# Reconciling hundreds of partner restaurants across public data
> A global beverage company wanted a directory of its partner restaurants, but four sources disagreed on names, addresses, hours and cuisine across more than ten languages. Twistag built a pipeline on Google Cloud that cleans the data with conventional ETL and uses OpenAI only for judgement: matching listings, resolving conflicts field by field and writing profiles in the brand's voice. Hundreds of restaurants were reconciled in three months, and new markets run on the same pipeline.

Client: Global beverage company · Industry: Retail & consumer · Practices: AI & Agents, Data & Cloud · Timeline: 3 months (phase one)
## Key results
- **100s** partner restaurants reconciled
- **10+** languages indexed
- **3 months** to a fully automated, repeatable pipeline
## The challenge

**Same restaurant, four different records.**

Twistag built an AI data pipeline for a leading global beverage company that reconciled hundreds of European partner restaurants across four conflicting data sources and more than ten languages into a single directory, fully automated and repeatable within three months. Records on TripAdvisor, TheFork, Google Reviews and venue websites disagreed on names, addresses, opening hours and cuisine categories, and deciding whether two listings describe the same physical restaurant is a judgement call. Twistag used conventional ETL for normalisation and deduplication, and reserved OpenAI calls for the judgement work: entity matching, field-by-field conflict resolution, and description synthesis in the brand's voice.

A global beverage company's on-premise presence is its restaurant network — the venues where consumers encounter the product in context. The company wanted that network navigable: a directory where someone could find partner restaurants by city, cuisine or dining style. The data made that harder than it sounds. Information for hundreds of venues sat scattered across TripAdvisor, TheFork, Google Reviews and individual restaurant websites, and no two sources agreed on anything — names, opening hours, cuisine categories and addresses all conflicted, often across more than ten European languages. "Ristorante Da Mario" on TripAdvisor, "Da Mario Restaurant" on Google, "da mario" on the restaurant's own site: sometimes the same place, sometimes not. Manually verifying each restaurant would have required a team of researchers working for months on the first market alone. The client is not named publicly.

## The approach

**AI only where judgement is needed.**

1. **Collect from every source.** Cloud Run workers gather TripAdvisor, TheFork, venue websites and Google Search results, with Pub/Sub between stages. A new regional source is a new container. 
2. **Clean with conventional ETL.** Addresses, cuisine categories and languages are standardised, obvious duplicates removed, and each field scored for confidence by source, recency and agreement. 
3. **Reconcile with judgement.** OpenAI decides whether listings describe the same place, takes each field from its most reliable source and writes a coherent profile. Every decision has explicit inputs, so it can be audited. 

## The impact

**A pipeline that travels.**

The first phase covered hundreds of partner restaurants across multiple European markets in three months, indexed across more than ten languages and reconciled across four major data sources. Each profile is the output of AI-driven reconciliation across at least two independent sources, with confidence scores on every field showing the team where manual review is worth spending. The more durable result is the architecture: adding a new European market means deploying scrapers for any region-specific sources and running the pipeline unchanged. Twistag has not published a figure for directory usage, partner engagement or verified coverage accuracy.

> Phase two and beyond are operational expansions.

## Technologies used

- OpenAI
- GCP
- Puppeteer

> Twistag built an AI data pipeline reconciling hundreds of European partner restaurants across four conflicting sources into a single directory.

---

Source: https://twistag.com/case-studies/ai-restaurant-directory-european-beverage-brand