We use cookies

We use cookies to ensure you get the best experience on our website. For more information on how we use cookies, please see our cookie policy.

By clicking Accept, you agree to our use of cookies.
Learn more.

Company logo
Customer Success Story
Navlungo logo

From Idea to Insight in 30 Seconds

Navlungo is Turkey's cross-border logistics marketplace connecting e-commerce sellers with 100+ carriers across 130+ countries. Thousands of daily shipments. Multiple operational systems. One recurring problem: answering business questions meant hours of manual work.

“Which carrier has the best delivery rate to Wonderland?” meant querying the shipment database, pulling tracking logs, checking pricing records, cross-referencing customer data—then reconciling mismatched carrier codes, country names, and currency conversions by hand. Half a working day from question to answer.

AI agent + clean data = 30 seconds from question to decision
720x
Faster insights
4hr → 30sec
5+
Data sources unified
<30s
Idea to insight
Real-time
Self-serve access
Before Datazone
~4hr
Manual analysis per question
Query multiple systems
Export to spreadsheets
Reconcile data by hand
Wait for analyst availability
After Datazone
<30s
Idea to insight, every time
Type natural language question
Agent queries clean fact table
Get structured, accurate answer
Make decision instantly

The Problem

Before: Manual Cross-System Analysis
Question: “Which carrier has the best delivery rate to Germany this month?”
Step 1: Pull shipment data (1 hour)
Export from shipment database, filter by destination
Step 2: Query CRM records (1 hour)
Get customer complaints, match carrier codes manually
Step 3: Check ERP cost data (30 min)
Pull pricing records, reconcile date formats
Step 4: Reconcile everything (1.5 hours)
Fix mismatched carrier codes, country names, date formats
Total time: ~4 hours per question
Operations team fielded 3-5 of these daily

What Datazone Built

1. Single Operational Fact Table

Shipment data, tracking checkpoints, pricing records, customer segments—all joined into one wide fact table. One row per shipment with everything needed for analysis.

Automated joins upstream
Complex multi-system joins handled in ETL pipeline. Dashboard and agent query clean, pre-joined data.
Normalized metrics
Business-specific calculations computed once. Multi-currency revenue normalized to base currency. Provider codes standardized across systems.
Business logic in data layer
Lead times, performance rates, delivery estimates—all pre-computed. No interpretation errors.

2. AI Agent + Dashboard on Same Data

Agent and dashboard both query the same fact table. Agent converts natural language to SQL, applies metric definitions, returns structured answers. Dashboard shows pre-built KPIs with filters.

Operations team asks ad-hoc questions via agent. C-level monitors trends in dashboard. Same underlying data. Same metric definitions. Instant answers.
Operations AI Agent
Which carrier has the best performance to Wonderland this month?
Based on current month data across 1,850 shipments to Wonderland:
1. Acme Express
On-time rate: 94.1% | Avg. transit: 2.3 days | 845 shipments
2. Swift Global
On-time rate: 91.7% | Avg. transit: 2.7 days | 625 shipments
3. Rocket Logistics
On-time rate: 88.4% | Avg. transit: 3.1 days | 380 shipments
Query: 1.8s | Total time: 18s
What about cost per shipment?
Average cost per shipment to Wonderland (current month):
Acme Express$24.50
Swift Global$22.80
Rocket Logistics$26.90
Query: 0.9s | Total time: 11s
What Changed
Same question. 30 seconds from idea to insight instead of 4 hours. Type question, get answer, make decision. No analyst. No spreadsheets. No waiting.

How It Works

Data Pipelines

Three core pipelines: business metric calculation, multi-currency normalization, and main operational fact table. Complex joins handled upstream. Agent and dashboard query pre-joined, clean data.

Metric Definitions

KPIs defined once in data layer. Business metrics mean the same thing to agent and dashboard. Performance rates calculated consistently. No interpretation errors.

Natural Language Interface

Agent converts questions to SQL against fact table. Understands logistics domain terms. Operations team types questions, gets answers. 30 seconds from idea to insight.

What Changed

Before
  • ×4 hours per business question
  • ×Manual cross-system reconciliation
  • ×Analyst bottleneck for every query
  • ×C-level waiting for answers
After
  • ✓30 seconds per question
  • ✓Automated data harmonization
  • ✓Ops & C-level self-serve
  • ✓720x faster insights

Navlungo's operations team goes from question to decision in under 30 seconds. Carrier performance, pricing trends, fulfillment lead times—type the question, get the answer, make the call. No analyst bottleneck. No spreadsheet reconciliation. No waiting half a day. Same data that powers their dashboard, accessible through natural language.