What Is an AI Agent? A Plain Definition for Business Owners

An AI agent is software that is given a job rather than a question. A chatbot answers "is the 2-bedroom available?" when you type it. An agent reads the email that asked, checks your listing, writes the reply and proposes it for your drafts, and does the same for the next forty you hand it, or, with Start from new email turned on, for each new inquiry within about a minute.

That is the whole difference: an agent acts in your tools toward a goal you set, across many steps, and keeps going. Everything else about the definition is detail. The detail that matters to a business owner is what stops it from acting badly.

The three parts of a useful agent

A job it understands. "Log each tenant's maintenance request with its urgency and draft a reply that says what was logged." Not "help with maintenance". The narrower the job, the better an agent does it and the easier it is to check.

Tools it can use. Your inbox, your calendar, your spreadsheet. An agent that can only talk is a chatbot; an agent that can read the email and write the row is doing the work.

Limits. What it may read, what it may change, and when it must stop and ask. This is the part most "AI agent" pitches leave out, and it is the part that decides whether you can leave the agent running while you are at a showing.

How an agent differs from automation you already know

A rule in your email client ("if the subject contains 'invoice', move to Billing") is automation. It does exactly one thing, the same way, and breaks the day someone writes "inv." An agent reads the message and decides it is an invoice the way you would, then does the next step. That flexibility is why agents handle the messy inbox work rules never could, and why they need limits rules never needed.

Zapier-style connectors are automation too: when this happens in app A, do that in app B. They are good at moving data between apps on a trigger. An agent is good at the work a person would do inside the apps, with judgment, which is where the time goes in a small office.

What an agent can do for a small business today

Tensor Autonomous is one agent that does office work inside limits you set. Its Library holds the jobs that have been proven on real inboxes and calendars, each with its exact line:

Reply to rental inquiries: drafts a short, friendly reply to each rental inquiry from your listing facts, ready for you to send.
Log maintenance requests: logs each tenant's maintenance request with its urgency and drafts a reply that says what was logged.
Book showings: books each showing request on your calendar at a free time within your showing hours, or says why it can't.
Log new leads and draft a reply: logs each new client inquiry in your leads sheet and drafts a reply that asks for a short call.
Follow up on open quotes: drafts a friendly follow-up for each quote with no answer after 5 days, and notes it in your quotes sheet.
Handle cleaning booking requests: logs each cleaning booking request and drafts a reply with the price range from your price list.
Voicemail to callback list: turns each voicemail email into a row on your callback list, with the caller, number and reason.
Answer event date inquiries: checks your calendar for each date inquiry, holds a free date for 7 days and drafts the reply.

Anything else you describe in your own words, and Tensor sets it up the same way.

What makes it safe to run

Three things, and they are the questions to ask of any agent:

Does it ask before it changes anything? Tensor shows you each change before it makes it, drafts included; nothing it drafts is sent until you send it. When you trust a routine, you set limits (which kinds of change may run without asking (adding a row, creating a draft, adding an event), a daily cap, a days-and-hours window, and for each write an amount ceiling and a limit on recipients) and it runs within them. Sending email, payments and refunds, and moving or cancelling a calendar event always ask; limits can't cover them.
Does it stop when it is unsure? A request that does not fit the job, records that disagree, a question outside what it knows: it stops and asks instead of guessing.
Does it leave a receipt? After it writes, Tensor reads the record back and compares it with what it sent. Every run shows what it read, what it changed, who approved it and when.

What an agent is not

It is not your property-management, field-service or invoicing software, and it does not replace them; it does the work around them. It is not an answering service; it works in Gmail, Google Sheets and Google Calendar today, with Microsoft 365 to come. And it is not autonomous by default. It earns autonomy one routine at a time, as you watch it get the routine right and widen its limits.

Where to start

Product: responsibilities, approvals, limits and receipts in one tour.
Try it live: watch an agent do a real task on sample data, with your approval before the change.
Plans from $149 a month: Starter, Operator and Operations Team.

See it in a demo

Watch one agent do one job, end to end, with the approval and the receipt.

Try it live

What is an AI agent? A definition for people who run a business