What is SparQL?
Reports, dashboards, and BI tools fail to address the real problem...
Most dashboards, BI, and visualization tools show business metrics you already know you should be monitoring. SparQL autonomously asks the questions nobody even thought to ask and surfaces answers before unforeseen problems become a crisis and hidden opportunities before they become an oversight.
Ask any question in plain English. Receive the kind of analysis that once took a senior analyst days — delivered in minutes.
Organizational Intelligence Platform
How It Works
From plain-English question to actionable intelligence
SparQL does not translate questions into SQL. It investigates — the same multi-step reasoning a skilled analyst would perform, completed autonomously in minutes.
Ask in Plain English
Type any business question — no SQL, no schema knowledge required. SparQL understands intent, not just keywords. "Why did revenue drop last quarter?" is a complete, valid query.
Works even when you don't know which schema, table, or column holds the answer.
SparQL Investigates Autonomously
Rather than guessing at a single SQL translation, SparQL iterates as many reasoning cycles as the investigation demands. Each cycle explores the schema, forms a hypothesis, tests it against real data, and refines — exactly as a senior analyst would.
Failed queries auto-correct with full error context. Real-time progress logs show every step the AI takes.
Root Causes Surface, Not Just Data
SparQL returns conclusions and recommended actions — not a table of numbers. "Revenue decline stems from a batch defect causing 135% refund spike in product line B, not a demand problem. Four recommended actions: …"
Multi-table investigations complete in minutes instead of days.
SparQL Gets Smarter Over Time
After every investigation, SparQL analyzes what it learned — ambiguous column names, hidden join paths, data-type casting patterns — and writes that knowledge back into its documentation layer.
50–70% reduction in iteration cycles after the learning phase. Every query makes the next one faster.
Platform Capabilities
Intelligence that works while you sleep
SparQL isn't a query tool. It's a proactive intelligence engine that autonomously monitors, investigates, and reports — every night, for every role in your organization.
Autonomous Investigation Engine
Fully configurable investigation depth — SparQL iterates as many reasoning cycles as your question demands, forming hypotheses and testing them against live Oracle data until root causes surface.
Morning Intelligence Briefings
Every morning before you arrive at the office, SparQL delivers a role-tailored intelligence briefing — anomalies detected, root causes investigated, actions recommended. Written like NPR, narrated by AI audio.
100% On-Premises Oracle Native
SparQL lives entirely inside your Oracle database. No ETL pipelines, no external SaaS, no data governance risk. Your data never leaves your infrastructure.
Multi-LLM Architecture
Deploy with Claude, GPT-4, GPT-5, Llama, or any compatible model. Switch providers without changing your Oracle schema. No vendor lock-in for your AI layer.
Security & Compliance Built-In
Air-gap capable, full Oracle-native audit trail, role-based access control, and read-only query execution. Data residency by architecture, not by policy.
Role-Scoped Intelligence
A CFO, a Sales Analyst, and a Facilities Manager each receive different briefings. Intelligence is scoped to hierarchy level, department, and authority — not a one-size dashboard.
Product Tour
See SparQL in action
Real screens. Real data. Every view shown here was produced by asking a plain-English question — no SQL, no configuration, no data preparation.
Your daily operations briefing — written by AI overnight, waiting when you arrive
Each morning, SparQL delivers a structured briefing scoped to your role and authority level. Situations requiring immediate attention surface first — highlighted and ranked by urgency. Emerging risks, anomalous metrics, and notable activity from your organization follow in order of priority. The right panel automatically generates ten targeted follow-up investigation questions, so the next analysis is always a single click away — no query required.
Audio Delivery
The "Play Briefing" button delivers your morning briefing as AI-narrated audio. Hands-free intelligence on your commute.
Click image to expand
Real-World Impact
Anomalies found before they become crises
These aren't hypothetical examples. SparQL surfaces findings like these every night — findings that exist in your Oracle database right now, invisible to every dashboard you own.
Identified product batch driving February sales decline — traced to supplier quality event
Flagged regional terminal anomaly before it appeared in any scheduled report
Cash flow risk surfaced automatically — 90 days past due across multiple vendors
Intelligence waiting before you arrive — every day, every role, like having analysts working overnight for you
SITUATIONS REQUIRING IMMEDIATE ATTENTION
Potential compliance exposure identified: Multiple POS orders within the last 30 days indicate minors purchased age-restricted items. Immediate review and remediation recommended.
EMERGING RISKS AND TRENDS
February sales declined approximately 30%, and early March indicators suggest the trend may be continuing. Investigation traced decline to a single product batch with 135% refund spike — supply chain quality event, not a sales problem.
SparQL found a compliance exposure we would have discovered in an audit six months later. It was waiting in our briefing before we even started the day.
Chief Compliance Officer
Regional Healthcare System
We replaced three dashboards and a weekly analyst meeting. The morning briefing tells us what matters and why.
VP of Operations
National Retail Chain
The Oracle-native architecture was the deciding factor. Data residency isn't negotiable for us.
