It drafts the scope and the gap/fit; the consultant approves.
AI and human control
AI agents run the work; people own the decision and the responsibility.
Where the AI does the work
Each line says both what the AI produces and whose hands the output passes through before it reaches you. Separating the two would make the output look more certain than it is.
It only reads your installed system; it changes none of your settings. Measured directly on Odoo, questionnaire-based on Zoho and abas.
Approval is stamped onto the document with who approved it and when.
No change reaches your system without your approval: first a preview, then written approval. Every step is recorded, and steps that cannot be undone are identified before approval.
It opens no ticket; it points you to the right place. Answers come from the knowledge base of the product you bought.
Before an output reaches you
- 01The AI produces a draft
- 02A person reviews it
- 03The approval is recorded on the document with who and when
- 04A change reaches your system only after approval
The level of control varies by output type, and we say so product by product: your measurement reports are reviewed by an expert consultant, while some documents are generated by your own entitled user. Which is which is stated plainly in the assurance table.
Frequently Asked Questions
About the platform and the AI agents
What is MindDX Platform?
MindDX Platform is a digital transformation platform that governs enterprise software transformation with evidence, AI agents and human control. It brings measurement, discovery, design, training and support together in one evidence chain; at every step the decision and the responsibility stay with people.
Does «agentic» mean fully autonomous?
No. Agents run the work; people own the decision and the responsibility. The only path that writes into your system follows preview, written approval and push, in that order; the audit is read-only. We neither promise nor perform a fully autonomous ERP implementation.
Do the AI agents change the system on their own?
No. The only path that writes into your system follows preview, written approval and push, in that order; every step is logged and steps that cannot be undone are identified before approval. The audit side is read-only — that code path contains no call that writes to the system.
Which AI models do you use?
We use leading foundation-model providers; model choice, parameters and cost stay under MindDX's control, and models do not make decisions. Provider names and data processing details are published on the AI and Data Use page.
Does every output pass an expert approval?
No — and we say so plainly. The level of control varies by output type. Where an approval is given, it is stamped onto the document with who approved it and when; where it is not, that is recorded too.
How can AI be used safely in an ERP project?
Through three separations. First, what the AI PRODUCES is separated from what it DECIDES: the model drafts, a person decides. Second, the approval itself is recorded — who approved and when stays on the document, and where no approval is given that is recorded too. Third, writing to the system is a separate gate: a change is shown as a preview first, applied only after written approval, and steps that cannot be undone are identified BEFORE approval. The level of control varies by output type; which outputs pass human review is stated product by product in the assurance table.