Portfolio
AI Development

DataPulse — Business Intelligence Dashboard

Featured Project
DataPulse — Business Intelligence Dashboard
72h→4h Month-End Close Time
Real-time Data Freshness
6 Source Systems Unified
4 days Board Pack Time Saved
About this Project

Meridian Capital Partners had rich data locked in 8 disconnected systems. Their executives relied on weekly analyst reports instead of real-time insights. DataPulse connects all data sources through a unified API layer, then wraps it in a natural-language interface powered by Claude AI — executives type questions like "Which clients are at churn risk this quarter?" and get back charts, tables, and written analysis in seconds.

Report preparation time dropped from 16 hours/week to zero. Board meetings now run from live data.

AI Development
Delivered by RapideKops
Project Details
Client Meridian Capital Partners
Category AI Development
Stack
Python FastAPI Claude AI React D3.js PostgreSQL dbt
The Challenge

The Problem We Solved

A PE-backed professional services firm with £45M in revenue was running its entire reporting function from a tangle of Excel workbooks, manual SQL pulls, and a 72-hour month-end close process. Department heads were making decisions on data that was 3–5 days old; the board pack took the finance team 4 full days to compile; and there was no single version of truth — finance, ops, and sales all had their own numbers.

Our Solution

How We Approached It

DataPulse is a real-time business intelligence platform built on a modern data stack: dbt for data transformation, Snowflake for the warehouse, and a custom React dashboard layer powered by a GraphQL API. It consolidates data from ERP, CRM, project management, and billing systems into a single semantic layer. Auto-refreshing board packs, role-scoped KPI views for each department, and anomaly detection alerts give every stakeholder the right data at the right time — with a 72-hour month-end close now taking 4 hours.

Key Features

Unified Semantic Layer

Single version of truth across ERP, CRM, billing, and project management — no more competing spreadsheets.

Real-Time Data Pipeline

dbt-powered transformation with incremental models refreshing every 15 minutes — decisions based on data hours old, not days.

Role-Scoped Dashboards

CFO, COO, and department heads each see the KPIs relevant to them — with drill-through from summary to transaction level.

Anomaly Detection Alerts

Statistical process control monitors key metrics and alerts the right person when variance exceeds expected bounds.

Auto-Generated Board Packs

Board packs assemble automatically from live data — formatted, branded, and emailed to stakeholders on a schedule.

Row-Level Data Security

Snowflake row-level policies ensure each user can only see the data their role permits — fully auditable.

Project Timeline

Phase 01
Data Discovery & Mapping
3 weeks
Audited all 6 source systems, mapped business logic encoded in spreadsheets, and defined the semantic layer.
Phase 02
Warehouse & dbt Build
5 weeks
Snowflake warehouse setup, dbt models for all business domains, and data quality tests implemented.
Phase 03
Dashboard & API Layer
5 weeks
React dashboard, GraphQL API, role-based access, anomaly detection, and board pack automation built.
Phase 04
Validation & Adoption
3 weeks
Data reconciliation against existing reports, user acceptance testing, and department-by-department rollout.

"We used to spend the first week of every month scrambling to reconcile numbers. DataPulse gave us our month-end in 4 hours and a board pack that generates itself. It's completely changed how we run the business."

T
Tom Reeves
CFO, DataPulse Professional Services
Let's Build Together

Have a similar project in mind?

We turn ambitious ideas into products that ship, scale, and matter. Let's talk.