AI-Powered Analytics Dashboard

Real-time insights and natural-language answers. Unified sales, marketing, and operations data into a single dashboard with "ask your data" queries and anomaly alerts.

8 weeks delivery

Omnichannel Retail (PT/ES)

−35% reporting time

The Challenge

An omnichannel retail company operating across Portugal and Spain had data scattered across multiple systems: ERP, POS terminals, advertising platforms, and e-commerce. Leadership was making critical decisions based on manual reports that were often outdated by days.

They needed a unified analytics platform that could aggregate data from all sources, provide real-time insights, and enable teams to ask questions in natural language without waiting for data analysts.

Fragmented data sources

Manual

Slow reporting process

Days

Decision latency

Our Solution

We built a central data lake with a semantic layer that connects all data sources and enables natural language queries. The platform processes data in real-time, generates automated alerts for key metrics like ROAS and stockouts, and provides an intuitive dashboard for business users.

Technology Stack

Data & Backend

BigQuery/ClickHouse for data warehouse
dbt for data transformation
Airbyte for data ingestion
FastAPI + Postgres + pgvector
OpenAI/Anthropic for NLQ

Frontend & Orchestration

Next.js 15 for dashboard UI
LangGraph for agent workflows
Supabase for auth & storage
Grafana for monitoring
n8n for workflow automation

Natural Language Queries

"Ask your data" interface powered by AI that understands business questions and generates accurate SQL queries.

Automated Alerts

Smart anomaly detection for KPIs, ROAS thresholds, and stockout predictions with instant notifications.

Unified Data Lake

Central repository connecting ERP, POS, ads platforms, and e-commerce with semantic layer for consistency.

Real-Time KPIs

Live dashboards tracking sales, marketing ROAS, inventory levels, and operational metrics across all channels.

The Results

The unified analytics platform delivered measurable improvements across operations, marketing, and sales within the first 8 weeks of deployment, transforming how the retail organization makes data-driven decisions.

−35%

Reporting Time

From days to real-time

+18%

Paid Media Margin

Improved ROAS optimization

−22%

Stockouts

Better inventory predictions

"AUTNEX unified our fragmented data landscape and gave us the tools to ask questions we couldn't answer before. The natural language queries have democratized data access across our team, and the automated alerts have prevented countless stockouts and wasted ad spend."

Miguel Rodrigues

Chief Operations Officer, Retail Group

The Process

W1-2

Discovery & Data Mapping

Audited all data sources (ERP, POS, ads, e-commerce), mapped data schemas, and defined KPIs and alerting logic.

Data Source Audit KPI Definition Architecture Design

W3-5

MVP Development

Built data lake, semantic layer, NLQ interface, and core dashboard with initial connectors and alert logic.

Data Lake Setup NLQ Engine Dashboard MVP

W6-8

Hardening & Launch

Refined models, optimized queries, added monitoring, trained users, and deployed to production with full documentation.

Testing & Optimization User Training Production Launch