Featured Project · March 2026

SAP × AI
Query Intelligence

Natural language to SAP HANA SQL in ~5 seconds. A tool I built that eliminates the 2–5 day ad-hoc reporting cycle — no ticket, no waiting, no guesswork.

Python 3.12 Groq AI · Llama 3.3 70B Flask SAP HANA SAP Datasphere

From days to seconds

Every time a business user needed an ad-hoc report, they raised a ticket, waited for IT, reviewed drafts, and repeated. SAP Query AI removes that entire loop.

✕  Before — Manual Process
  • Business user needs a report
  • Raises a ticket to IT
  • Developer writes SQL manually
  • Review → test → send back
2–5 days
per ad-hoc request
✓  With SAP Query AI
  • Type question in plain English
  • AI generates SAP HANA SQL
  • View results + explanation
  • Copy SQL → run in HANA Studio
~5 seconds
from question to query

End-to-end Architecture

A four-layer stack — Consumption, Engine, Metadata, and Source Systems — designed to translate plain English into validated SAP HANA SQL with schema-aware context injection.

SAP Query AI Architecture Diagram — four-layer system design

See it in action

A walkthrough from plain-English question to generated HANA SQL — with live query results and a plain-language explanation of the logic.

Tech Stack

Language
Python 3.12
Backend & CLI · Free
Framework
Flask
REST API · Web UI · Free
AI Model
Llama 3.3 70B
via Groq AI · Free tier
Database
SAP HANA
Target query engine
Data Layer
Datasphere
Virtual data models

SAP Data Models Supported

Covers the full SAP data landscape — from native HANA tables to Datasphere analytic models and BW/4HANA providers.

Platform Data Objects Status
SAP HANA Native Tables, Calc Views Built
SAP Datasphere Analytic Models, Views Built
SAP S/4HANA CDS Views (I_, C_, Z_) Ready
SAP BW/4HANA CompositeProviders, ADSOs Ready
BW[4] PCE Federation via Datasphere Ready
BEx Queries MDX generation Roadmap
SAP Analytics Cloud Stories, Models Roadmap

The Roadmap

Live HANA Connection

Direct connection to SAP HANA databases — execute queries and return live results in-tool.

BEx Query MDX

Support for legacy BW MDX generation, enabling natural language over BEx reports.

Datasphere Auto-Views

Generate Datasphere graphical views from natural language descriptions — design as you describe.

Multi-language Support

German, Hindi, and beyond — making data accessible in the language of the business user.

Role-based Access

Enterprise-ready query history, role permissions, and audit trail for governed environments.

Curious to see
more in action?

Whether you're an SAP team looking to reduce reporting overhead, a consultant exploring AI add-ons, or just want a live walkthrough — I'd love to show you more.