Resolon Start a conversation

InsightsAI-native ERP

Published Updated 5 min read

Eight questions to ask your ERP vendor about AI

Ask these of the vendor you are considering, and of the ERP you run today. The answers show where the AI works and who stays in control.

Chad Neptune, CEO and Chief AI Architect

Contents

Ask these eight questions before you buy AI from your ERP vendor, and ask them of the ERP you run today. The answers show whether the AI works on live records, who approves what it does, and whether you can leave with your data.

The label “AI-native” matters less than the answers. Three tests matter more: does the AI run inside the workflow on live records, can it act with a person approving, and does the data still pay off if you switch the AI off?

Three tests before the eight questions

The first three questions below are these tests. The rest check what happens between the AI’s draft and your books, and what you keep if you leave.

A weak answerA good answer
Inside the workflow on live recordsIt works on a copy exported beside the systemThe AI works on the same records your people use
Can it act, and who approvesIt acts with no named approver, or nobody can say who approvesIt can act, and a named person approves each decision
The data work if the AI is switched offThe data work only makes sense if the AI staysAgreed definitions and a named owner for each field are useful either way
Figure 1. The three tests, and how to read the answer

For leaders Put the three tests to every vendor in writing, and to the ERP you run today.

Colleagues talk on a balcony inside a bright glass office atrium

The eight questions

Ask these of the vendor you are considering, and of the ERP you run today. The answers show where the AI works and who stays in control.

  1. Does the AI run inside the workflow on live records, or beside it on an export? Ask whether the AI works on the same records your people use, or on a copy exported beside the system. An assistant on an export can answer questions but cannot do the work inside the steps.
  2. Can it act, or only answer? Who approves what it does? Some tools only answer. Others act. If it can act, ask who approves what it does before it takes effect. Our baseline is a named person approving each decision.
  3. If you switched the AI off tomorrow, would the data work still pay off? The data work should pay off whether or not the AI stays. Agreed definitions and a named owner for each field are useful either way.
  4. Does an AI-drafted record wait for approval before it reaches the live books? An AI-drafted record should wait for approval before it reaches the live books. Ask where the draft sits, who sees it and what moves it forward.
  5. Does the log show who or what acted, when, why, on which model, and who approved? The log should show who or what acted, when, why, on which model, and who approved. If an entry cannot be traced, it cannot be reversed.
  6. What happens if the data changes between the AI’s draft and the approval? A draft is made at one moment and approved at another. Ask what the approver sees if the record changes in between.
  7. Can you widen what it does in steps: suggest, queue for approval, act alone under a limit? Ask whether you can start with suggestions, move to queuing work for approval, and let the AI act alone only under a limit you set.
  8. If you leave, can you take the ledger, the history and the AI steps with you? Ask whether you can take the ledger, the history and the AI steps with you if you change vendors. Systems you own are the aim.

If an entry cannot be traced, it cannot be reversed.

For leaders Ask for the answers in writing, and ask the same eight of the ERP you run today.

Weighing AI from your ERP vendor? Talk to a principal.

Four more worth asking

What is live today, and what is on a roadmap? How is the AI priced? What data reaches the model, and may the provider keep it? Can we test with our own data before we sign?

For leaders Do not sign until you have tested with your own data.

What to do this quarter

Before you renew or buy, put the eight questions to the vendor and to the ERP you run today.

  1. Ask all eight in writing.

    Compare the answers side by side, vendor against vendor and against what you run now.

  2. Run the three tests first.

    Live records, a person approving, and data work that pays off with the AI switched off.

  3. Check the trail.

    Look at where an AI draft sits, who sees it, and what the log records.

  4. Confirm the way out.

    Ask what you can take with you: the ledger, the history and the AI steps.

Frequently asked questions

What should I ask my ERP vendor about AI?

Ask eight questions. Does the AI run on live records? Who approves what it does? Does the data work pay off without the AI? Does a draft wait for approval before it reaches the live books? What does the log show? What happens if the data changes before approval? Can you widen what it does in steps? Can you leave with your data?

What makes an ERP “AI-native”?

The label matters less than the answers. Three tests matter more: does the AI run inside the workflow on live records, can it act with a person approving, and does the data still pay off if you switch the AI off?

Should AI be allowed to act on live ERP records?

It can, with a person approving. If it can act, ask who approves what it does before it takes effect. Our baseline is a named person approving each decision, and an AI-drafted record should wait for approval before it reaches the live books.

Can I keep my data if I change ERP vendors?

Ask before you sign. Whether you can take the ledger, the history and the AI steps with you is one of the eight questions, and systems you own are the aim.

Do I need to replace my ERP to use AI?

Not necessarily. Resolon resells no ERP and redesigns the work on the ERP you run, with AI inside each step and a person approving what matters. If a replacement would serve you better, we will say so.

Share this article

New each month. Follow by RSS

Senior-led, start to finish.

A senior Resolon lead is accountable for every engagement, from the first conversation to handover, with a team built for the work. For an active engagement or a new requirement, start a conversation.