Ask and analyse
- Natural-language questions across your business data
- Analysis across connected modules
AI assistant
Nexa AI
Powered by a locally hosted large language model (LLM), Nexa’s AI is embedded in the ERP. Ask a question in plain English and get an answer grounded in your own business data. Nexa AI explains why the numbers moved, flags issues before they escalate and suggests what to do next.
Ask your ERP
Questions you can ask in plain English
Illustrative examples across finance, inventory, manufacturing and operations, with the ERP data each question relates to. These are not actual AI outputs. See how they apply to manufacturing ERP and wholesale and distribution ERP.
Finance
Why did our margins decrease?Relevant ERP data: sales prices, supplier costs, freight and stock adjustments.
Which customers have overdue invoices?Relevant ERP data: sales invoices, receipts and customer records.
What changed in our monthly expenses?Relevant ERP data: ledger movements compared with previous periods.
Inventory and purchasing
Where are stock shortages developing?Relevant ERP data: stock levels, open sales orders and purchase orders.
Which products are underperforming?Relevant ERP data: sales by product, margins and stock movement.
Which purchasing patterns require attention?Relevant ERP data: purchase orders, supplier prices and invoice matching.
Manufacturing and operations
Which production orders used more material than planned?Relevant ERP data: bills of materials, manufacturing orders and material consumption.
Which orders are waiting to be picked or dispatched?Relevant ERP data: sales orders, picking, packing and dispatch records.
How AI works with your ERP data
- You ask a question in plain English, inside Nexa.
- AI works from the data held in Nexa: the same connected finance, sales, purchasing and inventory records your teams use. AI requests are mediated through a controlled AI gateway, with controlled prompts, source context and guarded outputs.
- You get an answer that shows the evidence and reasoning behind it, so the people reviewing it can see where it came from.
- Recommendations are reviewed by your team, who approve, edit or reject them. Where AI proposes an action, it is carried out only after a user approves it.
Illustrative example
“Why did gross margin drop this month?”
AI links the change to connected signals, for example:
- Supplier cost variance
- Freight cost increase
- Stock adjustment impact
- Delayed invoice capture
Recommended action: review top suppliers with cost increases and renegotiate affected contracts.
Illustrative example, not a real customer output.
Types of analysis
- Explaining why a figure moved
- Flagging exceptions and risks to targets
- Comparing signals across modules
- Suggesting practical next actions, with source context
Permissions and human approval
Access in Nexa is governed by role-based permission models aligned to responsibilities and approval limits. How AI access is configured for each role is agreed during scoping. AI does not act on its own: actions it proposes are carried out only after a user approves them, and your team stays responsible for the final decision.
Capabilities
What Nexa AI does today
All of the capabilities below are operational. Example questions elsewhere on this page are illustrative.
Explain and alert
- Explanations of why figures have moved
- Proactive alerts when targets are at risk
Recommend and act, with approval
- Recommendations your team approves, edits or rejects
- Approval-controlled actions: carried out only after a user approves them
Customer perspective
“What truly sets Nexa apart is its embedded AI. By simply interacting with the AI in natural language and asking questions about our own business data, we’ve been able to uncover inefficiencies, identify opportunities to reduce costs and receive intelligent recommendations to improve our processes and decision-making.”
Farhat Shaikh, Director, One Media Optical
FAQ
Nexa AI: common questions
What data does Nexa AI use?
AI answers using the business data held in your Nexa ERP, such as finance, sales, purchasing and inventory records.
Can AI make changes on its own?
No. AI can propose actions, but an action is carried out only after a user approves it. Your team remains responsible for the final decision.
Are the example questions on this page real outputs?
No. They are illustrations of the kind of question AI handles. Book a demo to see AI working with sample data.
See Nexa AI with your questions
Book a demo and bring the questions your team asks every month.