This policy explains what Speriential does with information, and what it does not do with it. It is written to be read rather than skimmed, and it describes how the platform actually works rather than the widest thing the law would allow.
Speriential is a business product. It is not intended for consumers and not directed at children.
1. Two kinds of information, treated differently
Almost every question about privacy here turns on which of these two is meant, so they are separated throughout.
Account information is what we hold because you are our customer: the name and work email of each person with a seat, your organisation, your plan, your billing details, and a record of what was done in your workspace and when.
Customer data is what you bring for the platform to work over: the files you upload, the rows you send through the API, the messages and documents pulled from systems you connect, and the recordings you ask us to transcribe. It also covers everything derived from those, including answers, evidence references and transcripts.
We are the controller of account information. For customer data we are a processor, acting on your instructions. If your customer data contains personal information about other people, you remain responsible for having a basis to process it and for telling those people, and we handle it only as you direct.
2. What we do with customer data
We read it in order to answer the questions you ask of it, and to produce the evidence behind those answers. That is the whole purpose.
We do not sell customer data, share it with advertisers, or use it to train models. We do not use one customer's data to improve results for another. There is no aggregated pool.
The platform's default posture is not to send text to any model at all. Its core is deterministic: it computes figures from your rows and quotes passages from your documents. A model tier exists and ships switched off. If you turn it on for a workspace, text from that workspace is sent to the model provider we name in section 5 for that request only, and is not retained by them for training.
Two processing paths send data outside our own infrastructure by necessity, and both are named in section 5: transcription of audio and video, and reading from the systems you choose to connect.
3. Where your data lives
Every workspace is assigned to one cell, in either the United States or Brazil, and its data is stored and read in that cell's region. Cells are separate deployments with separate databases, separate storage and separate encryption keys. A cell does not reach into another, and a workspace does not move between them without your instruction.
Within a cell, every row carries the workspace it belongs to, and the database enforces that separation rather than leaving it to application code.
If you need a specific region for legal or contractual reasons, tell us before you load data. We cannot move it afterwards without your involvement.
4. Systems you connect
You may connect Google Workspace, Microsoft 365 or WhatsApp Business so the platform can read from them. Three things are true of every one of those connections.
They are read-only. The permissions we request are read permissions. The platform cannot send, reply, delete, move or alter anything in a system you connect. WhatsApp receives messages sent to your business number and cannot send one.
You grant the scope, and you see it. Consent happens on the provider's own screen, in their words, listing what is being granted. We never ask for or hold your password to another service.
The credential is sealed. The token a provider issues is encrypted with a key held for that cell alone, and is never returned to any screen or API response after it is stored. You can disconnect at any time, which stops further reading immediately. Data already brought in stays in your workspace until you remove it.
5. Who else is involved
We use a small number of providers, and only these.
Amazon Web Services hosts the platform, stores your data, sends our email, and performs transcription when you upload audio or video. Data stays within the region of your cell.
Cloudflare provides our domain name service and the bot challenge on our public forms. It does not receive customer data.
Google, Microsoft and Meta are involved only where you have connected their services, and only to the extent of reading what you authorised.
Our payment processor handles card details for self-serve plans. We do not see or store full card numbers.
We do not add a provider that touches customer data without updating this page first.
6. How long we keep things, and what you can delete
Customer data stays until you remove it. You can delete a collection, a record or a whole workspace, and deletion removes the data from the platform's working store.
Account information and the activity record are kept while your account is open and for a period afterwards where we are required to keep them, for example for tax and accounting.
Access logs are kept for one year.
One exception, and it is important. A workspace can turn on write-once sealing, which keeps an unalterable copy of each original file or payload as it arrived, as evidence that an answer rested on what it claims. This is off by default and only applies where you switch it on. Where it is on, those sealed originals are held under a seven-year retention lock and cannot be deleted before that period ends, by you, by us, or by anyone else. That is the point of the feature, and it is the one place where a deletion request cannot be honoured in full. Consider it before enabling it.
7. Security
The measures below are the ones this platform actually relies on.
Data is encrypted in transit and at rest. A workspace may supply its own encryption key, in which case everything stored for it is encrypted under that key.
A tool reads a collection only where you have granted that specific pairing. Nothing is granted by default.
New capabilities ship switched off and are enabled per workspace rather than for everyone.
Suspending or closing an account cuts access at every door immediately, including the ones that carry no API key.
A result shared outside your workspace is never public. The recipient verifies their email address and signs in, you choose whether the underlying records travel with it, links expire, and you can see who opened it and when.
No system is perfectly secure, and we do not claim a certification we have not earned. If you need to know where we stand on a specific standard, ask us and we will answer plainly.
8. Your rights
If you are covered by the GDPR, the LGPD, or a similar law, you have rights of access, correction, deletion, portability and objection.
For account information, contact us and we will act. For customer data we act on the instruction of the customer whose workspace holds it. If you believe a business using Speriential holds information about you, approach that business: they control it, and we will support them in responding.
The one limit on deletion is the sealed originals described in section 6.
9. Changes
We will update this page when what we do changes, and the date at the top will tell you when. If a change materially affects how customer data is handled, we will tell account administrators directly rather than relying on you to notice.
10. Contact
Write to contact@speriential.ai.