The Estée Lauder Companies (ELC) has wired AI into all three gates of agentic commerce in under two years: whether agents can reach its brands, whether they choose them, and whether the sale completes. The open question is what happens when an agent asks the brand for a judgment rather than a product listing.
The leader
Stéphane de La Faverie became President and CEO of ELC in January 2025, after 14 years inside the company. He took over a business with falling sales and launched Beauty Reimagined, a turnaround plan he describes as the biggest operational, leadership and cultural transformation in ELC's history.
The early numbers have moved. ELC returned to organic sales growth of 3% in fiscal 2026 (July 2025–June 2026), with operating margin expanding. Department stores now account for 30% or less of the business, as he shifts weight to Amazon, TikTok Shop, Sephora and brand-owned digital.
His read of the consumer is the thread that runs through everything below. Younger shoppers drive beauty trends and pull older generations with them, and he has said ELC was too slow to follow them into new digital touchpoints. Agents are the next of those touchpoints.
He has built the leadership to match. Brian Franz, ELC's first C-suite leader for technology, data and analytics, now runs technology and transformation as Chief Technology and Transformation Officer. Aude Gandon, from Nestlé, joined as Global Chief Digital and Marketing Officer. Both report directly to him.
Why it matters
Beauty is one of the categories where AI agents are changing the purchase fastest. A shopper can now describe a skin concern, a budget and a scent preference to ChatGPT or Gemini and get a shortlist back, often with a link to buy.
That moves the first conversation away from the counter, the brand site and the search results page. For a prestige house built on beauty advisors and department stores, the question is not whether to follow the shopper, but how to stay visible, accurate and chosen when a machine is doing the first round of advice.
ELC's answer so far can be read through the three gates every brand now has to pass: Access (can agents reach and read us?), Choice (do they recommend us?) and Execution (can the sale complete?).
Gate 1 — Access: making the catalog agent-readable
ELC's biggest Access move is plumbing. Shopify runs its direct-to-consumer omnichannel business, and Fortune reported in September 2026 that ELC would switch on Shopify's agentic feature in October, starting with M·A·C. Shopify's Agentic Storefronts syndicate a merchant's catalog to ChatGPT, Microsoft Copilot, Google AI Mode and Gemini, with checkout staying on the brand's own store.
Underneath that sits a data project. Franz's three priorities are a frictionless omnichannel experience, unifying data across an organization with more than 80 years of history, and "AI everywhere." The One ELC operating model is pulling media, digital operations and supply chain onto a shared global data, AI and technology base, and ConsumerIQ, built with Microsoft, consolidates consumer and market data.
Franz has called agentic shopping "still early." The aim he describes is simple: shoppers find products through AI assistants and buy in one or two clicks.
Gate 2 — Choice: being represented accurately
In September 2026, ELC signed a global strategic partnership with Profound and became the first prestige beauty company to deploy it across its full brand portfolio. The goal is a single view of how each brand appears and is recommended on platforms such as ChatGPT and Gemini.
The second half of the deal is content. ELC will use Profound's insights and agentic tools to rework product description pages, blogs, YouTube channels and content at scale. The aim is to make ingredients, claims and benefits clearer to large language models. Those insights feed into ELC's "One Operating Ecosystem."
Gandon framed it as shaping the shift rather than reacting to it, with every brand in the portfolio able to show up distinctly, consistently and accurately. It is the clearest public statement yet that ELC treats AI recommendation as a brand-building channel, not an SEO side project.
Gate 3 — Execution: converting the conversation
At Shoptalk in March 2026, ELC announced it would deploy Rezolve Ai's Brain Suite across its brand sites in 70 EMEA markets, including Estée Lauder, Clinique, Bobbi Brown, M·A·C and Tom Ford. The suite brings AI-native search, conversational discovery and automated merchandising to the brands' own storefronts.
The design choice matters. Retailers such as Ulta have gone the other way, putting their catalogs into Google's AI surfaces. ELC's EMEA move keeps the conversation, and the customer relationship, on brand-owned ground.
Jo Malone London shows the same instinct at brand level: a consumer-facing, AI-enabled scent advisor built with Google Cloud. Add Shopify's agentic checkout, and ELC has conversational selling running both on its own sites and inside third-party assistants.
