Shopify
genAI generation of merchant content
In Q3 2025, Shopify's genAI agent Sidekick onboarded 750,000 new merchants and accumulated nearly 100 million interactions, with the Magic suite generating descriptions, images, and emails delivered by default on all plans.
Key points
- GenAI generation of merchant content (descriptions, images, emails) via Shopify Magic.
- Sidekick conversational agent that drives these functions in natural language.
- 750,000 merchants onboarded and nearly 100 million interactions in Q3 2025.
- Evidence level B, confirmed status, figures communicated by Shopify's president.
Objective
Roll out AI content generation to all merchants, without a separate subscription, to reduce the time spent on production tasks (listings, emails, visuals) and retain merchants on the platform.
The deployment
Shopify Magic groups the genAI functions built into the back office: generation of product descriptions, images, emails, and blog articles from the merchant's data. The description generator accepts detailed inputs (competitor product links, target personas, SEO keywords). Sidekick is the conversational agent that drives these functions in natural language, by voice or screen sharing, and covers task automation, analytics, and support. Magic is delivered by default to every merchant on every plan, with no separate AI subscription. Adoption figures communicated by Shopify's president in Q3 2025.
Results Proof B
Adoption figures communicated by Shopify president Harley Finkelstein in the context of the Q3 2025 results, reported by the press. Platform-scale usage metrics, but the quantified activity concerns the agent more than a consolidated financial result.
How it works
Inferred typical approachThe internal detail is not public. Here is a proven approach that leads to the same result, to adapt to your stack.
The stack in detail
- plateforme Shopify Magic genAI functions built into the admin: product descriptions, images, emails, and articles, generated from the merchant's data
- outil Shopify Sidekick conversational agent that drives these functions in natural language (voice, screen sharing) and covers automation, analytics, and support
- llm LLM de fondation underlying text and image generation models; the provider is not named in the case's sources
- infra Shopify admin back office where Magic is delivered by default on all plans, with no separate AI subscription
How it runs, concretely
For ops teams-
1Natural language request customer
The merchant describes what they want (copy, image, analysis) to Sidekick or activates Magic in the admin.
-
2Draft generation AI
Magic generates a description, image, email, or answer from the product data and the inputs provided.
-
3Acceptance or editing customer
The merchant accepts the draft or adjusts it before publishing.
The quality of the merchant's product data. The generator produces a weak description if the source listing lacks attributes or context.
How your customers perceive this type of use
Sourced studiesUn ecart net separe les annonceurs des consommateurs : 77% des annonceurs voient l'IA positivement contre 38% des consommateurs (Yahoo/Publicis, 2024). Les mesures implicites confirment le rejet declare : en EEG, les pubs generees par IA produisent une activation memorielle plus faible que les pubs traditionnelles et sont decrites comme agacantes, ennuyeuses et confuses (NIQ, 2024). La disclosure a un effet ambivalent : elle augmente fortement la confiance quand elle est remarquee (Yahoo/Publicis), mais 27% des jeunes consommateurs disent faire moins confiance a une entreprise dont la pub est creee par IA (IAB, 2024).
Acceptance conditions
- Une disclosure visible : quand la mention IA est remarquee, la confiance globale envers l'entreprise augmente de 96% (Yahoo/Publicis 2024)
- Une qualite visuelle suffisante : les visuels IA de basse qualite augmentent l'effort cognitif et distraient du message (NIQ 2024)
Red lines
- Le contenu IA non declare puis identifie : 72% des consommateurs disent que l'IA rend l'authenticite difficile a etablir (Yahoo/Publicis 2024) et les marques utilisant des pubs IA sont plus souvent jugees inauthentiques ou non ethiques par les consommateurs que par les dirigeants (IAB 2024)
- Les mannequins et personnes generes par IA : 46% des consommateurs n'en veulent pas dans la publicite, l'inquietude premiere etant les standards de beaute irrealistes (Attest 2025)
Sources: Yahoo / Publicis Media (terrain Ebco) 2024 · IAB (avec Attest) 2024 · NIQ (NielsenIQ) 2024 · Attest 2025
How to replicate
Inference, not sourcedData prerequisites
- structured product data
- store history
- merchant content templates
Org prerequisites
- back-office integration
- a default model with no activation friction
Possible stack
- foundation LLM via API
- image generation
- conversational agent on the back office
The plan, step by step
- Step 1List the highest-volume content types (listings, emails, visuals) and the product data available to feed themDeliverable: Backlog of use cases ranked by volume and value
- Step 2Integrate description generation into the listing-creation flow, as an editable draft enabled by defaultDeliverable: Generator in production on a first flow
- Step 3Add the enriched inputs: SEO keywords, target persona, competitor referencesDeliverable: Contextualized generation, quality assessed on a sample
- Step 4Expose the functions through a single conversational interface (agent)Deliverable: Agent covering several tasks, usage instrumented
- Step 5Track draft acceptance rate, time saved, and retention of users of the functionDeliverable: Adoption dashboard and prioritization of the next functions
First step: Integrate a content generator into the listing-creation flow and offer it enabled by default, with adoption following draft acceptance.
Sources
- S1 Shopify President: AI Tools Central to Holiday Sales Strategy as Sidekick Adoption Soars Secondary archive pending
- S2 Sidekick - AI-enabled commerce assistant Interested party 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.