CISO
Federal Contracting Agency
Technical Features
The intelligence layer beneath the question box
SparQL’s investigation quality comes from three compounding capabilities — each of which makes the others more effective with every query.
Your database teaches SparQL how to query it
On first contact with a schema, SparQL reads the Oracle data catalog — table names, column names, data types, existing comments, and foreign key constraints. It samples actual rows to detect real data formats: date patterns, numeric precision, hidden NVARCHAR2 columns that require CAST() before aggregation.
Outcome
Users with zero schema knowledge ask "What data do we have?" and receive a structured answer covering every schema in the database — before asking a single business question.
77% token reduction after schema documentation is built (90K → 21K tokens per query)
How it works
- Queries DBA_TABLES, DBA_TAB_COLUMNS, DBA_CONSTRAINTS for structural metadata
- Samples representative rows to infer data patterns and encoding quirks
- Reads user-uploaded schema documentation (PDF or Markdown) when available
- Weighs the pre-built catalog of 66 schemas / 7,532 tables as a starting point
Distribute intelligence automatically — on a schedule or when your data demands it
SparQL investigations can be automated and their results distributed to any recipient list — either on a recurring schedule or triggered automatically the moment a custom SQL condition is met. Finance leaders receive a weekly receivables aging report every Monday morning. Operations managers receive an instant alert the moment overdue shipments cross a threshold. No manual queries, no dashboard monitoring — intelligence delivered to the right people the moment it is relevant.
Outcome
A procurement team configures a SQL trigger: when contracts expiring within 30 days exceeds 10, automatically run the contract pipeline investigation and email results to the legal team. The first time the threshold is crossed, the report is waiting in their inbox — before anyone knew to ask.
One-time, daily, weekly, and monthly schedules — plus event-driven Custom SQL Triggers with configurable check frequency
How it works
- Schedule any NLQ investigation to run on a one-time, daily, weekly, or monthly basis — full results delivered automatically by email
- Define a Custom SQL Trigger with a threshold condition: when your SQL query returns a value that breaches the threshold, SparQL fires the investigation automatically
- Set the check frequency for SQL triggers — hourly or custom interval — so time-sensitive conditions are caught as soon as they occur
- Each distributed report includes the complete investigation output: written analysis, visualization charts, and paginated data tables
- Manage multiple active schedules simultaneously — each scoped to a different question, trigger condition, and recipient group
Choose the right model for the right task — or let SparQL decide
SparQL is not locked to a single AI provider. Every investigation session specifies the model at query time. Organizations can select based on complexity, cost, compliance requirements, or preference.
Outcome
Federal agencies with strict data-egress restrictions can run Llama locally on their own infrastructure. No query result, no schema detail, and no row of data ever leaves the network boundary.
Supports any model accessible via an OpenAI-compatible API endpoint
Supported models
claude-sonnet-4-6
Recommended — best quality/cost balance for most investigations
claude-haiku-4-5-20251001
Fastest and lowest cost — ideal for simple lookups and high-volume scheduled reports
claude-opus-4-6
Highest capability — complex multi-schema investigations and adversarial analysis
GPT-4 / GPT-4o / GPT-5
Available for organizations with existing OpenAI agreements or compliance constraints
Local Llama (self-hosted)
Air-gapped and fully on-premises — no external API calls, no data egress of any kind
Universal Compatibility
Works with any Oracle data — any ERP System, Data Warehouse, Custom Application/Data Model
SparQL's investigation engine adapts to any schema automatically. Whether you're running Oracle Financials, a retail data warehouse, a custom operational database, or all three — SparQL queries them in plain English without configuration.
Works With Any Oracle Schema
SparQL adapts to your data — ERP, data warehouse, custom operational database, or mixed environments. No per-schema configuration required.
Self-Discovers Joins & Relationships
Undocumented foreign keys, cross-schema joins, implicit lookup tables — SparQL finds and validates them automatically through iterative test queries.
Zero Schema Knowledge Required
Business users ask questions in plain English. SparQL identifies the relevant tables, columns, and relationships autonomously — without help from a DBA.
Full Audit Trail
Every investigation, briefing, and query logged to Oracle-native SPARQL tables. Complete chain of custody, queryable from any tool that speaks SQL.
Role-Scoped to Your Org Hierarchy
Intelligence scales to authority level. A CFO, a department manager, and a line analyst each receive briefings appropriate to their scope and role.
100% On-Premises
All logic lives in PL/SQL inside your Oracle database. No cloud sync, no data egress, no new infrastructure. Air-gap capable for classified environments.
Supported Data Environments
Oracle ERP
Financial Management
Data Warehouses
Analytics & Reporting
Retail & POS
Point of Sale
Healthcare EMR
Clinical Data
HR Systems
Workforce Data
Custom Schemas
Any Oracle Data
All logic lives in PL/SQL packages inside your Oracle database. No microservices, no external pipelines, no new infrastructure to secure, monitor, or patch.
