Acceptable Use Policy

Who this applies to

This policy is part of the terms of service and applies to everyone who uses the product: the account owner, every person on a seat, and anybody they let use their sign-in.

What you must not do

Do not use the product to break the law, to infringe somebody's rights, or to handle information you have no right to hold.

Do not attempt to reach another customer's workspace, probe the service for weaknesses without our written permission, or interfere with how it runs for anybody else.

Do not resell the product, share one seat between several people, or automate the product in a way that is really a second product.

Do not upload malware, and do not use the product to store or distribute it.

Do not use the product to collect details of people from a site whose terms forbid it. Research reads a company's own public pages, and you must not point it at a site that forbids that.

Where web forms are switched on, a form you publish opens at an address on this site. Do not use one to ask for a password, payment details or any other credential, or to pretend to be us or anybody other than your own business. A form sends its visitors only to a page on one of the websites you list for it, and every form page carries a link anybody can use to report it to us.

Where quote pages are switched on, the link to a quote you issue opens at an address on this site. Do not use a quote to ask for a password, payment details or any other credential, or to pretend to be us or anybody other than your own business. Every quote page carries a link anybody can use to report it to us.

Using the AI features

These obligations come from our AI provider's usage policy and they bind you because we pass them on. Breaking them puts our access at risk as well as yours.

Do not use AI features to generate spam, bulk unsolicited messaging, or content designed to deceive the person who receives it about who is writing to them or why.

Do not present AI output as something it is not. If a message was drafted by a model, do not claim a person wrote it from scratch, and do not use the product to impersonate a real person or organization.

A person reviews every AI draft before it is sent. The product is built that way and this policy requires it: nothing this product generates goes to a recipient without somebody reading it first.

Do not use AI features for decisions that materially affect somebody's rights, safety, employment, credit or health without a person reviewing the decision.

Do not use AI features to produce content that harasses, defames, or sexualises anybody, or that promotes self harm or violence.

The mail you send, and the law that governs it

When you send outreach through this product, you are the sender. The obligations below are yours under the CAN-SPAM Act and its equivalents elsewhere, and the law is explicit that you cannot contract them away, including to us.

Your headers and your subject line must be honest. The from address, the reply address and the routing must identify you, and the subject must reflect what is in the message.

Every message must carry a valid physical postal address for you. Ours is not yours, and this product will not substitute one for you.

Every message must carry a clear and conspicuous way to opt out, and you must honor an opt-out within ten business days. The product adds an unsubscribe link and records what it receives; acting on it in time is still your obligation.

Say clearly when a message is an advertisement, and keep your own records of consent where the law where your recipient lives requires it.

Messages you write into a firm's contact form

Where writing to a firm's contact form is switched on, the product drafts the message and you submit it yourself, in your own browser. You are the sender, and the obligations in the section above apply to that message as they apply to an email: be honest about who you are and why you are writing.

The product adds a paragraph under your message naming the business your workspace sends on behalf of (your workspace's own name unless you set another) and its postal address, and saying that a reply is enough to stop. Keep it. When a firm asks you to stop, record it on the firm so nobody in your workspace writes to it again, and do not use a form to reach anybody the mail rules above would not let you write to.

The product leaves every tick box on their form for you to decide. Tick one that agrees to their terms or signs you up to anything only if you mean it.

The records you load

Only load records about people you are entitled to hold and to contact. You are the controller of those records and you decide what is collected and why; we process them for you.

That includes the people our AI research proposes and you accept. Accepting a person is your decision to hold their details, and what comes with it is yours: telling them where you got those details within the time the law where they are allows, and stopping when they object.

That includes a page you paste into the product to find the people it names. Paste only the text of a page the company itself publishes on its own site, give that page's own address, and paste nothing from a site whose terms forbid collecting it.

Do not load special category information, payment card numbers, government identifiers, or health information into a workspace. The product is not built to hold them and this policy forbids it.

Fair use of the service

The product applies rate limits, sending limits and daily caps. Do not work around them, and do not run the service at a volume that degrades it for other customers. If you need more headroom, ask us.

How we enforce this

Where we can, we ask you to fix a problem before we do anything else. Where a problem is urgent, or where it puts other customers or our providers at risk, we may suspend a capability or an account first and tell you immediately afterwards. Serious or repeated breaches end the agreement.