intercom, Freshdesk and the rest are built against your account during a deployment.
FAQ (verbatim)
Q: Will customers know they are talking to AI?
A: Today a person reads and sends every answer, so what reaches the customer is your team writing with a head start. When a role starts replying directly, disclosure goes in the reply template and we set that up with you.

### https://unistaff.ai/solutions/hr — AI HR and recruiting agent
Answers: how does an AI employee screen applicants and draft candidate messages?
Applicants are scored against the must-have and nice-to-have criteria in your job description, with the reason written next to every score. Invites, reschedules and rejections are written and wait for a send. Onboarding turns into a day-by-day list with the owner named on each item, and collected reference answers become a structured summary with the quotes kept alongside. A person picks who moves forward; the machine never makes that call. Writing the criteria down is the setup work and usually takes under an hour. No ATS connector ships today — applications arrive through a form that posts to the lead endpoint, and a Greenhouse, Lever or Workable connection is scoped during a deployment.
FAQ (verbatim)
Q: Does it make hiring decisions?
A: No. It screens, schedules, and summarizes. Final decisions stay with humans.

### https://unistaff.ai/solutions/accounting — AI accounting assistant
Answers: what does an AI employee do with receipts, invoices and reconciliation, and what does it never touch?
It keeps the list of open invoices and writes each reminder in the tone you set, reads a forwarded receipt and proposes the category with the merchant, amount and date it read, matches transactions against invoices and puts the ones that refuse to match in front of a person, and holds payroll dates, missing timesheets and filing deadlines on a list. The platform has no payment path at all, so no configuration of it can move money. Stored credentials are encrypted before they reach the database, and every outbound message waits for a person. The ROI calculator models 25 routine hours a week for this role. Decisions handed back to a human: capitalize or expense, recognition and timing, cross-border tax treatment, whether an anomaly is fraud or an error, whether to send a reminder at all, and the signature on a filing. No accounting-software connector ships today; a QuickBooks or Xero connection is scoped and built as part of a deployment.
FAQ (verbatim)
Q: What does one accounting employee cost?
A: $149 a month on Starter, $599 on Scale, custom on Enterprise, charged per employee per month. The ROI calculator models 25 routine hours a week for this role; you set the share you expect to hand over and the loaded hourly cost of the person doing it today.

### https://unistaff.ai/solutions/operations — AI operations assistant
Answers: how does an AI employee dispatch work, write status updates and report the day?
It proposes the assignment from your own rules — location, skill, capacity — and names the rule it applied. It writes the update for each job the moment its state changes, keeps the list of open purchase orders and missing items with the next chase written for each, and rolls completions, stalls and open loops into one morning summary with the age of every stuck item. Sending stays with your coordinator. An urgent after-hours job is marked urgent and put at the top of the queue with the escalation message already written; waking a person up remains a person's decision. No field-service connector ships today — Jobber, Housecall Pro, ServiceTitan and the rest are scoped and built against your account during a deployment — and the platform carries no SMS provider and no voice connector.
FAQ (verbatim)
Q: Does it integrate with field-service tools?
A: Not out of the box today. Jobber, Housecall Pro, ServiceTitan and the rest are connectors we scope and build against your account as part of a deployment.

### https://unistaff.ai/vs/human-hire — AI employee vs. hiring a human: what each one costs
Answers: what does an AI employee cost against a human hire, line by line, and when does each one win?
A Unistaff AI employee costs $149 a month on Starter and $599 for up to five on Scale. A full-time human on the same desk models at $6,062 a month at a loaded $35 an hour, of which $4,243 sits in the routine 70% an AI employee takes over — a modelled difference of $4,094 a month. The cost table runs in two halves: everything in the AI column is the published price, everything in the human column is a model built from public statistics with its source under each line (BLS March 2026 for benefits at 30.1% and paid leave at $2.24 per hour worked; Gallup 2019 for replacement cost at one-half to two times annual salary). Beyond the invoice, the page compares time to hire, cost floor, management load, availability, judgment and creativity, and scaling. Hire AI first when the work is repetitive, rules-based and high-volume, when coverage is needed outside business hours, when cost must be predictable, and when audit logs matter. Hire a human when the role demands empathy, trust or complex negotiation, when the work changes constantly, when brand voice needs taste, and when leadership or coaching is the job.
FAQ (verbatim)
Q: What does an AI employee cost?
A: $149 a month for one AI employee on Starter, $599 a month for up to 5 on Scale — $120 per employee at that size — and custom pricing for Enterprise. The price covers the role itself: its brief, memory, workflows and logs. Human teammates in the workspace cost nothing extra, and usage stays unmetered.

