B&Q et Castorama (Kingfisher)
personalized recommendations plus a genAI conversational DIY assistant
At Kingfisher, AI-driven recommendations and personalization generated about 80m GBP in sales in the first half of 2025/26, up about 37% year over year, and the Hello Casto DIY assistant has about 60,000 users per month.
Key points
- Personalized recommendations and the conversational DIY assistants Hello Casto and Hello B&Q.
- In-house Athena orchestration on Google Cloud Vertex AI and the Nucleus data lake.
- About 80m GBP in AI sales in H1 2025/26, +37% year over year.
- Hello Casto has about 60,000 monthly users, evidence level A confirmed.
Objective
Grow online sales by pushing the right product recommendations and answering DIY questions in natural language, in order to turn a technical catalog into a guided journey.
The deployment
Kingfisher built Athena, an in-house orchestration framework that connects several language models to its retail brands. On this foundation, the group launched Hello Casto, its first conversational genAI DIY assistant, on Castorama France on November 14, 2023, then Hello B&Q in the United Kingdom. The customer asks a DIY question in natural language and receives step-by-step advice and suitable products. In parallel, recommendation and personalization engines feed the product pages. According to the group's 2025/26 half-year results, AI-driven recommendations and personalization generated about 80 million pounds in sales in the first half, up about 37% year over year. Hello Casto has helped about 350,000 customers and has on the order of 60,000 users per month.
Results Proof A
The central figure, about 80m GBP in sales up about 37% year over year, comes from Kingfisher's 2025/26 half-year results (financial document from the brand). It is corroborated by specialized press (Digiday on Hello Casto users) and by the Google Cloud partnership announcement. Its presence in the financial results puts it at A.
How it works
Documented architectureThe stack in detail
- plateforme Athena (orchestration in-house) Kingfisher's orchestration framework that connects several language models to the group's retail brands
- plateforme Google Cloud Vertex AI AI foundation of the Google Cloud partnership, including Vertex AI Search for Commerce for search and assistants
- infra Nucleus (data lake) group data lake: browsing and purchase behavior, catalog, foundation for the recommendations
- outil Moteurs de recommandation et assistants Hello Casto / Hello B&Q personalized recommendations on product pages and conversational genAI DIY assistants per brand
How it runs, concretely
For ops teams-
1Question or browsing client
The customer asks a DIY question to Hello Casto or Hello B&Q, or browses the product pages.
-
2Orchestration AI
Athena routes the request to the right model and composes advice and recommendations.
-
3Personalized recommendation AI
The recommendation engines rank products according to the customer's behavior and context.
-
4Measurement and attribution data team
The teams track sales attributed to recommendation and personalization and arbitrate between the models.
The browsing and purchase behavior that feeds the recommendations, and the quality of the catalog for the assistant. Without clean browsing data or a structured catalog, recommendation loses relevance and the assistant answers off target.
How your customers perceive this type of use
Sourced studiesLe paradoxe est documente des deux cotes : 71% des consommateurs attendent des interactions personnalisees et 76% sont frustres quand elles manquent (McKinsey, 2021), mais 75% declarent ne pas acheter aupres d'organisations auxquelles ils ne confient pas leurs donnees (Cisco, 2024). La « creepy line » est localisee : messages recus quelques secondes apres une recherche et suivi de localisation sont les pratiques qui mettent le plus mal a l'aise (Periscope by McKinsey, 2019).
Acceptance conditions
- La confiance dans le traitement des donnees precede l'achat : 75% ne achetent pas sans elle (Cisco 2024)
- Un cadre legal protecteur rassure : 59% des consommateurs disent que des lois fortes sur la vie privee les rendent plus a l'aise pour partager des informations dans des applications IA (Cisco 2024)
- La personnalisation elle-meme est attendue quand elle est consentie : environ la moitie des consommateurs (US 55%, UK 52%) disent s'inscrire souvent ou parfois a des services personnalises (Periscope by McKinsey 2019)
Red lines
- Le message declenche quelques secondes apres une recherche ou un achat : deuxieme ou troisieme cause de malaise selon les pays (Periscope by McKinsey 2019)
- Le suivi de localisation percu comme de la surveillance : 40% de malaise en Allemagne et au Royaume-Uni (Periscope by McKinsey 2019)
- Le mesusage des donnees personnelles par l'IA, devenu la premiere inquietude des consommateurs, a 53% et en hausse (Qualtrics 2025)
Sources: McKinsey & Company 2021 · Periscope by McKinsey 2019 · Cisco 2024 · Qualtrics 2025
How to replicate
Inference, not sourcedData prerequisites
- Unified data lake of browsing and purchase behavior
- Structured product catalog for the assistant and recommendation
- Attribution of sales to recommendations
Org prerequisites
- Multi-model orchestration framework
- Group data team able to serve several retail brands
- AI transparency rule
Possible stack
- LLM orchestration layer like Athena
- Behavior-based recommendation engine
- Commerce semantic search
The plan, step by step
- Step 1Unify browsing data and the product catalog in the data lakeDeliverable: Structured catalog + queryable browsing data
- Step 2Connect a recommendation engine to the product pages with sales attributionDeliverable: Personalized recommendation in production on a pilot brand
- Step 3Build the conversational assistant on the catalog (step-by-step advice + products)Deliverable: Assistant in beta on a scope of DIY questions
- Step 4Launch the assistant on the pilot brand and measure usage and attributed salesDeliverable: Monthly users + sales attributed to recommendation and the assistant
- Step 5Pool through a multi-model orchestration layer and extend to the other brandsDeliverable: Common framework like Athena serving several retail brands
First step: Unify browsing data and the catalog, connect a recommendation engine, then expose a conversational assistant on a pilot brand.
The same pattern, proven in other industries
Function of Beauty
personalized product formulation by algorithm from a quiz
Henkel
genAI content supply chain for one-to-one personalization at scale
Hims & Hers
AI-driven personalized treatment routing on intake data (MedMatch)
Sources
- S1 Kingfisher and Google Cloud Partner to Deliver AI-Powered Shopping Across UK and Europe Primary archive pending
- S2 As brands look for AI edge, B&Q retail owner Kingfisher is expanding in-house development Established press archive pending
- S3 B&Q and Screwfix owner Kingfisher taps GenAI as it builds data led omnichannel customer experience Secondary archive pending
- S4 Kingfisher plc - Half year results for the six months ended 31 July 2025 Primary archive pending
An error, newer info, a source?
This page lives on its accuracy. If a figure has moved, if the deployment has changed, or if you have a higher-quality source, tell us. Every sourced correction is verified before publication.