What not to share

What not to share

SUPERWISE Chat is built to be safe for real work — your conversations stay inside your organization’s private workspace, your privacy is protected, and your administrators have strong controls. So why have a page about what not to share?

Because good judgment about sensitive information is something you bring to every tool you use, not something any product can fully decide for you. This page explains the kinds of information worth pausing over before you paste them into a chat, why that pause is worth it, and the simple habits that let you get the most out of Chat without second- guessing yourself.

The short version: share the problem, not the secret. You almost never need to paste a password, a full customer record, or a regulated document to get a great answer — and once you see why, you’ll find it natural to leave those bits out.

Why this matters at all

Chat is genuinely private. Each organization’s conversations, files, and knowledge are kept completely separate from every other organization’s, and your one-to-one chats aren’t visible to your colleagues. (See Your privacy for the details.)

But “private to your organization” is not the same as “invisible to everyone in your organization.” Two everyday realities are worth keeping in mind:

• Your conversation is data. Anything you type becomes part of a stored conversation in your workspace. That’s what lets you scroll back, search, and pick up where you left off — but it also means a secret you paste is now written down somewhere, just like it would be if you’d put it in a shared document.
• Some content is shared on purpose. When you share a conversation, post in a channel, or save something into a folder or your organization’s shared knowledge, you’re inviting other people in. That’s a feature — but it means a sensitive detail can travel further than the moment you typed it.

None of this is cause for worry. It’s exactly the same care you’d apply to email, a wiki, or a shared drive. The point is simply this: treat a chat the way you’d treat a written note that others on your team might one day read.

The data worth pausing over

Here’s a practical list of the categories most people decide to keep out of a chat, and the plain reason for each.

Don’t paste	Why	Do this instead
Passwords, API keys, access tokens, secrets	A credential in a conversation is a credential written down. Describe the task, not the key.	”I have an expired API token for service X — what’s the safe way to rotate it?”
Full customer or employee records	Names tied to emails, addresses, account numbers, or health/financial details are personal data you’re responsible for.	Strip the identifiers. Ask about the pattern (“how do I phrase a refund apology?”) rather than the person.
Government IDs, payment card numbers, bank details	Highly regulated and rarely needed to get help.	Share the type of problem, not the actual number.
Confidential documents you weren’t cleared to redistribute	Pasting moves a controlled document into a new place.	Summarize the question you have about it, or check your team’s policy first.
Anything covered by a contract, NDA, or regulation (legal, medical, financial under specific rules)	Your obligations follow the data wherever it goes.	Use anonymized or hypothetical examples.
Other people’s private messages or info shared with you in confidence	It isn’t yours to redistribute.	Speak in general terms about the situation.

A simple test covers almost every case: “Would I be comfortable if a teammate read this exact message six months from now?” If yes, share it. If you hesitate, that hesitation is the signal to anonymize first.

You usually don’t need the secret to get the answer

This is the part that surprises people: removing the sensitive bits almost never weakens the help you get. Chat is reasoning about your problem, and the problem is rarely the literal data.

Compare these two ways of asking the same thing:

Over-shares: “Customer Jane Doe, jane.doe@acme.com, account #44192, called furious about a double charge of $1,420 on her Visa ending 7781 on June 3 — write her an apology.”

Shares only the problem: “A customer was double-charged and is upset. Write a warm, professional apology that acknowledges the mistake, explains we’re refunding it, and offers a goodwill gesture.”

The second version gets you an equally good draft — often a better one, because it’s not cluttered with details the answer doesn’t need. Then you fill in the name and amount yourself, in the final document, where they belong.

The same trick works everywhere:

• Debugging: paste the error and the relevant code, not the production credentials in your config file. Replace real keys with YOUR_API_KEY.
• Analysis: describe the shape of your data (“a spreadsheet of monthly sales by region”) rather than uploading the row with every customer’s name.
• Drafting: use placeholders — [Customer Name], [Order #] — and swap in the real values afterward.

The anonymize-then-personalize habit

Ask with placeholders, get a great answer, then paste the real details into your own document. You keep the quality and keep the secret out of the chat. It quickly becomes second nature.

What Chat already does to protect you

You’re not doing this alone. SUPERWISE Chat is designed so that the careless mistake is caught, and the data you do share is governed — your good judgment is the first layer, not the only one.

• Your work stays in your workspace. Conversations, files, and knowledge are isolated to your organization and kept separate from every other organization’s.
• Sensitive content is reviewed automatically. As Chat works through a request, the information flowing through it is screened, and personal or confidential material that doesn’t belong can be filtered out before it’s used. This happens quietly in the background — see Why it said no for the times you’ll actually notice it.
• You’re told when something is being handled for safety. If a message is reviewed before it’s processed, Chat tells you plainly rather than acting silently.
• You control optional data uses. Things like long-term memory of your preferences are yours to turn on or off. Essential, safety-related processing stays on, and Chat is open with you about why.

Think of it as a guardrail on a mountain road: it’s there for the moment you’d benefit from it, but the safest drive is still the one where you don’t need to lean on it.

A few good habits

1. Pause before pasting bulk data. A single paragraph is easy to scan; a thousand-row export is easy to forget you shared. If you’re about to paste a lot, ask whether you need all of it.
2. Strip identifiers by default. Names, emails, account numbers, and IDs are the usual culprits. Remove them unless they’re genuinely necessary to the question.
3. Use placeholders for secrets. [NAME], YOUR_API_KEY, [ACCOUNT #] — Chat understands the intent perfectly well.
4. Mind where it’s going. Be extra careful with anything you’ll share to a channel, save to shared knowledge, or post where colleagues can see it — that content travels.
5. Follow your own organization’s rules. Your company’s data and acceptable-use policies always come first. When in doubt, check with your admin or security team.

If you shared something by mistake

It happens. You can delete a conversation, and your organization’s administrators can help if sensitive content needs to be removed more thoroughly. For passwords or access keys you pasted, the safest move is to rotate (change) them — once a secret has been written down anywhere, treating it as exposed is the responsible habit. See Your privacy for what you can delete yourself.

The takeaway

You don’t need to be anxious about using Chat at work — that’s the whole point of the privacy and safety built into it. You just need one simple instinct: share the problem, not the secret. Describe what you’re trying to do, leave out the credentials and personal records you don’t need, and fill in the real specifics yourself at the end.

Do that, and you get all the help with none of the worry.

Related reading

• Your privacy — what’s kept, what’s private to you, and what you can delete or turn off.
• Safe at work — why Chat is built to be trusted with real company work.
• Why it said no — what’s happening when Chat reviews or holds back a response.
• How Chat is different from a public chatbot — why your work stays your organization’s.
• How your data flows — where your message goes and what’s recorded along the way.