### https://unistaff.ai/vs/chatbot — AI employee vs. chatbot vs. assistant
Answers: what separates the three categories, mechanism by mechanism?
A chatbot answers the message in front of it. An assistant does what you ask while you stay in the thread. An AI employee holds a role: it keeps your facts between runs, works a queue on a schedule, obeys written limits on what it may do alone, and leaves money and outbound messages to a person. Each of those parts is a record you can open — a role is a prompt with a validator kept in code, your facts are knowledge items read into the prompt on every run, a timer wakes the work loop every 15 minutes, and each run lands in a journal with its inputs, its outcome and its error text. Memory works two ways: knowledge items stored per brand, and the output of one successful run becoming the input of the next role in the chain. Six axes are compared on the page: memory between conversations, initiative, queue, limits of authority, reporting, and human control.
FAQ (verbatim)
Q: What is the difference between a conversational AI chatbot and an AI assistant?
A: A chatbot sits on a channel and answers whoever writes to it, using the material it was given. An assistant sits in a thread you drive and does what you ask of it there. Both hold context for the length of the conversation. Neither one carries a role, a work queue, or a rule about what it may do without you.

### https://unistaff.ai/vs/outsource — AI employee vs. outsourcing
Answers: how do control, start speed, monthly cost and quality consistency compare with an outsourced team?
Outsourcing gives you hands; an AI employee gives you a playbook that runs on its own. The comparison covers control (you set rules, thresholds and approvals rather than briefing and waiting for revisions), speed to start (hours against weeks of procurement, contracts and onboarding), monthly cost (one flat fee per AI employee against hourly rates, project fees, markups and management overhead), quality consistency (the same process, tone and audit log every time against output that depends on the assigned person), scaling and security. The page's own recommendation is a split: the AI employee absorbs the repeatable rules-based work while outsourced partners keep strategy, creative and specialist tasks.
FAQ (verbatim)
Q: Will an AI employee replace my outsourced team entirely?
A: Plan a split: the AI employee absorbs the repeatable, rules-based work, while your outsourced partners keep the strategy, creative, and specialist tasks only they can do.

### https://unistaff.ai/vs/virtual-assistant — AI employee vs. human virtual assistant
Answers: how do cost, coverage and consistency compare, and when does a human VA still win?
A digital employee starts at $149 a month per role, with the $149–$599 flat band covering Starter and Scale and custom pricing above that. A full-time human VA usually runs well over $1,500 a month after salary, tools, taxes and onboarding. Availability, speed, consistency and scaling are compared on the page. Its recommendation is to run both: give the AI employee the repeatable share of the role and keep the human VA for the judgment-heavy part. Choose a human VA when the work is undefined, highly personal, or changes every day; choose an AI employee when the work is repeatable and tied to a clear KPI.
FAQ (verbatim)
Q: Will an AI employee replace my virtual assistant?
A: Run both. Give the AI employee the repeatable share of the role and keep the human VA for the judgment-heavy part — costs drop and the person stays.

### https://unistaff.ai/resources/compare — Unistaff.ai vs. chatbot builders vs. agent frameworks
Answers: how does Unistaff.ai compare with chatbot builders and agent frameworks, feature by feature?
A feature grid across three columns — Unistaff.ai, a chatbot builder, an AI-agent framework — covering pre-built roles, KPIs out of the box, business memory, no-code setup, human approvals and monthly cost.
---
## Explainers and deep dives

