Cognitive Query
Ask your city in plain language. Get grounded answers—tables, charts, maps—and trigger safe actions when you need them.
What It Is
Cognitive Query is the natural-language interface to the LaplaceX Intelligence Core. It understands domain ontologies (transport, safety, energy), maps questions to schemas and live feeds, performs hybrid retrieval (semantic + keyword + graph), and returns verifiable answers. Every response includes citations and policy checks, with optional "Act" steps routed through guardrails.
Why It Matters
Faster Insight
Turn ad-hoc questions into answers in seconds—no shadow SQL or complex query languages required.
Trust by Design
Citations, data lineage, and human-override ensure safe use in public agencies and critical operations.
From Ask → Act
Escalate to actions (alerts, tickets, signal plans) via Incident Orchestrator with full audit trails.
How It Works
Ingest
Connectors (SQL, APIs, streams, IoT)
Understand
Schema linking + city ontology
Retrieve
Hybrid search (BM25 + vector) with time/geo filters
Reason
Tool-using LLM with function calling
Answer
Charts/maps/tables + citations
Act
Safe actions behind guardrails & approval
Key Capabilities
Natural-language → SQL/Graph/Time-series
Cross-source joins & geospatial filters
Multilingual queries (EN + major locales)
Citations & data-lineage panel
Row-/column-level access control
Saved queries, subscriptions, and alerts
SDK & REST API for embedding
Audit log: question → rationale → action/override
Try It Yourself
Experience the power of natural language queries with live city data
Try Cognitive Query
Proven Outcomes
For Your City
Mobility/Parking
Predict overflow and reroute demand with natural language queries about parking patterns, traffic flows, and mobility trends.
Learn MorePublic Safety
Surface multi-agency context within 1 query, combining incident reports, camera feeds, and response resources for comprehensive situational awareness.
Learn MoreSimple Integration
Embed Cognitive Query into your existing systems with our REST API
// POST /api/cq
{
"question": "Predicted congestion on M4 corridor tomorrow 7–9am?",
"context": {
"area": "west",
"horizonHours": 24
}
}
// Response
{
"answer": "High congestion expected (85% above normal)",
"chartData": [...],
"citations": [
{
"source": "Traffic Sensors API",
"timestamp": "2024-01-15T13:45:00Z",
"reliability": "99.2%"
}
]
}Frequently Asked Questions
How do you ensure accuracy?
Grounded retrieval + citations; no answer when confidence low. Every response includes data lineage and source reliability scores.
Does it write to systems?
Only via approved actions with human-in-the-loop. All write operations require explicit approval and go through our Incident Orchestrator with full audit trails.
Data security?
Row-level access control; on-premises or VPC peering options. All queries respect existing database permissions and access controls.
See Cognitive Query on Your Data
Experience the power of natural language queries with your city's data. Get answers, insights, and actions in seconds, not hours.