How Kering uses AI: luxury clienteling and conversational commerce

KEY TAKEAWAYS Summary by the editors

1. Gucci was the first pilot customer for Salesforce's Einstein GPT after its launch in March 2023, using generative AI to support client service advisors rather than to replace them with a chatbot.
2. In April 2023 Kering launched KNXT, a US platform whose ChatGPT-based assistant Madeline recommended products from Kering houses such as Gucci, Bottega Veneta and Balenciaga.
3. At its Capital Markets Day on 16 April 2026 Kering said it is consolidating proprietary and external data into a unified client base, powered by AI, to inform decisions from creation to clienteling.
4. Kering's 2026 plan also includes a technology hub built on cloud-native systems, agentic AI and digital twins, while individual houses keep responsibility for client relationships.
5. Kering has not published group-level results for its AI initiatives; the only reported figure on Gucci's pilot comes from the technology vendor.

Kering uses AI mainly to support the people who sell and serve its clients: Gucci's client service advisors were the first pilot users of Salesforce's generative AI in 2023, Kering tested a ChatGPT-based shopping assistant called Madeline the same year, and in April 2026 the group announced an AI-powered unified client base as part of its ReconKering strategy. The common thread is clienteling, the personal relationship between a luxury house and its customers.

What has Kering built?

Kering's public AI activity falls into three phases:

According to VentureBeat, an Italian delegation from Gucci met Salesforce's AI leadership in November 2021 to discuss AI assistance for customer service. Gucci wanted "high-touch, very accurate" support for its advisors and did not want a chatbot. When Einstein GPT launched in March 2023, Gucci became its first pilot customer.

Madeline was presented in April 2023 on KNXT, a Kering platform that combined an AI assistant with NFTs and cryptocurrency payment. The Interline reported that the assistant was derived from ChatGPT and trained on Kering product data from houses including Gucci, Bottega Veneta and Balenciaga.

Why did Kering invest in AI?

In luxury, the client relationship is the product as much as the item. Advisors need deep product knowledge and a consistent brand voice, and top clients expect a personal service across stores, phone and digital channels. Gucci's interest in AI started with a practical problem: during the pandemic its client service teams worked remotely and were understaffed, according to VentureBeat, and it needed a way to support advisors without months of product training.

The 2026 strategy widens this logic to the whole group. In its Capital Markets Day release of 16 April 2026, titled "ReconKering: True Luxury. Next Luxury.", Kering said it is "developing the luxury industry's most advanced client intelligence platform" and "consolidating proprietary and external data into a unified client base, powered by AI, to inform decisions upstream, from creation, sales & operations planning, and downstream across media, activation and clienteling".

Read also

How adidas uses AI: from design archives to generative content

How does it work (data, models, process)?

Kering has disclosed the structure more than the technology. The 2026 plan organises the group into hubs, including Industry, Client and Technology. The Technology hub is described as "a modern, scalable tech foundation, powered by cloud-native systems, agentic AI and next-generation digital twins". Luxury Tribune reported that the houses keep responsibility for creativity, desirability and managing client relationships, so the central platform supplies data and tools while advisors in each house remain the face to the client.

For Gucci's service pilot, VentureBeat reported that the AI went beyond filling staffing gaps: it coached representatives so they could "become revenue producers" and "brand storytellers" without months of product training. Analysis: a client intelligence set-up of this kind typically requires:

• Data needed: client profiles and purchase history across houses, product information, service history and consent records.
• Human role: advisors review and use AI suggestions; the 2023 pilot explicitly avoided an autonomous chatbot.
• Governance: a central client base must respect each house's brand and the privacy rules of each market.

What results has Kering reported?

Kering has not disclosed group-level results for AI in clienteling or conversational commerce. The only quantified figure on the Gucci pilot comes from the vendor side: Salesforce Ben, a Salesforce community site describing Salesforce's pilot programme, reported that case handling efficiency for users of the generative AI tool was 30 percent higher than for users without it. This is a vendor-reported pilot metric and should not be read as an audited business result. No usage or sales figures have been published for Madeline, and the unified client base was announced in 2026 as a plan, so results cannot yet be assessed.

What are the limits and open questions?

The first question is privacy. Merging client data from several houses and from external sources into one AI-powered base is ambitious under European data protection law, where purpose limitation and consent matter. The second is brand integrity: generative tools that write for advisors must sound like Gucci or Bottega Veneta, not like a generic assistant. The third is focus. KNXT combined AI with NFTs and crypto payments in 2023, and Kering has not publicly reported on its outcome, which illustrates how quickly experimental formats can fade. Analysis: the 2026 plan places AI in core operations rather than in showcase projects, which is a more durable position but also harder to execute.

Read also

How Nike uses AI: direct-to-consumer data and generative product creation

What can other fashion companies learn?

Analysis: Kering's path, from a narrow advisor pilot to a group-wide client platform, offers lessons for any brand that sells through personal relationships, including wholesale and private client teams:

1. Augment advisors first. Gucci's choice to support people rather than deploy a chatbot fits high-touch selling, where trust matters more than automation.
2. Unify client data before scaling AI. Kering's 2026 plan starts with a consolidated client base, because clienteling AI is only as good as the profile behind it.
3. Keep relationships local. Central platforms can supply insight while brands or sales teams own the client conversation.
4. Treat vendor metrics with care. Pilot figures published by technology suppliers are useful signals but not proof of business impact.

Frequently asked questions

How does Gucci use AI in client service?

Gucci was the first pilot customer for Salesforce's Einstein GPT in 2023. The generative AI coaches client service advisors so they can act as brand storytellers without months of product training, instead of replacing them with a chatbot.

What was Kering's Madeline assistant?

Madeline was a ChatGPT-based personal shopping assistant launched in April 2023 on Kering's KNXT platform in the US. It recommended products from Kering houses such as Gucci, Bottega Veneta and Balenciaga. Kering has not publicly reported on its later status.

What is Kering's AI-powered unified client base?

Announced at Kering's Capital Markets Day on 16 April 2026, it consolidates proprietary and external client data into one base, powered by AI. Kering says it will inform decisions from creation and sales and operations planning to media, activation and clienteling.

Has AI improved Kering's results?

Kering has not published results for its AI initiatives. A vendor-reported figure from Gucci's 2023 pilot showed 30 percent higher case handling efficiency for advisors using the tool, but this is not an audited business outcome.

GuideThe complete guide to AI in fashion e-commerce, marketing and retailRead the complete guide