This is some text inside of a div block.

From 1,600 to 8,500+ active restaurants in 15 months with Nekt

About

Brendi lets restaurants run their own WhatsApp delivery channels instead of paying marketplace commissions. To scale, they had to unify data that lived in four disconnected systems.

Company size

1-100

Industry

Foodtech

“Thanks for solving our pain of centralizing data so well, in a way that’s simple and accessible to everyone in the company.”

Daniel Frageri
CEO

The challenge

Brendi’s data was fragmented across Firestore, MongoDB, Supabase, and BigQuery: dozens of collections, billing systems, campaign engines, and bot conversations, all disconnected. There was no single view to run the business on.

What they built

  • 84 pipelines turning raw JSON documents into a clean dimensional model
  • An AI bot analytics engine tracking conversation state, intervention confidence, and post-order analysis
  • A dual marketing engine separating scheduled from AI-driven campaigns, with attribution across WhatsApp touchpoints
  • Reverse ETL pushing curated data back into Supabase and AlloyDB, turning the lakehouse into an operational hub
  • AI observability for bot-to-human escalations, with confidence scores and explanations
  • 111 active pipelines today, feeding everything from abandoned-cart recovery to ban-prediction models

What they built