The engine room
The outward-facing bets rest on a broad internal rollout. About 14,000 employees have Microsoft Copilot and ChatGPT, and more than 20 AI champions across the brands now share ideas on building agents. Younger staff "reverse mentor" senior leaders on the tools.
The narrower uses carry hard numbers:
| Initiative | What it does | Reported result |
|---|---|---|
| Formula navigator | Speeds new formulations in R&D | 20% gain in development efficiency |
| Ella (Estée Lauder Line Assistant) | Helps plant operators fix equipment | Faster production-line changeovers |
| AI media optimization | Optimizes North American campaigns | 31% better media ROI (fiscal 2025) |
| Adobe generative AI | Scales campaign content | Not disclosed |
Franz sums up the philosophy: technology doesn't adopt itself; culture and change drive the outcomes. De La Faverie made the same point when he expanded Franz's role, saying transformation at ELC has no end date.
The next layer: the brand as the expert
ELC has covered the three gates from the outside in: syndicate the catalog, optimize the content, convert on the site. What the agentic economy adds next is a question coming the other way: an agent asking the brand directly.
A shopping agent working for one person needs a judgment as well as a listing. Is this serum right for reactive skin? Is this foundation available in a shade for deep olive undertones? Which Jo Malone scent fits someone who dislikes florals? Today it pieces that answer together from product pages, reviews and third-party content. Better content improves the raw material. It does not let the brand give the answer itself.
That is the gap AIVO, Inc. built AIVO Assist to close, and the reason for this series: Assist begins deployment today. Assist is a brand-side product expert that AI agents consult directly. The agent passes the shopper's need (no personal data) and receives a judged answer in about two seconds: Fits, Fits, if…, No suitable option or Need a little more information. Each answer comes with the deciding facts and sources, drawn word for word from a knowledge repository the brand has signed off. Every answer carries a reference code, so the sale it influences can be attributed.
AIVO, Inc. controlled test, October 2026. Working paper.
For a portfolio of more than 20 prestige brands, the fit is direct:
- Accuracy where claims matter. Skincare and fragrance claims are regulated and brand-sensitive. A signed-off repository with a never-say list keeps agents from improvising them.
- Judgment, not just presence. A brand that can tell an agent "not for you" when it isn't earns trust the next time it says "this fits."
- Demand intelligence. Agents' questions arrive in shoppers' own words, a live read on unmet needs by brand and market.
- Attribution. Agent-driven visits are largely invisible to standard analytics. Reference codes close that loop.
It sits alongside what ELC already runs. Shopify gets the catalog to the agent, Profound improves what the agent reads, and Rezolve converts on the brand site. Assist answers the agent's question on the brand's behalf.
What to watch
- M·A·C as the test case. As the first brand on Shopify's agentic channel, its results will show whether agent-driven sales justify rolling out across the portfolio.
- From visibility to verdicts. Whether ELC moves from tracking how agents describe its brands to answering agents directly.
- EMEA as the template. If the Rezolve rollout across 70 markets lifts conversion, expect the brand-led model to travel to the Americas and Asia.
- Fragrance. It is ELC's growth engine and its most subjective category, which makes it the hardest test of whether an agent can recommend well.
- Proof in the numbers. De La Faverie has tied Beauty Reimagined to sustained growth and a solid double-digit operating margin. Agentic commerce will be judged against that, not against pilots.
See how AI agents choose in your category, and what your brand's own expert would change.
Talk to AIVO, Inc.Sources
- Fortune: Estée Lauder's AI shopping technology and Brian Franz (Sept. 9, 2026)
- Business Wire: ELC partnership with Profound (Sept. 2026)
- Stellagent: ELC and Rezolve Ai across 70 EMEA markets
- Cosmetics Business: expanded roles for Brian Franz and Amber English
- Invezz: Aude Gandon appointed Chief Digital and Marketing Officer
- Digital Commerce 360: ELC online sales and AI, fiscal Q2 2026
- Glossy: Inside ELC's reset, AI and media ROI (Aug. 2025)
- ELC fiscal Q4 2026 earnings call transcript
- IADS: ELC and Microsoft build ConsumerIQ
- AIVO Assist working paper, Sheals & de Rosen (Zenodo, Oct. 2026)