### https://unistaff.ai/resources/faq — AI employee FAQ
Answers: everything in one place — what an AI employee is, how a launch runs, what it costs, where the data lives, which connectors exist, and how the KPIs are counted.
Thirty-six answers in six groups, each one carried over from the page that owns the fact: pricing, security, product, integrations, and the role pages. The hub's own summary: an AI employee is a digital worker with one role, a written brief, memory and a journal of every run; Unistaff.ai staffs five desks — sales, support, HR, accounting and operations; billing runs per employee at $149 a month for one, $599 for up to five, custom above that. Where a control or a connector is missing, the answer says so and names what exists instead.
FAQ (verbatim)
Q: What is an AI employee?
A: An AI employee is a digital employee: a role with its own memory, workflows and analytics operating in one or more channels. It arrives with the job written down — the brief, the tone, the rules, what a good outcome looks like — and every run it makes is written to a journal with its outcome.

### https://unistaff.ai/resources/ai-front-desk — AI front desk for small business
Answers: what does an AI front desk handle for a café, a salon or a clinic, which channels are live, and what does it cost?
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 — while telephony, SMS and calendar connections are scoped per deployment. The four moves in one pass: intake, draft, qualify, journal. The loop runs every fifteen minutes in the reference deployment and each pass picks up at most five enquiries with no draft yet. Draft length is two to four sentences, 100 to 500 characters; any price is marked pending your approval. The full name, phone number and email address stay out of the prompt — the role works from an initial and the form answers. Three venue scopes are modelled: coffee shop, salon or barber, clinic or practice.
Channel status as published on the page:
| Channel | Status | What that means |
|---|---|---|

### https://unistaff.ai/resources/ai-accounting-assistant — AI accounting assistant: what it does, and how to choose one
Answers: what does an AI accounting assistant do, which decisions stay human, and which questions should you ask a vendor?
An AI accounting assistant reads your financial documents and drafts the bookkeeping around them: it pulls merchant, amount, date and tax off a receipt, proposes a category, matches payments against open invoices, and writes the chasing email. Everything it produces is a draft; a person reviews it, approves it, and signs the filing. That property — draft in, human out — is what separates the software from neighbouring categories: a tool that joins your calls and transcribes them is a meeting assistant, and a tool that answers questions about numbers already in your ledger is a reporting layer. The honest unit of value is entries a human accepted. The page splits the work into the routine it takes (field extraction, category proposal, transaction matching and more) and the decisions it hands back, distinguishes an assistant for bookkeepers (records) from one for accountants (a review packet arranged for sign-off), and lists eight questions to ask any vendor.
FAQ (verbatim)
Q: What is an AI accounting assistant?
A: Software that reads financial documents and drafts the bookkeeping around them. It extracts merchant, amount, date and tax from a receipt, proposes a category, matches payments against open invoices, and writes the reminder for an overdue one. The output is a draft; a person reviews it, approves it and signs the filing.

### https://unistaff.ai/resources/ai-hr-assistant — AI HR assistant: what it does and what it costs
Answers: what does an AI HR assistant do across screening, scheduling and onboarding, and what does it cost?
It handles the repetitive parts of recruiting and onboarding: resume screening against must-have and nice-to-have criteria, interview scheduling, reference collection and new-hire checklists. Human HR keeps final decisions, offer strategy and employee relations. Pricing runs per digital employee: $149 a month for one, $599 for up to five, custom above that, with a 14-day free pilot on Starter and human seats at no cost. Configuring the criteria takes under an hour. One platform connector ships today, Meta ads; applications reach the role through a form that posts to the lead endpoint, and Greenhouse, Lever, Workable and the rest are built against your account during a deployment.
FAQ (verbatim)
Q: What does an AI HR assistant actually do?
A: It handles the repetitive parts of recruiting and onboarding: resume screening, interview scheduling, reference collection, and new-hire checklists. Human HR keeps final decisions, offer strategy, and employee relations.

### https://unistaff.ai/resources/ai-sales-assistant — What an AI sales assistant does
Answers: what does an AI sales assistant do at the top of the funnel, and where does a human rep pick up?
An AI sales assistant is an AI employee that handles the top of the funnel — inbound lead response, qualification, meeting preparation, record hygiene and follow-up — so human reps spend more time closing. It asks the three questions that determine fit, scores leads against your ICP, and escalates edge cases with the full context: transcript, lead score and recommended next step. Human reps keep discovery calls, negotiations and strategic account conversations. One platform connector ships today, Meta ads; leads arrive through a form or quiz that posts to the lead endpoint, and CRM connections are scoped per deployment.
FAQ (verbatim)
Q: What is an AI sales assistant?
A: An AI sales assistant is an AI employee that handles the top of your sales funnel — inbound lead response, qualification, meeting booking, CRM updates, and follow-up — so human reps spend more time closing.

