Privacy Policy
Valid from 19.11.2026.
This English text is the binding version.
This platform works with the customer's business mail archive and the tools the customer turns on. So this page states exactly what is processed, where the data lives, who reaches it and how it is deleted — no boilerplate.
Who is responsible for the data
The platform is provided by Well Chained d.o.o. The customer (the agency that connected its mailbox and accounts) is responsible for the data entered into the instance: they decide whose messages enter the system and who gets written to. We process that data on the customer's instructions, to run the platform. In data protection terms the agency is the controller of its clients' data and we are the processor; the terms are in our Data Processing Addendum, agreed before any client account is connected.
What data is processed
- Email messages from the mailbox the customer connected: sender, recipients, date, subject, message body and attachment metadata (name, type, size).
- Derived data: contacts and companies recognised from addresses, conversations, classification (whether anyone replied) and commitments extracted from message text, with a verbatim quote from the source message.
- Social media accounts the customer connects for publishing and comment replies, the posts, and the comments and messages that come in through them.
- Lead search results: company names and public professional data of people (name, title, company), and the public Instagram profile of a brand the customer asks to analyse.
- Brand material and briefs used in the Studio, and the images, videos and voice-overs made from them.
- Platform account data: username, password as a scrypt hash (never stored in readable form), time of last sign-in and last activity.
- Operational records: actions of the agents that import mail and of the platform's assistant, processing errors, language model usage (token counts and estimated cost).
The system collects no visit data, uses no analytics tools or ad networks, and does not profile visitors. See Cookies.
Access requests (the form on this site)
Whoever fills in the “Request access” form leaves their name, their agency's name, a work email and, optionally, a website address and a short note on what they want to automate. We also record the page language, the site the form was sent from, the version of this policy they agreed to and when. Well Chained d.o.o. is responsible for this data.
The data is used only to answer the request and judge whether we can help. It is not used for ads, not sold and not added to any mailing list. The system sends no automatic email: a person replies from their own inbox.
The request is kept in the shared part of the platform, outside any customer's space, and only authorised operators see it. It is kept for 12 months from the last contact and then deleted; if the request turns into a contract, the data follows the contract. You can ask for earlier deletion at the data contact below.
Abuse protection works without outside services and without extra cookies: the form has a hidden field and a timing check, and the number of requests per hour is limited.
Measurement in campaign messages
When the platform sends campaign messages on the user's behalf, measurement elements may be embedded: a pixel that records opens and redirected links that record clicks (time, message address, browser technical data). Measurement is a setting per instance; instances for UK agencies start with it switched off, and the agency decides whether to switch it on. Every message carries a visible unsubscribe link and the standard List-Unsubscribe header — unsubscribing is honoured immediately: the platform excludes that address from every campaign for good, and the customer must not add it back by hand.
The previous paragraph concerns recipients of campaign messages. For visitors of this site: no analytics and no tracking. The site sets only three cookies of its own, listed on the Cookies page.
Where the data lives
Every customer gets their own instance — a separate space in the database and separate settings, with a separate project for each of the agency's clients. Data is stored in Belgrade, Serbia, on servers run by Leniventus LLC, our subprocessor; backups are kept on a second server in Belgrade and as an encrypted copy at a second site in Čačak, Serbia. Serbia has no UK adequacy decision, so transfers from the UK are covered by the IDTA in our DPA. One customer's data is never visible to another. Details on the Security page.
Language model processing (AI)
To extract commitments, prepare drafts, write captions and answer in chat, the system sends parts of message text to an external language model provider. That is a material fact, so it is stated here explicitly:
- Each instance runs one AI provider: Anthropic by default, or a local model on the platform's own servers that sends nothing outside. Details on AI and your data.
- How a provider is added or changed, and the consent needed for one without a DPA, is set in section 6 of our DPA.
- The message text needed for processing is sent, without attachments. The result (commitment + quote) is stored in the customer's instance.
- The customer may switch this processing off — the platform then runs only its deterministic parts (import, threading, classification), sending no text outside.
Other providers
Features that need an outside provider list that provider, with what it receives, its country and its transfer safeguard, on the Subprocessors page. That page lists the providers in use and the optional ones an offer can choose.
Who has access
- Account users the customer created on their own instance.
- The assistant (software agent) that feeds in messages — it uses a separate key and can only send messages and start processing.
- A member's own AI client (for example ChatGPT or Claude), if the member connects one. It acts with the member's rights only, and what it receives is processed by that AI provider under the member's own account — see AI and your data.
- Two groups can see it. Engineers of Leniventus LLC, our subprocessor, run the servers, hold the backup key and maintain the platform's software; they reach your data only to run the service, fix faults, deploy updates and answer support requests, on servers in Serbia, without moving anything out of Serbia. Our staff at Well Chained d.o.o. have no access to the servers; they see your data only through a member account you give them, for onboarding or support. Both are bound by confidentiality.
How long it is kept and how it is deleted
Data is kept for the term of the contract. The customer may at any time ask for:
- an export of their data in machine-readable form,
- deletion of a single contact or message, or
- closing a client project or the whole instance — we send an export within 30 days, then delete it; backup copies expire about two months later.
Requests go to ivan.slovic@wellchained.com and are carried out without delay, and within 30 days at the latest.
Rights of people whose data is in the archive
People who appear in the correspondence (contacts) may ask to see, correct or delete their data. Since the archive belongs to the customer, such a request goes to them; we provide the technical support to carry it out. If someone asked not to be written to again, the system permanently excludes them from every campaign.
Security
Access to the platform goes through password sign-in; traffic is protected by HTTPS; agent access uses separate scoped keys, so a key for one purpose does not open another. Passwords are stored only as scrypt hashes. More on the Security page.
Changes
When processing changes under section 6 of our DPA (a different AI provider at the customer's written request, a new subprocessor after notice), this text and the valid-from date at the top change with it.
Questions about data and the DPA: ivan.slovic@wellchained.com (Well Chained d.o.o.). Other documents: Trust centre · Security · AI and your data · Subprocessors · DPA · Support · Usage · Terms · Cookies.