How we use data

The short version

• Your company’s data is yours. We hold it to run the product for you and for nothing else.
• We do not sell data, and we do not use it to train any AI model — ours or anybody else’s.
• Nothing is shared with another company on tewo. One workspace never sees another.
• You can export it at any time, and ask us to erase it when you go.

What each part holds

Employees and Organization

The roster: names, contact details, employment terms, documents the company or the person uploads, departments and reporting lines. Pay figures sit behind their own permission, separate from being an admin — a company can give somebody the roster without giving them salaries.

Jobs & Candidates

Postings, and what an applicant sends: name, contact details, CV, answers and any cover letter. A public posting is reachable without an account; an application is not. Where AI screening is switched on, the CV and the posting are sent to the model to be read against each other. See below.

Leave, Timesheets and Performance

Requests and balances; hours against projects; self-reviews, manager ratings and appraisals. Reviews are read by whoever the review names and by whoever signs it off — never by the whole company. An employee is shown a letter grade rather than the number they were scored, and the server does not send the number to their screen.

Clients, Leads & Sales, Projects

Your customers’ and prospects’ details, agreements, invoices, payments and the pipeline. Where a company puts a tewo form on its own website, what people type into it arrives here — and that company is answerable to those people for it, not us.

Drive, Chat and Calendar

Files and folders, conversations, calls and meetings. A conversation is not erased when one person leaves, because everybody else in it was part of it too — we say so plainly before anybody presses erase.

Assets and Support

What the company owns and who holds it; tickets and anything attached to them.

What we hold for ourselves

Who created the workspace, who signs in and when, billing details, and the ordinary records a server keeps — addresses, times and errors — so that we can keep it working and see when somebody is attacking it. These are ours to answer for.

Where AI is used, and where it is not

AI appears in seven places. Each runs only when somebody uses that feature, and only what that feature needs is sent:

• Reading a CV against a posting. The document and the posting go to the model; it returns a score, the lines from the CV it scored on, and what the reading cost. The result sits on the application where you can check the working.
• Pulling the details out of a CV when somebody imports one, so a person does not retype it.
• Turning a recording into text — a voice note, or a call somebody chose to record.
• Summing up a recorded call from its transcript.
• Drafting text where somebody asks for a draft, from what they typed.
• Reading an enquiry into columns — a lead form’s answers, or a pasted message, into name, need and budget.
• Suggesting acceptance criteria for a task from its title and details. It proposes; a person decides.

The providers are named on the privacy page. Their terms with us say inputs are not kept beyond the request and not used to train their models. Three things are true of all of them. Nothing is used for training. Nothing is decided by a model— a score orders a list, and no part of this product rejects anybody automatically. And every output shows its working, because a number nobody can check is worse than no number.

Who else touches it

Only what the product genuinely calls: the hosting and database provider, object storage for files, an email provider for the messages the product sends, and the AI providers named above when those features are on. Each is bound to use it only to provide that service. The current list is in the privacy notice and changes to it are announced rather than quietly made.

What we will never do with it

We do not sell it. There is no arrangement under which anybody pays us for access to what is in your workspace, and there will not be one.

We do not pool it. No benchmarks, no industry averages, no “companies like yours pay X” — nothing of yours is mixed with anybody else’s to make a product. Your pay bands, your margins and your pipeline stay yours.

We do not train models on it. Where the product uses AI — reading a CV against a job description, writing a draft — the text goes to the provider for that one request and comes back; it is not kept by us for training and we do not send your records anywhere to improve a model, ours or anybody else’s.

We do not look at it for curiosity. Somebody at tewo may open your data when you ask us to help with something, or where we must to keep the service running or to meet a legal obligation. Not otherwise.

How long it is kept

For as long as the workspace is open. Deleting something puts it in Recently deleted for thirty days, where it can be brought back; after that it is gone for good. Invoices and tax records are kept for as long as the law requires. When a workspace closes, its data is erased within ninety days unless the law says otherwise.

Asking us something

Write to hello@amionyx.com. If the data belongs to a company’s workspace we will pass the request to them, because it is theirs to decide, and we will tell you we have.