We gave our ERP's AI assistant the same permissions as the staff. Here's how it works with Claude over MCP
Most "AI in the ERP" demos have the same problem. The assistant talks to the database through a service account, so it can see and change far more than the person typing the prompt. That's fine in a demo. It isn't fine w
Most "AI in the ERP" demos have the same problem. The assistant talks to the database through a service account, so it can see and change far more than the person typing the prompt. That's fine in a demo. It isn't fine when the data is payroll, invoices and customer records.
When we started building ERPoida, we made one rule early: the AI never gets its own permissions. It works as the user, under the same checks, and every action is logged. This post walks through how that works and what it means once you connect Claude or ChatGPT over MCP.
The rule: one permission system, for people and agents
ERPoida has three levels of access control:
- Modules: which apps a user can open at all (CRM, Accounting, Payroll, and so on)
- Rows: which records they can see inside a module (a sales rep sees their own deals, a manager sees the team's)
- Fields: which fields they can read or edit (salary is visible to HR, not to the whole company)
The AI assistant doesn't sit next to this. It sits inside it. Every tool the assistant can call goes through a registry, and the registry checks the caller's permissions before the tool runs. If a sales rep asks the assistant to "show me everyone's salary," the request fails the same way it would fail in the UI. The refusal is logged too.
That gives you three useful properties:
- No privilege escalation through prompts. A clever prompt can't unlock data, because the data was never reachable.
- Audit you can actually read. Every change has a who, a what and a when, whether a person or an agent made it.
- The same rules everywhere. The built-in assistant, an automation and an external agent connected over MCP all go through the same gate.
Customization without a developer
The second thing we cared about is that every business runs a bit differently. In most ERPs, adapting the system means a consultant, custom code or both.
ERPoida is metadata-driven. Objects, fields and views are definitions stored as data, so:
- Views can be a table, board, calendar, timeline or chart, with your own columns, filters and grouping
- Fields can be added to any record, including invoices and contacts
- New modules can be created for things that don't exist yet: an equipment register, a grants tracker, a project type only your company has
Because the AI uses the same tools as the user, it can also do this work. You can ask it to add a field, set up a view or create an automation, and it does that inside your permissions, not around them.
Money is different: the database says no
Flexible data is great for CRM and projects. It's a bad idea for money.
So we split storage into two tiers. Things like CRM, HR and helpdesk live in a flexible store. The ledger, invoices, inventory and payroll live in strict relational tables, and the rules are enforced in the database itself:
- An unbalanced journal entry is rejected by the database, not just by the API
- Posted invoices can't be edited or deleted. Corrections go through credit notes, which keeps the books audit-proof
This matters even more with AI in the loop. You don't want "the model was confident" to be the only thing between you and a broken ledger.
Connecting Claude or ChatGPT over MCP
ERPoida exposes an MCP server, so you can connect Claude or ChatGPT to your workspace and work with your data in plain language:
- "Which invoices are overdue by more than 30 days? Draft a reminder for each."
- "Create a deal for Acme, link it to the contact I met yesterday and set a follow-up for Friday."
- "Add a 'Contract end date' field to companies and show it in the table view."
The assistant signs in as you. It can only do what you can do. If you can't approve payroll, neither can Claude.
ERPoida is also an MCP client. You can connect other MCP servers you already run, and their tools become available to ERPoida's assistant and automations, again under the same permission checks.
The setup guide is here: Connect Claude or ChatGPT.
Automations, with a record of every run
Automations are saved instructions that run on a schedule or when a record changes, for example when a deal moves to "Won" or an invoice becomes overdue. Each run is logged, so you can see what happened and why.
What's in the box
ERPoida covers CRM, sales, accounting, payments, inventory, purchasing, manufacturing, people (HR), payroll, projects, helpdesk, point of sale, marketing and contacts. The interface is available in English, German and Arabic, and the finance side includes German localisation.
Try it, or tell us where we're wrong
We'd love feedback from people who have fought with ERPs before, especially on the permission model and on what you'd want an agent to do inside one.
- Website: erpoida.com
- Docs and guides: erpoida.com/help
- Tutorials: youtube.com/@erpoida-official
Written by the ERPoida team, with help from AI tools for drafting and editing.
Originally published by Dev.to Security. Aggregated on AIWithGhost for educational purposes — full credit and traffic to the original publisher.

