Privacy

Privacy notice

PLACEHOLDER: have this reviewed for your jurisdictions before launch (Canada's PIPEDA + CASL for the author's base; GDPR/ePrivacy and CAN-SPAM where readers are). See PRE-DEPLOYMENT.md.

What we collect

If you subscribe, we store your email address, the date and source of your signup, your consent record, and your subscription status. That's it.

How we use it

Only to send you The Fourth Stage and to honor your preferences. We use a double opt-in: you must confirm before we ever email you content.

Your choices

Processors

Delivery is handled by Resend; the site runs on Cloudflare. We do not sell your data or use invasive tracking. Traffic is measured with privacy-friendly Cloudflare Web Analytics.


Free tools

This section covers the free tools at /tools, including the AI Use Case Brief. It is deliberately specific, because these tools ask about your own organization and you deserve to know exactly where that goes.

AI Discovery Session

The AI Discovery Session works differently from everything else on this page, so read this rather than the sections below. It runs in your browser, and your notes are held on your own device. Unless you press Save online, nothing you type is ever sent to us.

If you press Save online, the session is encrypted in your browser before it is sent, and what we store is that encrypted copy and nothing readable alongside it. The key that unlocks it is part of the link you are given, after the #, and browsers do not send that part of a link to the server, so we do not hold the key. We keep the encrypted copy for 90 days from the last time it was saved: saving it again starts the 90 days over. Once that has passed, the link stops opening for anyone, and the copy is then deleted from our database, a few at a time, whenever the tool saves a session.

Anyone you give the full link to can read the session, so share it the way you would share the notes themselves.

What it does send is a count of eight events, as daily totals: a session started, a session resumed from a saved draft, a session saved to a file, a session saved online, a readout printed, a readout exported as a PDF, the question editor opened, and a question added to it. That is the complete list at the time of writing, and a test fails the build if code is added that sends anything else. It tells us how much the tool is used and which parts of it; it tells us nothing about what was discussed, and the text of a question you write yourself is never sent.

There is no identifier, no IP address kept, no cookie, and none of your notes, figures or organization name. Not truncated or hashed — they are never sent at all, so every row is a count that could not be about one person. There is a switch in the tool to turn it off, and it is off automatically if your browser sends a Do Not Track or Global Privacy Control signal.

What you give us (AI Use Case Brief)

Whatever you type into the interview: a description of a task, rough volumes and durations, and multiple-choice answers about judgment, data and governance. You are never asked for a customer name, a colleague's name, or anything about an identifiable individual. please don't enter any. There is no account and no login.

Who processes it

The interview and the written brief are produced using Cloudflare Workers AI, which runs on Cloudflare's own network. Your answers are not sent to OpenAI, Anthropic, Google or any other third-party AI provider, and they are not used to train any model. The site, its database and the inference all sit with the same processor: Cloudflare.

Your answers are not written to request logs, and the wording of your brief is produced from a fixed instruction set that we control.

What we store, and for how long

If you publish a result card

Sharing is optional and off by default. If you create a result card, the only things published are your maturity stage, the use-case pattern, the readiness, value, feasibility and effort scores, and the delegation decision. Your company, your sector, your free-text answers and the brief itself are never published. The table behind that page has no column to put them in.

You can revoke a published card at any time from the tool page on the device you created it on. Revoking takes effect immediately: the page returns a 404 from that moment.

What the tool is, and is not

The brief is generated from your own answers. The scoring, the maturity stage, the governance call and every figure are produced by fixed rules, not written by a model; the AI conducts the interview and writes the prose around those findings. It is a structured start to a conversation, not professional, legal or financial advice, and you should check every figure against your own numbers before quoting them.

Deleting your data early

Everything expires on its own within 30 days (for a Discovery session saved online, 90 days from the last save). To remove something sooner, email hello@benedictnwankwo.com with the link to your brief or result card. For a Discovery session, send the link with everything from the # onwards removed: that part is the key, and we do not need it to delete the copy.