### https://unistaff.ai/resources/ai-support-assistant — What an AI support assistant does
Answers: what does an AI support assistant resolve, how does escalation work, and how is the result measured?
It answers repeat questions from your help docs, knowledge base and past tickets, detects sensitive or complex cases and routes them to the right agent with transcript, history and a suggested next step, and keeps the queue tagged, prioritized and accurate. Measurement runs on first-contact resolution, average handle time, customer satisfaction, and the share of tickets where the draft went out unedited — a count you can pull, because every run is journaled with its outcome. One platform connector ships today, Meta ads; tickets arrive through a form that posts to the lead endpoint, and helpdesk connections are scoped per deployment.
FAQ (verbatim)
Q: Will customers realize they are talking to AI?
A: Disclosure is built in by default. More importantly, the assistant replies in your brand voice, uses your help docs, and stays inside approved policy. The speed is what stands out first.

### https://unistaff.ai/resources/ai-employee-vs-chatbot — AI employee vs chatbot
Answers: what is the real difference between an AI employee and a chatbot?
The explainer version of the comparison: scope, memory, setup, oversight and cost set side by side, with the mechanism named behind each row.
FAQ (verbatim)
Q: Can I run a chatbot and an AI employee at the same time?
A: Yes, and it is a common split. The widget answers on the page in the same second, and the work that continues after that answer goes to the role. Today the handoff is a form post to our lead endpoint, which lands the request as one record and emails you when it arrives.

### https://unistaff.ai/resources/ai-vs-human-hire — AI vs human hire, in depth
Answers: when should a small team hire an AI employee, and when is a person the right call?
The long-form companion to the cost comparison: which work is repeatable enough to hand over, which stays with a person, and how the two run on one team.
---
## Industries
Each industry page describes a deployment shape — the desks staffed and the roles scoped — for that kind of business. Connector status for all of them is the one published on the integrations page.

### https://unistaff.ai/industries/saas — AI employees for SaaS
Answers: which desks does a SaaS team staff with AI employees?
Trial onboarding, support coverage and renewal follow-up scale with signups while headcount stays flat. Roles scoped: SDR, customer success, support. Measurement runs on trial-to-paid rate, time-to-value, support ticket deflection and expansion-signal conversion.
FAQ (verbatim)
Q: Can it read product usage data?
A: Not out of the box. One platform connector ships today, and it is Meta ads. A Segment, Mixpanel or product-database connection is scoped and built as part of a deployment, and it appears on the integrations page when the code behind it ships.

### https://unistaff.ai/industries/agency — AI employees for agencies
Answers: how does an agency serve more clients with the same team, and how does white-label work?
Client intake, project updates and delivery capacity run under your brand. Roles scoped: intake, project management, reporting. The partner program lets an agency deploy white-label AI employees with its own branding and pricing, set the retail price and keep the margin, with partners running month to month.
FAQ (verbatim)
Q: Can I resell this to my clients?
A: Yes. Our partner program lets you deploy white-label AI employees with your branding and pricing.

### https://unistaff.ai/industries/ecommerce — AI employees for e-commerce
Answers: which store questions does an AI employee draft answers to?
Order-status questions, returns and exchanges, cart follow-up and 24/7 coverage. Roles scoped: support, retention. Refunds are initiated with human approval or drafted for a person to confirm.
FAQ (verbatim)
Q: Which e-commerce platforms connect?
A: None out of the box today. One platform connector ships, and it is Meta ads; store questions reach the role through forms and exports. Shopify, WooCommerce, BigCommerce and Magento are connectors we scope and build as part of a deployment.

