Yours, And Only Yours

Your organisation has a home region: the European Union, the eastern or western United States, or Singapore. It is set when you join. Everything of yours is stored and processed there, and it moves only with your written consent.

Your system is separated from every other member’s where it matters most: in the database itself. A person signed in to one organisation can only ever see that organisation’s records, and sharing between organisations is switched off. We test that separation on every region, and we keep the dated results for your reviewer.

A member of our team can open your content only during a support session, for a stated reason, and for four hours at most. Each time that happens it is recorded in a log held in your own region, which your owners can read: who looked, at what, and when. Nobody on our team can add themselves to your organisation.

Your data is never used to train an AI model. IN DEPTH is not built on one AI company: each task goes to the model best suited to it, chosen from a fixed list of named providers, and we record which model did the work. Every training option on our accounts is switched off, and our providers state that what we send them is not used for training and is kept for up to thirty days, for abuse monitoring only.

A small number of outside services help us run IN DEPTH, each for one specific job: the model providers that write each output, two web-search services that return public web pages, a service that sends your sign-in emails, a product-analytics service hosted in the EU, and the database and hosting providers the platform runs on. We have read and filed the terms of every one of them.

Every connection to the platform is encrypted, and so are your data and its backups where they are stored. Your password is chosen by you and never sent by us.

If there is a person or an organisation you never want approached, put the name on your never-contact list. The system will not research them or draft anything to them, and it will not let the name be added again by mistake.

You can leave with everything. An owner can export all that we hold for your organisation as a single document, at any time, whether you are leaving or not. And if you ask us to erase your data, we erase all of it: your reports, the evidence behind them, your uploaded files, your usage record and your accounts. Backups are kept for seven days, for recovery only, so nothing you erase outlasts that.

If you apply and your application is not granted, the information you gave us and the research we built from it are deleted immediately.

Every commitment on this page is written into the Terms of Use you accept when your application is granted. A fuller account, How IN DEPTH Protects Your Data, and a technical overview for your reviewer are both available on request at apply@in-depth.io.

How do you know the output is right? Every claim made for AI’s speed rests on that question, and most never ask it.

It starts from sources, not memory. Every claim in a report carries the source it came from, and you can open it. If we cannot cite it, we do not say it.

Every report says how many sources it read and carries a confidence note. Where the sources may be describing two different organisations with the same name, the report says so rather than choosing.

A person on your team reviews at a named point, and nothing leaves without them. There is always a human on the other side.

And the machinery between a model’s output and that person’s review, the checking and verification the system runs before a human ever sees the work, is engineered, specific and written down.