What challenges did Filum face before AI implementation?

Fragmented customer data and hundreds of client integrations became unscalable with manual Python scripts and spreadsheets. The team lacked DevOps engineers to manage infrastructure and operations at scale.

What was the impact of the AI solution?

Reduced infrastructure maintenance to zero. Two engineers now manage 100+ integrations. Enabled real-time high-velocity data handling. Reduced operating overhead and helped support low churn for key retail clients.

Source

GCPScaled productionEvidence: Medium55/100

Filum | Zero-Ops AI engine using BigQuery, Cloud Run, Gemini/Vertex AI

Use case typeMarketing analyticsUpdated Jun 13, 2026

Filum is an AI-powered Unified Customer Experience Platform that helps enterprises transform fragmented customer data into actionable business intelligence for customer loyalty and growth. The company originally managed pipelines with manual Python scripts and spreadsheets, but as it onboarded large retailers such as ZUS Coffee and GS25, the volume and number of integrations became unscalable. To address the lack of in-house DevOps capacity, Filum moved to a fully serverless Google Cloud architecture using BigQuery for schema-less ingestion and Cloud Run / Cloud Run functions for isolated client integrations that auto-scale in real time. The customer story says Filum also uses Gemini for instant customer insights and plans to integrate Vertex AI into internal workflows as part of an AI-first product strategy.

Organization: Filum
Industry: Retail
Location: Singapore
Published: May 2026

Reported outcomes

Strategic outcomes

Scale & capacityManaged integrations with small teamSpeed & agilityEnabled real-time data handlingCost efficiencyEliminated infrastructure maintenanceCustomer experience & trustSupported lower customer churn

Why do we believe this deployment?Customer identity, provider attribution, maturity, and source checks

Customer: Filum
Provider: GCP
Maturity: Scaled Production
Linked source: Google Cloud Customer Story

The team lacked DevOps engineers to manage infrastructure and operations at scale

Customer identity supportedSource describes one deploymentMaturity supported

Primary read

Use case focus

Showing 3 of 4

1Serverless Integration Platform
2Customer 360
3Operational Automation

Fragmented customer data and hundreds of client integrations became unscalable with manual Python scripts and spreadsheets.
The team lacked DevOps engineers to manage infrastructure and operations at scale.

Adopted BigQuery for schema-less data ingestion without complex provisioning.
Moved client integrations to Cloud Run and Cloud Run functions so each integration runs as a standalone serverless function.
Used orchestration to automate infrastructure management and generate customer insights.
Uses Gemini for instant customer insights and has a planned Vertex AI integration for internal workflows.

Technologies

BigQuery Cloud Run Cloud Run functions Dataflow Pub/Sub Looker Studio Gemini Vertex AI

Reduced infrastructure maintenance to zero.
Two engineers now manage 100+ integrations.
Enabled real-time high-velocity data handling.
Reduced operating overhead and helped support low churn for key retail clients.

Architecture

A fully serverless Google Cloud stack built around BigQuery, Cloud Run, Cloud Run functions, Dataflow, Pub/Sub, Looker Studio, and Gemini, with planned Vertex AI integration for internal workflows.

Sources & evidence1

Evidence: Medium55/100Evidence strength