### https://unistaff.ai/industries/realestate — AI employees for real estate
Answers: how fast does a listing enquiry get a reply, and how are showings handled?
Instant lead response, showing coordination and follow-up that keeps every listing inquiry warm. Roles scoped: SDR, coordinator. Routing rules can be round-robin, territory-based or team-based. Legal and financial questions route to a licensed agent.
FAQ (verbatim)
Q: Does it integrate with MLS?
A: Not today. No MLS or RETS connector ships, so the role works from the listing material you give it — your feed export, your listing documents, your brief. An MLS connection is scoped at the start of a deployment.

### https://unistaff.ai/industries/healthcare — AI employees for healthcare
Answers: what does an AI front desk cover in a clinic, and what stays with the practice?
Booking, reschedule and cancellation requests drafted against the rules you set; intake details carried off your form into one structured record; reminder and pre-visit instruction drafts; inbound requests sorted with the urgent ones marked for the staff line you define. Roles scoped: reception, intake. Clinical records and clinical decisions stay with your team and your systems, and the role gives no medical advice. Health-data obligations are scoped before launch rather than assumed.
FAQ (verbatim)
Q: What does it handle, exactly?
A: The administrative side of the front desk: booking, rescheduling, reminders, intake paperwork, and routing messages to your staff. Clinical records and clinical decisions stay with your team and your systems.

### https://unistaff.ai/resources/guides — AI employee guides
Answers: which playbooks exist for deploying an AI employee?
Four step-by-step guides plus two cross-links: choosing a first role, writing the brief, measuring ROI in week one, rolling out to a team, the security checklist (on the security page), and the connector setup playbook (on the integrations page).

### https://unistaff.ai/resources/guides/how-to-choose-your-first-ai-employee
Answers: which desk should you staff first, and how do you pick the KPI?
Pick the desk that is losing money today, define one KPI, and launch in a day.

### https://unistaff.ai/resources/guides/writing-the-perfect-brief
Answers: how do you write instructions that keep an AI employee on-brand and on-policy?
Plain-language instructions, tightened by reading real conversations every day for the first week and turning every mistake into a brief update.

### https://unistaff.ai/resources/guides/measuring-roi-in-week-one
Answers: what do you track in the first seven days?
What to track, what to ignore, and how to decide whether to add another role after the first seven days.

### https://unistaff.ai/resources/guides/rolling-out-to-a-team
Answers: how do you introduce an AI employee so the team uses it?
How to introduce an AI employee so your team treats it as a colleague.
---
## Modeled scenarios

### https://unistaff.ai/resources/case-studies — What a deployment looks like
Answers: what does a deployment look like, and how is every number calculated?
Three deployment shapes, each built from the inputs of the ROI calculator, with the arithmetic printed on the page. These are models: no client, no logo, no measured result. The method has five parts — four inputs (role, weekly hours in the routine loop, share handed to the machine, loaded hourly cost), the arithmetic (hours × rate × 4.33), targets agreed at scoping and measured in your own systems, the connections each scenario assumes, and the boundary between model and measurement.

### https://unistaff.ai/resources/case-studies/b2b-saas-ai-sdr — A 22-person B2B software company
Answers: what does an inbound-SDR deployment target, and what arithmetic produces those targets?
Starting point: demo requests arrive around the clock and one SDR answers during business hours, so an overnight lead waits four hours. Deployment: the website form posts to the lead endpoint, the AI SDR is briefed on the ICP, the pricing page and ten sample objection answers, and it writes the first reply as a draft. Model inputs: 40 hours a week, one employee, 45% handed off, $35 an hour loaded — 18 hours a week, $2,728 a month of routine time against $149 a month. Targets a pilot of this shape runs against: +34% meetings booked, time to first reply from four hours to minutes, 87% of first replies starting from a draft, 18 human SDR hours freed per week. Targets, not measured results.

### https://unistaff.ai/resources/case-studies/agency-ai-support-ops — A 14-person marketing agency
Answers: what does an agency hand to two AI employees?
Two AI employees draft the client-status and project-ops answers, with a target of 28 hours a week handed off and account managers checking the escalation queue twice a day. The page prints its inputs and the multiplication behind that figure.