Industry Use Cases
Built for every Oracle environment
Whether you're running Oracle Financials, a retail data warehouse, a healthcare database, or a custom operational system — SparQL adapts automatically. No connectors. No data movement. Query the live Oracle tables in plain English.
Plain-English answers from your Oracle Financials data
AP aging, cash flow, budget vs. actuals — without writing SQL.
Finance teams spend hours extracting insights from Oracle ERP — AP, AR, GL, project accounting, and procurement tables. SparQL sits directly on your Oracle Financials schema and answers complex financial questions in minutes, without SQL, without ETL, and without the data ever leaving your environment.
Schema coverage
SparQL adapts to standard Oracle ERP schemas (AP_INVOICES, AR_CUSTOMERS, GL_JE_LINES, PO_HEADERS, PA_PROJECTS) as well as custom extensions. Self-learning documentation builds over time — each query makes the next one faster and more accurate.
Example questions your team can ask — today
- ›“Which vendors have outstanding invoices over $50,000 that are more than 90 days past due?”
- ›“Show budget vs. actuals variance by department for Q1 — flag any department over 15% over budget.”
- ›“Which purchase orders were approved this month that exceed the department spending limit?”
- ›“What is our accounts receivable aging summary for the top 20 customers by open balance?”
- ›“Compare this quarter's operating expenses against the same quarter last year, broken down by GL category.”
Anomaly detection across supply chain, inventory, and POS data
Spot problems before they appear in any scheduled report.
Operations teams managing retail, distribution, or supply chain data often discover problems too late — in weekly reports, quarterly reviews, or audits. SparQL monitors your operational Oracle database overnight and surfaces anomalies proactively, before the business day starts.
Overnight monitoring
SparQL's DBMS_SCHEDULER jobs run overnight across your operational data. Anomalies are investigated automatically — root causes traced, findings written up in plain English — and waiting for the right people before they arrive at the office. No dashboard, no manual query, no meeting required.
Example questions your team can ask — today
- ›“Which stores had void rates more than 2 standard deviations above their 30-day average yesterday?”
- ›“Show me products where refund rate exceeded 10% in the last 7 days — group by SKU and region.”
- ›“Which suppliers have open purchase orders past their committed delivery date by more than 14 days?”
- ›“Flag any inventory locations where on-hand quantity dropped below safety stock threshold this week.”
- ›“Compare weekend revenue per location against the prior 4 weekends — identify underperformers.”
Why organizations choose SparQL
100% On-Premises
SparQL is a PL/SQL package deployed inside your Oracle database. No SaaS tier, no cloud sync, no data egress. Your data stays where it belongs.
Zero Schema Configuration
SparQL reads your Oracle data catalog and samples representative rows on first contact. No manual column mapping, no ETL setup, no integration work before the first query.
Self-Learning Documentation
After every investigation, SparQL writes what it learned — ambiguous columns, join paths, data-type quirks — back into its documentation layer. 50–70% fewer LLM iterations over time.
Oracle-Native Audit Trail
Every query, iteration, generated SQL, and model response is logged in SPARQL_LOGS and SPARQL_SESSIONS — standard Oracle tables, visible to your DBA, queryable in any audit tool that speaks SQL.
Integration
Call SparQL from any Oracle context
SparQL is a unique PL/SQL package, installed directly into the Oracle codebase. No complicated installation required. If you've ever run a SQL script, you already know how to install SparQL.
Patterns
About this pattern
The simplest integration: call SPARQL.COORDINATOR from any PL/SQL context — SQL*Plus, a scheduled job, a stored procedure, or a third-party tool. The call blocks until the investigation completes and returns the final answer and generated SQL.
Best for scripts, testing, and short queries. Complex investigations can take 2–10 minutes; for production use the async pattern below.
Session statuses
-- Connect via SQL*Plus or any Oracle client
-- SET SERVEROUTPUT ON SIZE UNLIMITED
DECLARE
v_session_id VARCHAR2(100) := NULL; -- auto-generated if NULL
v_answer CLOB;
v_sql CLOB;
BEGIN
SPARQL.COORDINATOR(
p_nlq => 'Show IIJA infrastructure spending by state for FY2025',
p_model => 'claude-sonnet-4-6',
p_session_id => v_session_id, -- IN OUT: populated on return
p_final_answer => v_answer,
p_final_sql => v_sql
);
DBMS_OUTPUT.PUT_LINE('Session: ' || v_session_id);
DBMS_OUTPUT.PUT_LINE('Answer: ' || SUBSTR(v_answer, 1, 4000));
DBMS_OUTPUT.PUT_LINE('SQL: ' || SUBSTR(v_sql, 1, 4000));
END;
/Package
SPARQL.COORDINATOR
Version
5.0.1
Prerequisite
Oracle19c
Get Started
See SparQL in your Oracle environment
Schedule a live demonstration with real Oracle data. See morning briefings, live investigation, and the role-scoped intelligence delivery that replaces your existing dashboards.
Sicore Consulting LLC · Enterprise Natural Language Querying for Oracle