The Home Depot
genAI conversational project assistant backed by an in-house knowledge base
Launched in March 2025, The Home Depot's genAI suite Magic Apron answers project questions across millions of product pages on homedepot.com and extended in January 2026 with Google Cloud toward multimodal input and the in-store experience.
Objective
Bring the expertise of in-store associates to the digital channel, to give the customer the confidence to take on their DIY project and remove decision friction on technical products.
The deployment
Magic Apron is The Home Depot's suite of genAI tools, launched on March 6, 2025. To start, an assistant present on millions of product pages of homedepot.com and in the app answers DIY questions and summarizes customer reviews, drawing on an in-house knowledge base that blends the retailer's data and its product expertise. The customer describes their project in natural language and gets advice and recommendations, from fixing a faucet to remodeling a kitchen. In January 2026, at NRF, Home Depot extended Magic Apron with Google Cloud: richer conversational capabilities, multimodal features (image upload), and an in-store experience in testing with local stock and aisle guidance. Home Depot then describes an assistant that already produces better engagement and resolution results.
Results Proof C
Official press releases from The Home Depot and Google Cloud naming the brand, plus major retail press on the launch. The scope (millions of pages, national extension) establishes scale, but no results figure is published, only a stated improvement in engagement and resolution, which caps at C.
How it works
Documented architectureThe stack in detail
- outil Magic Apron (assistant in-house) The Home Depot's custom genAI suite: project conversation, review summarization, RAG over the in-house base, multimodal features (image upload) added in 2026; the underlying LLM at launch is not named
- infra Base de connaissances proprietaire Home Depot Product data, project expertise, customer reviews, and local stock, the foundation of the RAG and the condition for accurate answers
- plateforme Google Cloud Partner for the agentic extension announced at NRF 2026: richer conversational capabilities, multimodal input, in-store experience in testing
- plateforme Site et app The Home Depot Deployment surfaces for the assistant, across millions of product pages
How it runs, concretely
For ops teams-
1Project question customer
The customer describes their project or asks a question on a product page, in text or via an image.
-
2Retrieval from the in-house base AI
The assistant retrieves advice, product specifications, and reviews from the proprietary knowledge base.
-
3Answer and recommendation AI
It composes project advice, summarizes reviews, and recommends products; in store, it adds local stock and aisle.
-
4Follow-up and extension data team
Teams measure engagement and resolution and extend the scope (pros, stores, voice).
The coverage and freshness of the product and project knowledge base. If it is not up to date, the assistant answers off the mark or steers toward an unsuitable product.
How your customers perceive this type of use
Sourced studiesLes consommateurs n'acceptent pas les chatbots par defaut : 64% prefereraient que les entreprises n'utilisent pas d'IA dans leur service client (Gartner, 2024) et pres d'un utilisateur sur cinq du service client par IA n'en retire aucun benefice (Qualtrics, 2025). L'acceptation se construit sur trois conditions mesurees par Salesforce : savoir qu'on parle a une IA, pouvoir escalader vers un humain, comprendre la logique de l'agent.
Acceptance conditions
- Etre informe qu'on parle a une IA et non a un humain (pres de 75% le demandent, Salesforce 2024)
- Un chemin d'escalade clair vers un agent humain (45% plus enclins a utiliser l'agent IA, Salesforce 2024)
- Une logique de l'agent clairement expliquee (44% plus enclins, Salesforce 2024)
Red lines
- Rendre l'humain injoignable : c'est la premiere inquietude des consommateurs sur l'IA dans le service client (Gartner 2024) et 50% craignent que l'IA les coupe du contact humain (Qualtrics 2025)
- Remplacer le service client par l'IA sans alternative : 53% envisageraient de partir chez un concurrent (Gartner 2024)
Sources: Salesforce 2024 · Gartner 2024 · Qualtrics 2025
How to replicate
Inference, not sourcedData prerequisites
- Structured product and project knowledge base
- Usable customer reviews for summarization
- Local stock and store layout for the in-store version
Org prerequisites
- AI transparency rule toward the customer
- Team to maintain and refresh the knowledge base
- Engagement and resolution measurement loop
Possible stack
- Generative LLM via API
- RAG layer over the in-house knowledge base
- Multimodal module for image upload
The plan, step by step
- Step 1Consolidate the knowledge base: product specifications, project guides, customer reviews, with a refresh process.Deliverable: Indexed, queryable knowledge base
- Step 2Connect a RAG assistant (LLM via API) on high-traffic product pages, with AI transparency and guardrails on technical advice.Deliverable: Assistant piloted on a first batch of product pages
- Step 3Add review summarization and measure engagement and question resolution against pages without the assistant.Deliverable: Quantified engagement / resolution review on the pilot scope
- Step 4Extend to the full catalog and prepare the extensions: multimodal (image), local stock, and in-store guidance.Deliverable: Generalized assistant and framed in-store roadmap
First step: Build a product and project knowledge base, connect a RAG assistant on high-traffic pages, then measure engagement and resolution.
The same pattern, proven in other industries
Booking.com
genAI trip planning assistant
Burberry
genAI copilot for consultants (search, image recognition, complete-the-look)
Homes.com
natural-language voice and text search, then conversational assistant
Sources
- S1 Unveiling Magic Apron: The Home Depot's Smartest Tool Yet Primary archive pending
- S2 The Home Depot and Google Cloud Launch Agentic AI Tools to Help Customers and Associates Primary archive pending
- S3 The Home Depot Launches New Suite of Gen AI Tools for Customers Established press 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.