• Solo operators
  • 2–10 person teams
  • Agencies
  • Anyone about to make a hire

A receptionist is $3,400 a month. This one is not.

Hire a coworker with a name, a job description, and exactly the permissions you decide. It works your follow-up, answers the phone, chases the reviews and drafts the replies — and asks you before it does anything you told it to ask about.

Start 14 days free $0 today · no contract · cancel any time
  • Unlimited contacts; provider usage billed separately
  • Scale trial access; partner tiers are separate
$0charged today
14 daysfree, on the top plan
No contractcancel before day 14, never billed
Unlimitedcontacts on every plan

You are not short of software. You are short of somebody to actually work the list.

What it is costing you

The problem is not that it is hard.

It is that nobody has the hours, and the cost of skipping it never shows up as a line on any invoice.

The work is not hard, it is relentless

Follow up, chase, remind, reply, ask for the review. None of it is difficult and none of it gets done, because there is nobody whose whole job it is.

A hire is a five-figure bet

Salary, tax, equipment, six weeks of training, and the risk they leave in April. That is a big commitment to make against a maybe.

An AI tool that acts on its own is worse than none

The reason most people will not let AI near their customers is that they cannot see what it is about to do. Autonomy without an approval step is not delegation, it is gambling.

How it works

Set up once. It runs from then on.

  1. 1

    Define

    Write the job description

    Name the employee, say what it is responsible for, and hand it the Brain that holds your business. That is the whole hiring process.

  2. 2

    Scope

    Decide what it may touch

    Permissions are per employee. Give one of them the phone and nothing else; give another the pipeline but no ability to send a single message.

  3. 3

    Approve

    Keep your hand on the wheel

    Anything you flagged lands in an approvals inbox before it happens. The Revenue Operator ships in shadow mode by default — it shows you the stalled deals worth recovering and sends nothing on its own.

  4. 4

    Scale

    Add the next one

    The second employee costs what the first one did: nothing extra. Unlimited agents on Scale, all working off the same contact timeline.

What you get

Every piece of this is already live.

These are not roadmap items. Each one is running in production today, described here in the same words the full product catalog uses.

AI Employees

Hire a coworker with a name, a job, and permissions you set.

A five-step hiring wizard — face and name, personality, duties, permissions, then the offer letter — produces a named AI coworker with standing duties, a budget and an approvals inbox, instead of a chat box holding the keys to everything.

  • Access granted per domain at four levels: nothing, view, do the work, or propose — where propose lets it ask to send, spend or delete and a person approves each one
  • Standing duties on a schedule down to every five minutes, each borrowing a named person's authority, with a run log for every execution
  • A duty that fails five times running pauses itself instead of failing forever
  • A daily token budget per employee, checked before each run, so nothing quietly loops up a bill overnight

Revenue Operator (shadow mode)

See the stalled deals worth recovering, without an AI sending anything on its own.

A read-only recovery board over at-risk pipeline: prioritized next moves, coverage, and a flagged-to-won-or-lost ledger that does not pretend correlation is causation. It proposes. It does not text, book, or collect until a later approval phase ships.

  • Ranks open risks from stored deal intelligence — no extra model call and no spend
  • Proposed next moves sit beside the deal, with coverage KPIs for the book
  • Currencies stay separate; partial populations are marked instead of rounded away
  • Sends nothing, starts nothing, and changes nothing — Copilot can open the board, not act from it

AI Brain knowledge base

Teach it your business once, and every agent knows it.

Organize everything your agents should know into topics and items — hours, pricing, services, policies, objection handling — then drop it into any agent's greeting, persona, or goals.

  • Upload what you already have: PDF, Word, spreadsheets, HTML, CSV, Markdown, or plain text
  • Or point it at your website and let it read the pages — one page, a section, or the whole domain
  • Merge the whole brain, one topic, or one item into any prompt
  • Word counts and copy chips so you can see what you're spending on context

In-app Copilot

Ask for the outcome instead of hunting through menus.

A workspace-aware assistant can search Chirply, answer questions from live CRM context, and propose the same actions available throughout the app.

  • Searchable conversation history and workspace-specific context
  • Reads and actions come from the platform's documented capability catalog
  • Approval required before it sends, spends, publishes, or destroys
  • Past chats can be reopened or removed from one place

AI throughout the app

The blank page problem, gone.

AI is woven through the CRM, messaging, and the funnel builder — draft a reply, summarize a contact, or generate a whole landing page.

  • AI page and funnel generation
  • Message drafting and contact summarization
  • A searchable model picker with live input/output pricing and context windows, so you compare cost before committing

Tasks & activity timeline

Nothing falls through, and you can prove what happened.

A tabbed activity feed on every contact with multi-select filters by channel and direction, plus tasks with priorities and due dates.

  • Calls, SMS, email, AI sessions, payments, and automation runs on one timeline
  • Filter to just inbound calls, or just email, in a click
  • Tasks created by hand, by an automation, or by an AI agent taking a message

The shared action engine

Add a capability once and it shows up in every menu.

Actions are defined in one registry with their fields and the surfaces they're allowed on. Workflows, call dispositions, and bulk actions all render from it.

  • Same action, same options, whether you're in a workflow or a bulk selection
  • Pickers for numbers, IVR menus, campaigns, sales bridges, AI agents, lists, and stages
  • Registry-driven, so the menus can't drift out of sync

This is the part of Chirply that solves ai employees. See the other 14 areas — CRM, scheduling, messaging, automations, funnels, commerce, reputation and the rest — all included in the same workspace.

The offer

Try it on your own business for 14 days.

You do not have to migrate anything to find out whether this works. Start the trial, set up the one thing you came here for, and judge it on that.

  • $0 charged today. A card holds the workspace; nothing moves until day 14.
  • Unlimited contacts. Communication channels depend on your plan; calls, messages and AI are billed by your providers.
  • Your providers, your rates. Bring your own Twilio, Mailgun and AI keys and pay them directly — we do not mark usage up.
  • No contract. Cancel before day 14 and you are never billed.

Your trial runs on Scale, the top plan, so nothing is locked while you evaluate. If ai employees is all you need afterwards, Scale at $97/mo carries it — switch down before the trial ends and keep the lower price.

Your order

Start free for 14 days.

$0 today. A card keeps the workspace running after the trial, and Scale renews at $97/month unless you change plans or cancel first.

Scale plan — 14-day trialThe whole platform, nothing locked
$0
From day 14, unless you switch or cancelChange plans or cancel any time before then
$97/mo
Due today
$0.00

Cancel any time before day 14 and nothing is charged. Calls, texts, emails and contacts are unlimited on every plan — you bring your own Twilio, Mailgun and model keys and pay those providers directly.

No trial subscription charge today. Cancel before renewal.

Provider usage and optional app purchases are separate. The card only exists so a workspace you decide to keep does not stop working on day 31.

Before you start

The questions everyone asks.

Will it message my customers without asking?

Only if you tell it to. Permissions are set per employee and anything you mark for review waits in an approvals inbox. The Revenue Operator defaults to shadow mode — it surfaces the opportunity and takes no action at all.

How is this different from an automation?

An automation runs the path you drew. An employee has a remit and reaches for whatever action fits — the same action catalog your workflows use, so anything it can do, you can already see and audit.

What does each extra employee cost?

Nothing on Scale. Agents are unlimited; you bring your own model key, so the usage is billed at cost by your provider rather than marked up by us.

Can I see what it did?

Every action lands on the contact's activity timeline like any other, alongside the human ones. There is no separate AI log to go hunting through.

Weighing it against something specific?