### https://unistaff.ai/resources/case-studies/clinic-ai-front-desk — An 8-person outpatient clinic
Answers: what does a front-desk pilot in a clinic aim at?
Starting point: one front-desk staff member covers phones, check-ins and scheduling, and the scenario starts from an 18% no-show rate. Deployment: text-message scheduling and reminders, drafts against the clinic calendar, reminders at 48 hours, 24 hours and 2 hours before the visit, with anything clinical and every cancellation going straight to a person. Model inputs: 40 hours a week, one employee, 55% handed off, $35 an hour loaded — 22 hours a week, $3,334 a month of routine time against $149 a month. Targets: no-shows from 18% to 11%, 100% of reminders on a schedule, 22 front-desk hours handed off per week, a 24/7 booking window. Targets, not measured results.
---
## Blog

### https://unistaff.ai/blog — Blog
Notes on AI employees, hiring, support, sales and security. Ten articles.

### https://unistaff.ai/blog/ai-digital-employee-vs-chatbot — AI employee vs. chatbot vs. assistant: the real difference
Published 2026-08-10, updated 2026-08-17. A chatbot answers the message in front of it. An assistant works inside a thread you drive. An AI employee holds a role: memory between runs, a work queue, written limits, and a person on the send. The second memory mechanism described in the article: the output of one successful run becomes the input of the next role in the chain — the analyst's decision feeds the writer, the writer's drafts feed the proposer — so a decision made on Monday still shapes what gets drafted on Friday.
FAQ (verbatim)
Q: What is the difference between a conversational AI chatbot and an AI assistant?
A: A chatbot sits on a channel and answers whoever writes to it from the material it was given. An assistant sits in a thread you drive and does what you ask of it there. Both hold context for the length of the conversation. Neither one carries a role, a work queue, or a rule about what it may do without you.
Q: How does an AI employee remember things between conversations?

### https://unistaff.ai/blog/ai-receptionist-small-business — AI receptionist for small business: what it handles, what stays human
Published 2026-08-12, updated 2026-08-17. An AI receptionist is a digital front desk that reads every inbound request, answers the routine questions from your own material, qualifies what came in and logs the run. At Unistaff.ai it works over text channels, a person sends every reply, and one AI employee costs $149 a month on Starter. Phone answering sits outside that today: Unistaff ships no telephony connector, no SMS provider and no calendar connector, so the front desk that runs at 3 a.m. is the text one — a form submission at 3 a.m. is read, drafted and queued at 3 a.m. The four moves per request are intake, answer, qualify, journal; the loop runs every fifteen minutes and picks up at most five requests per pass.
FAQ (verbatim)
Q: Does an AI receptionist answer the phone?
A: On Unistaff, no. Text channels run today — a web form or quiz posts to the lead endpoint, the request is stored, and the first reply is drafted on the next run. A voice front desk is custom scope on an Enterprise plan and starts with a scoping conversation about telephony and consent. Call tracking is on the integrations roadmap.
Q: What does an AI receptionist cost?

### https://unistaff.ai/blog/ai-bookkeeper-replace-manual-bookkeeping — AI bookkeeper: can it really replace manual bookkeeping?
Published 2026-08-13. AI bookkeeping tools cut data entry, catch anomalies and close books faster. They still need human oversight for judgment, policy and compliance.

### https://unistaff.ai/blog/ai-customer-service-deflection-vs-resolution — AI customer service agent: deflection vs. resolution
Published 2026-08-14. Deflection measures how many tickets never reach a human. Resolution measures how many customers actually leave satisfied. The article argues for optimizing both and shows how each is counted.

### https://unistaff.ai/blog/ai-sdr-what-it-actually-does — AI SDR: what it actually does (and what it does not)
Published 2026-08-11. An AI SDR can research, personalize, follow up and prepare meetings. It cannot replace the human judgment that closes complex deals.

### https://unistaff.ai/blog/why-sales-teams-hire-ai-first — Why sales teams hire an AI employee before they hire another human
Published 2026-08-07. Humans close relationships. AI employees handle the repetitive share — qualification, follow-up, record updates, meeting preparation — so reps focus on deals that matter.

