Skip to content

Data ownership and portability

You own the custom code and the data built for you, and we keep our reusable tooling. This page spells out what falls on each side of that line, where your data lives, how you get it out and what we hold after an engagement ends.

What What it includes
Everything in your system of record Your ERP data was yours before we arrived and stays yours. Nothing we build changes that.
The four Discovery Sprint deliverables The Operations Map, the Systems and Data Audit, the Automation Backlog and the Working Prototype. You keep them whether or not you continue.
The custom code built for you The screens, queues, rules and integrations of the operations layer. The instructions, rules, limits and thresholds written for your agents. The ERP-like core, if we built one.
The data built for you The documents the software received, drafts, approvals, exceptions and how they were resolved, the audit trail, the measurements for each workflow and the records of promotion reviews.

We keep our reusable tooling: the general-purpose code and tools our forward-deployed engineers bring to every engagement. Nothing in it is specific to your company, and it holds none of your data, your rules or your documents.

We keep none of your data after an engagement ends. The last section of this page says what that means in practice.

Two agreements settle this before any of your data is copied anywhere. Where our engineers’ working copies and the Working Prototype run is agreed in writing with your IT manager before day one of the Discovery Sprint. The hosting location for the software is agreed in your contract.

We state no default here, because the right answer depends on your IT policies, your ERP and who will run the software in the long term. Stages and terms lists what your contract records.

Three things hold wherever it is hosted.

  • Your ERP data stays in your ERP. The operations layer holds what the workflows need: the working data listed above and, where it helps, a read-only copy of ERP data that is refreshed from the ERP.
  • Test data is treated like production data. A test copy of your data lives in the agreed location, under the same access rules.
  • Data sent to an AI model is limited to what a workflow step needs. AI providers and your data covers what is sent, to whom and on what terms.

Because the data is yours, you can ask for an export at any point in an engagement, not only at the end. One rule applies to every export: you can read it without any of our tooling.

What Export format
Records: work in queues, drafts, approvals, exceptions, rules and limits CSV files, one per table, with a description of every column
Documents the software received: purchase orders, invoices, emails The original files, exactly as they arrived
The audit trail CSV files with one row per action, and the documents each action refers to
The whole database A full export in the database’s own standard format, with the table layout documented
The custom code The complete source code, with its change history

CSV is the choice for records because it opens in Excel and loads into any database. If you run an ERP-like core that we built, its orders, inventory, purchasing and invoicing records are exported the same way.

You can stop after any stage. When you do, three things happen.

  1. You keep your code and data. That means the running software, a full export or both. Handover and exit covers what is handed over and how another team takes over.
  2. Our access is removed. Your IT team disables every account our engineers hold in your systems. If the AI model provider account or the hosting account was ours, it moves to you. Access and credentials covers the steps.
  3. We delete what we still hold. That includes the working copies our engineers made, such as notes, sample documents and data extracts. If you need a deadline for the deletion or a written confirmation of it, that goes into the security terms in your contract.

After that, we hold our reusable tooling and none of your data.