CASE STUDY
AI Wine Curation System for Bistro Wines
We built an AI-assisted wine curation and order management system that analyses customer taste profiles, wine attributes, budgets, stock and previous orders to generate personalised selections in under a minute.
CASE STUDY
AI Wine Curation System for Bistro Wines
We built an AI-assisted wine curation and order management system that analyses customer taste profiles, wine attributes, budgets, stock and previous orders to generate personalised selections in under a minute.
CASE STUDY
AI Wine Curation System for Bistro Wines
We built an AI-assisted wine curation and order management system that analyses customer taste profiles, wine attributes, budgets, stock and previous orders to generate personalised selections in under a minute.
THE BRIEF
From manual wine curation to an AI-assisted selection system.
From manual wine curation to an AI-assisted selection system.
From manual wine curation to an AI-assisted selection system.
Bistro Wines needed to turn customer preferences, budgets, previous orders and live stock into accurate monthly selections without Julius spending hours checking every bottle by hand. The goal was not to replace his taste. It was to give that judgement a faster, more reliable operating system.
Bistro Wines needed to turn customer preferences, budgets, previous orders and live stock into accurate monthly selections without Julius spending hours checking every bottle by hand. The goal was not to replace his taste. It was to give that judgement a faster, more reliable operating system.
Bistro Wines needed to turn customer preferences, budgets, previous orders and live stock into accurate monthly selections without Julius spending hours checking every bottle by hand. The goal was not to replace his taste. It was to give that judgement a faster, more reliable operating system.
Client
Bistro Wines
Industry
Wine Subscription / Hospitality
Focus
AI wine curation, inventory, orders
AI wine curation, inventory, orders
AI wine curation, inventory, orders
Year
2026
THE CURATION ENGINE
THE CURATION ENGINE
The AI helps interpret the wine. The system protects the rules.
The AI helps interpret the wine. The system protects the rules.
The AI helps interpret the wine. The system protects the rules.
The product combines customer taste profiles, inventory, previous orders, budget utilisation and wine attributes. AI helps analyse label information and generate useful customer notes, while the workflow keeps stock, pricing, exclusions and human review under control.
The product combines customer taste profiles, inventory, previous orders, budget utilisation and wine attributes. AI helps analyse label information and generate useful customer notes, while the workflow keeps stock, pricing, exclusions and human review under control.
The product combines customer taste profiles, inventory, previous orders, budget utilisation and wine attributes. AI helps analyse label information and generate useful customer notes, while the workflow keeps stock, pricing, exclusions and human review under control.
01
Customer preferences, budgets, exclusions and order history are pulled into one reviewable profile.
Customer preferences, budgets, exclusions and order history are pulled into one reviewable profile.
Customer preferences, budgets, exclusions and order history are pulled into one reviewable profile.
02
The AI analyses wine attributes, taste fit, stock and margins before suggesting the right bottles.
The AI analyses wine attributes, taste fit, stock and margins before suggesting the right bottles.
The AI analyses wine attributes, taste fit, stock and margins before suggesting the right bottles.
03
Julius can review the proposed selection, adjust edge cases and keep expert judgement in control.
Julius can review the proposed selection, adjust edge cases and keep expert judgement in control.
Julius can review the proposed selection, adjust edge cases and keep expert judgement in control.
04
Final selections, picking notes and customer-ready order messages are generated in under a minute.
Final selections, picking notes and customer-ready order messages are generated in under a minute.
Final selections, picking notes and customer-ready order messages are generated in under a minute.
Five hours of manual curation became an under-one-minute batch run.
Five hours of manual curation became an under-one-minute batch run.
Five hours of manual curation became an under-one-minute batch run.
A process that used to take around 5 hours for 10 clients can now run in under 1 minute. The owner moves from building every order by hand to reviewing and approving recommendations, creating more capacity without losing the expert judgement that makes the service valuable.
A process that used to take around 5 hours for 10 clients can now run in under 1 minute. The owner moves from building every order by hand to reviewing and approving recommendations, creating more capacity without losing the expert judgement that makes the service valuable.
A process that used to take around 5 hours for 10 clients can now run in under 1 minute. The owner moves from building every order by hand to reviewing and approving recommendations, creating more capacity without losing the expert judgement that makes the service valuable.
99.7% faster for a 10-client batch
99.7% faster for a 10-client batch
99.7% faster for a 10-client batch
Five hours became under one minute for the same batch size, removing the biggest bottleneck in the workflow.
Five hours became under one minute for the same batch size, removing the biggest bottleneck in the workflow.
Five hours became under one minute for the same batch size, removing the biggest bottleneck in the workflow.
AI with human control
AI with human control
AI with human control
The system recommends, explains and drafts, while Julius can swap bottles, add extras and approve the final selection.
The system recommends, explains and drafts, while Julius can swap bottles, add extras and approve the final selection.
The system recommends, explains and drafts, while Julius can swap bottles, add extras and approve the final selection.
Ready to scale beyond one person
Ready to scale beyond one person
Ready to scale beyond one person
The workflow creates a stronger foundation for more customers, restaurant use cases and supplier conversations.
The workflow creates a stronger foundation for more customers, restaurant use cases and supplier conversations.
The workflow creates a stronger foundation for more customers, restaurant use cases and supplier conversations.