How NgageChat works

Every step on screen: what it read, where, and what came back.

From a sentence to an ERP record, in four stages

Every stage shows on screen while it runs.

NgageChat 4 steps

Raise a requisition for 20 seals for pump P-114, needed by Friday.

  1. 1 Read the requestA purchase requisition, for a part on P-114 NgageChat
  2. 2 Find the asset and its partP-114 at site 100, seal 6205 IFS
  3. 3 Read the defaults and the real choicesPlant 1000 by default, one question for you SAP
  4. 4 Draft the purchase requisition20 seals for Friday, waiting for your yes SAP
Create the purchase requisition in SAP?
Illustrative example on a test landscape: IFS for assets, SAP for purchasing. Values are samples.
  1. 1

    You ask

    Type it the way you would say it. There is no screen to find and no field name to learn.

  2. 2

    It finds the record

    NgageChat already knows how IBM Maximo, IFS and SAP are structured, and finds the records the request needs.

  3. 3

    It fills the blanks with you

    Defaults come from the ERP. Anything missing comes back as a short question with your system’s real choices.

  4. 4

    You approve, it writes

    It shows what it is about to write, and you choose which writes wait for your yes.

How it connects to each ERP

IBM Maximo, IFS and SAP work out of the box. For any other application, including a homegrown one, we teach Ngage how it works in weeks.

ERPReads and writes throughRuns within
IBM Maximo Maximo’s official interfaces Your security groups
IFS OData projections to read, official IFS actions to write Your permission sets
SAP SAP’s official interfaces Your authorisations
Any other application Homegrown or specialist software. We teach Ngage how it works in weeks.

Nothing is installed in IFS, and on all three the database is never touched.

It never writes a value from memory

Allowed values come from your live ERP. If none fits, it says so. On IFS, these details decide whether a write succeeds.

Defaults
From IFS’s own _Default() function
Concurrency
Updates carry the record’s ETag
Actions
Bound actions and action imports, each used correctly
Field values
Enumerations, then lists of values, then free text

See the IFS integration

It stays inside your own ERP permissions

The ERP still enforces segregation of duties.

  • Signed in as you

    Your own ERP permissions, never a shared service account.

  • You choose what waits

    Writes can wait for your approval.

  • Every step on record

    Who asked, what was planned, what ran, what came back.

  • Figures checked

    1,000 and above, against what it read. A safeguard, not a guarantee.

Questions about how it works

Read all the questions

How does NgageChat connect to IBM Maximo, IFS and SAP?

Through each ERP’s official interfaces, and never the database. On IFS, reads use OData projections and writes call official IFS actions, and nothing is installed in IFS. On all three, every request runs under the signed-in person’s own permissions.

Can one request use two ERPs?

Yes. One request can work across two ERPs, for example the asset in IFS and the cost in SAP.

Which systems does it cover?

IBM Maximo, IFS and SAP work out of the box. On IFS, that is IFS Cloud and IFS Applications 10. For any other application, including a homegrown one, we teach Ngage how it works in weeks. NgageChat also answers from your knowledge bases, documents and policies.

Does it need a copy of our data?

No. Every question is answered by reading the ERP live, at the moment you ask. There is no replica, no warehouse and no nightly extract to go stale.

Why does it ask me questions back?

Because guessing is worse. If a required field is missing, or several records match the name you used, NgageChat asks. The options it offers come from your own system.

How long does an answer take?

The first step usually shows within seconds. Quick lookups take seconds, deep analyses take minutes, and every step shows as it runs.

Can I see what it did?

Yes, while it runs and afterwards. Each step shows on screen as it happens, and every step is written to an audit trail: who asked, what was planned, what ran and what came back.

Which AI model does it use?

NgageChat is built to work with more than one model. Enterprise deployments run on Google Cloud Vertex AI with zero data retention configured, in the region you choose.

Send us a question from your own ERP

We run it with you, live, on a test system, and you see every step.