### https://unistaff.ai/blog/training-ai-on-your-company-knowledge — How to train an AI employee on your company knowledge
Published 2026-08-06. Your documents and brand voice become knowledge items the role reads on every run, so the answers stay yours and the escalations stay honest.

### https://unistaff.ai/blog/customer-support-at-scale — Customer support at scale without losing the human touch
Published 2026-08-04. Speed and empathy can rise together: routine issues resolve from drafts while complex cases reach humans with full context.

### https://unistaff.ai/blog/security-and-compliance-for-ai-workers — Security and compliance checklist for AI employees
Published 2026-08-03. Data residency, audit logs, SSO and DPA belong on every AI-employee shortlist. The article is the checklist to run before granting access.

### https://unistaff.ai/blog/future-of-work-2030 — The future of work: humans managing AI employees by 2030
Published 2026-08-05. Team leads manage squads of specialized AI employees while humans set strategy, review edge cases and own creative judgment.
---
## Français

### https://unistaff.ai/fr/automatisation-du-recrutement — Automatisation du recrutement
Répond à : qu'est-ce que l'automatisation du recrutement, comment se construit un système de recrutement automatisé, et que reste-t-il au recruteur ?
L'automatisation du recrutement confie à un logiciel les tâches répétitives d'une embauche : notation des candidatures sur vos critères, rédaction des invitations, des relances et des refus, checklists d'onboarding, synthèses de références. Chaque message part après validation humaine, et la décision d'embauche reste au recruteur. Chez Unistaff.ai, un employé numérique RH lit chaque candidature arrivée par votre formulaire, pose une note avec la raison écrite à côté, et prépare les messages. Tarif : 149 $ US par mois et par rôle, 599 $ US pour jusqu'à cinq employés numériques, Enterprise sur devis. Un seul connecteur de plateforme est livré aujourd'hui : Meta Ads. Version anglaise de cette page : https://unistaff.ai/solutions/hr
FAQ (verbatim, en français)
Q: Qu'est-ce que l'automatisation du recrutement ?
A: C'est le transfert à un logiciel des tâches répétitives du processus d'embauche : réception des candidatures, notation sur des critères écrits, rédaction des invitations et des refus, checklists d'onboarding, synthèses de références. Le jugement sur la personne et la décision finale restent chez le recruteur.

### https://unistaff.ai/about — The team behind your first AI employee
Answers: who builds Unistaff.ai, and which legal entity stands behind it?
Unistaff.ai is built by Innova Consulting Group Inc., based in Ottawa, Ontario, Canada. Contact: hello@unistaff.ai. The product grew out of the marketing operating system the group built to run its own work — an agent runtime, a mission engine and a lead pipeline the group operates for itself — and it still shares a codebase with that system: a demo request from this site lands in the same lead table the group's own operators work from. Stated mission: hand the repetitive share of a desk job to a digital employee so founders and small teams spend their time on customers, product and growth.

### https://unistaff.ai/contacts — Book a demo or start your pilot
Answers: how do you reach Unistaff.ai?
Email hello@unistaff.ai, phone +1 613 867 7767, reply within one business day. The contact form asks for name, work email, company, the role you want to automate (sales assistant, support agent, HR screener, e-commerce operator, or something else) and a message. Built by Innova Consulting Group Inc. in Ottawa, Ontario, Canada.

### https://unistaff.ai/company/partners — Partner with Unistaff.ai
Answers: how do agencies, integrators and consultants work with Unistaff.ai?
Four partner models: agency resellers (white-label AI employees sold inside your service packages), integrators (custom connectors and workflows for clients on the Enterprise plan), technology partners (co-selling with CRM, helpdesk and e-commerce platforms), and consultants (helping clients choose roles, write briefs and measure ROI).

### https://unistaff.ai/company/careers — Careers
Answers: is Unistaff.ai hiring, and where are roles posted?
A small remote team, deliberately lean while the product takes shape. Hiring status as published: no open positions right now, and a role will be posted on this page first when one opens. Introductions are welcome at hello@unistaff.ai.
---
## Notes for citation
- Prices quoted anywhere in this file are the published list: $149/mo Starter, $599/mo Scale (up to 5 employees), Enterprise custom. The billing unit is the digital employee.