AI policy

1.What AI is and does today

AI is the assistant built into Scale Suite. This policy is part of the Terms of service. Today AI can:

• answer questions about Scale Suite from our help and legal pages
• look up information in your workspace that your role is allowed to see, like invoices, and for the owner, payroll totals and bank balances
• add a to do in Tasks and draft an email that opens in Mail for you to read and send yourself
• do small jobs inside apps, like writing form questions, a spreadsheet formula, notes from a Video call, or an image

AI never sends, spends, shares, deletes or signs anything on its own. Before any of these, it shows an approval card with the exact action (what, to whom, and how much), and nothing happens until the person who asked approves it. An approval works once, expires after 15 minutes, and is checked again on our server against that person's role. Spending also needs the owner and stays under a limit. Rules that let AI act on its own are off by default.

AI output is labeled "AI-generated" wherever it appears: answers, drafts, summaries, Chat catch up, Video notes and designer picks. When AI talks for your business with people outside it, it says it is AI.

2.Who can see what

AI only uses what your role allows. The server checks your role on every request, not the AI model, so a cleverly worded message cannot widen access. Payroll, salaries and contractor pay, and bank balances are for the owner only.

AI treats email, documents, web pages and page content as information, never as instructions, so text hidden in an email cannot make AI change its rules.

3.Our AI providers

• Anthropic, PBC: the Claude models behind AI, for answers, drafts, AI notes in Video (only when everyone in the call agrees) and Catch up in Chat. We use smaller and larger models depending on the question.
• Cloudflare, Inc.: Workers AI, for images AI makes.
• Browser speech services from Google and Microsoft: live captions in Video use your browser's speech recognition, which may send audio to its maker, such as Google for Chrome and Microsoft for Edge. Each person decides whether to turn captions on.

For each request we send only what the task needs: your question and recent messages in that chat, the matching help pages, what is on your screen if the app includes it, and workspace records a tool looked up for your role. See Subprocessors.

4.No training on your content by default

The AI improvement setting is off unless you turn it on. While it is off, nothing your business puts in Scale Suite, like emails, documents, chats, files and AI conversations, is used to train or improve any AI model, by Scale or by our AI providers. Anthropic's commercial terms say: "Anthropic may not train models on Customer Content from Services." Cloudflare's terms for Workers AI [Mike: confirm current wording] do not allow it to train on what we send.

Without your opt-in, we use only usage signals to improve Scale Suite, such as how many questions were asked, which help page answered, whether a tool failed, response times, and thumbs up or down, without the words of your questions, answers or content. We keep those signals de-identified, we do not try to re-identify them, and anyone we share them with must agree not to either.

Business content means everything your business puts into or creates in Scale Suite, including mail, files, contacts, messages, documents, money records, AI chats and AI memory.

5.The AI improvement opt-in

You can choose to help us improve AI. The setting is called "Use my content to improve AI", it is off unless you turn it on, and it is offered in onboarding and in Settings, AI, with this explanation. The account owner or an admin may change it. Saying no never changes your price, plan or features.

If you turn it on, Scale may use your AI conversations and the AI drafts you keep or change, with names, email addresses, phone numbers and account numbers removed first, to make AI give better answers. It is never sold or used for ads, and our AI providers still may not train their models on it. It is never shown to other customers or used to answer them.

Today Scale does not train or tune any AI model, so turning this on changes nothing yet. If that starts, only accounts that turned this on are included, and we update this policy first.

You can turn it off any time in Settings, AI. From then on nothing new is used, and anything kept to improve AI is deleted within 30 days.

We record each choice: who made it, their role, when, on or off, where it was made, and the policy version shown, which is ai-policy-2026-10-28 for this page. The records cannot be edited. A choice made under an older version does not count for a newer one: if we ever want to use content in a new way, we ask again, and we never switch it on through a change to our terms.

6.Support access, a separate choice

Support access is separate from the AI improvement setting, and it is off by default. Only the account owner can turn it on, in Settings, AI.

• With it on, a person on the Scale support team can read your account settings and your AI settings and rules, read only, to fix something you asked about.
• They cannot send, pay, sign, delete or change anything, and they cannot open your Mail, Drive, Docs or Sign content.
• AI memory stays encrypted on your device, so support cannot read it.
• The owner picks 1 hour, 24 hours or 7 days. Access ends on its own when the time is up, and the owner can end it sooner. Ending it stops access at once.
• Every grant, every visit and every end is logged and shown in Settings. The log cannot be edited.

Without a grant, people at Scale do not look at your content except to investigate abuse, fraud or a security threat, to fix an outage, or when the law requires it. Any such access is recorded and, when the law allows, reported to you afterward. [Being built: a log entry and notice for this emergency access.]

7.AI memory

AI keeps short notes about your work, like a client preferring afternoons. You can see, edit, export and delete each note in Settings, AI, Memory. Do not put passwords, card numbers or bank details in memory. When AI uses a note for a task, the note is sent with that task to our AI provider.

8.A person decides what matters

Do not let AI alone decide anything about a person's job or pay, housing or tenancy, credit or lending, insurance, health care, education or legal rights. AI can help draft, sort or summarize. A person who understands the decision must review it and make it.

If your business uses AI to help with decisions like these, you are responsible for the laws that apply, including notices, fairness testing and appeal rights under state automated decision laws that start applying in 2027 [Mike: confirm which apply to Scale's customers]. See the Acceptable use policy.

9.Limits of AI

AI can be wrong, out of date or incomplete, even when it sounds sure. Check what it writes before you send, publish or rely on it. AI does not give legal, tax, financial or medical advice, and nothing it says is a promise from Scale. If AI and these policies disagree, these policies win.

As between you and Scale, you own what AI makes for you, to the extent the law allows. Others may get similar results, and AI output may not be protected by copyright.

10.Limits on use

Scale Suite and its AI are for adults using them for a business. When you sign up or reserve, you confirm with two separate checkboxes that you are 18 or older and that the account is for a business, not personal or household use. We keep a record of those two answers, when you gave them and the wording version, and never ask for your date of birth.

AI is included in every plan. To keep it fast and fair, there are rate limits, currently 20 requests a minute and 300 an hour from one network address. [Decision for Cameron: the monthly AI allowance for each plan, and what happens when it runs out. Show it on the plan before purchase.]

11.Changes to this policy

Each version of this policy has a version string, shown at the top. If we change it in a way that matters, we tell you at least 30 days ahead, in the app and by email. A change that would use your content in a new way needs your new opt-in.

12.Contact

Questions about this page: legal@scalembs.com, or by mail to Scale Modern Business Solutions LLC, 2415 San Ramon Valley Blvd, Suite 4, San Ramon, CA 94583.