AI in ERP: Where It Helps, Where It Doesn't, and How to Start
A grounded look at AI in ERP: document processing, support and sales assistance, natural-language reporting — plus the limits, risks, controls and how to choose a first use case.
Solution · AI
Overview
The salesperson stays in control: AI prepares, people decide.
01Typical problems
Reviewing history before every call.
Deals without next steps.
Repetitive quote building.
Orders retyped from emails.
02Capabilities
Orders, issues and opportunities summarised.
Prompts based on activity gaps.
Drafts from requests and history.
Orders parsed into drafts.
Flags stale or inconsistent deals.
Follow-up emails with context.
03Workflow
Summary before meetings.
Drafted follow-ups sent after review.
Drafts priced from ERP rules.
Orders created from emails.
Pipeline cleaned weekly.
04 — Technical notes
Pricing must come from the ERP's rules, never from AI-generated numbers.
Email access should be scoped to sales mailboxes with clear retention policies.
Where it applies
Insights
A grounded look at AI in ERP: document processing, support and sales assistance, natural-language reporting — plus the limits, risks, controls and how to choose a first use case.
A practical look at Odoo 20 AI agents: creating records, automated triggers, MCP access, IAP credit billing, and where we would and would not use them.
FAQ
No. Prices come from ERP pricing rules; AI only drafts structure and text.
Yes, with ERP CRM or integrated external CRMs.
We use providers with appropriate data terms and limit data to what's needed.
Next step
We will identify AI use cases with clear payback.