AI analytics for Oracle EBS — ask analytical questions in plain English

[DuskoDelic]

Hey everyone,

Long-time lurker, first time posting about something we built.

We got tired of the same thing you're probably tired of: ad-hoc data
requests that require custom SQL or report building. "What's our AP
aging by supplier?" shouldn't require a DBA request that comes back
tomorrow.

So we built an AI assistant that connects to your EBS 12.2 database
and answers analytical questions in plain English. It runs as a Docker
container on your infrastructure — your data never leaves your network.

The real value isn't simple lookups (though it does those too). It's
the analytical questions your UI was never designed to answer:

- "Show me AP aging over 90 days by supplier" — cross-table query,
structured answer
- "What's our on-hand inventory value by organization?" — multi-org
aggregation
- "PO spend by category this quarter" — procurement analytics
- "GL trial balance for March across all ledgers" — finance summary

Questions that normally require SQL expertise, a custom report, or a
request to the DBA team — answered in seconds.

Before anyone asks (because I would):
- Read-only SQL only. SELECT statements. Never INSERT, UPDATE, DELETE.
- Full audit log. Every query logged — SQL text, timestamp, user,
row count. You can see exactly what it did.
- RBAC enforced. Users log in with their EBS credentials. They only
see data for their operating units.
- Air-gapped mode. Run with a local LLM and zero external network
calls if you want.

It also works with PeopleSoft HCM and IBM Maximo if you run those.

You can try it yourself right now — self-service demo, no install,
no signup: interpretos.ai/demo

Or if you want to run it yourself:

docker pull interpretos/local

Setup wizard walks you through SSH/DB connection. Should be up in
about 5 minutes.

Free 30-day evaluation. After that it's flat monthly pricing
($2K/$5K/$10K depending on tier — no per-user fees). 40% early
adopter discount before August 1.

Happy to answer questions. Would especially appreciate anyone who
tries the demo or runs it on a dev instance and reports back.

interpretos.ai