Lepota vođena podacima: kako The Estée Lauder Companies otključava uvide uz ChatGPT

The Estée Lauder Companies⁠(отвара се у новом прозору) (ELC) je globalni lider u prestižnoj lepoti, sa portfoliom od više od 20 kultnih brendova uključujući Clinique, La Mer, Bobbi Brown Cosmetics i Aveda. Kompanija i njeni brendovi poznati su po vodećim proizvodima za lepotu uparenim sa personalizovanim iskustvima visokog nivoa usluge za svoje potrošače širom sveta.

Kako se svet lepote kreće ka hiperpersonalizaciji, tako se razvijao i pristup kompanije istraživanju i tehnologiji. Široko usvajanje ChatGPT Enterprise u ELC-u pomaže kompaniji da osnaži kreativnost zaposlenih i otključa uvide kako bi bolje služila današnjim potrošačima.

Globalna industrija lepote kreće se brzo, a na ELC-ov proizvodni pipeline utiču ogromne količine potrošačkih podataka iz anketa, kliničkih ispitivanja i upotrebe proizvoda.

The ELC team adopted ChatGPT for data processing and analysis at scale. Today, ChatGPT is an integral part of ELC’s workflows, with more than 240 custom GPTs that give their employees greater time and insights to do what they do best: develop new and market existing products with precision, aligned to emerging skincare and beauty trends.

Central to ELC’s rapid adoption of AI was the company’s core belief that creativity lies in every chair. “When OpenAI came out with ChatGPT, we asked employees to tell us how they would use it,” says Raheel Khan, SVP of Foresight and Growth Intelligence at ELC. “Over one thousand people submitted ideas.”

This interest led ELC to establish their GPT Lab: A cross-functional group empowered to experiment and develop business solutions—“custom GPTs”—with ChatGPT. 

“Our job is to uncover high-impact use cases that drive value to our different brands and regions,” said Charmaine Pek, ELC’s Director of ChatGPT Enterprise Adoption. “The GPT Lab aims to identify patterns across meaningful use cases and amplify these successes to scale to more brands and regions.”

Za samo 10 nedelja, članovi Laba razvili su više prilagođenih GPT‑ova, svaki sa sopstvenom svrhom i obimom.

Jedna od istaknutih aplikacija koje je kreirao GPT Lab je Fragrance Insights GPT. Razvijen da pomogne timu Fragrance Foresight da izvuče uvide iz anketa potrošača, ovaj GPT analizira velike skupove podataka iz potrošačkih anketa kako bi otkrio trendove i preferencije, omogućavajući ELC-u da dizajnira proizvode koji odgovaraju različitim demografskim grupama.

„Ranije smo provodili sate ručno čisteći i organizujući podatke da bismo otkrili uvide“, kaže Yuan Zhan, direktorka ELC-ovog tima Fragrance Foresight. „Uz Fragrance GPT, možemo postavljati složena pitanja na običnom engleskom, a on trenutno pretražuje podatke.“

Članovi Laba su takođe pokrenuli Clinical Trial Data GPT, koji brzo izvlači uvide o efikasnosti proizvoda za negu kože — na primer, određivanje procenta trenutnog poboljšanja hidratacije proizvoda kao što je serum Estée Lauder Advanced Night Repair — iz hiljada izveštaja kliničkih ispitivanja jednostavnim upitom.                 

Ostali GPT‑ovi uključuju Copywriting GPT, prilagođenog asistenta za pisanje marketinških tekstova za različite brendove, osmišljenog da kreira detaljan, smislen sadržaj usklađen sa brendom na različitim platformama, kao i Vendor Snapshot Creator GPT, koji objedinjuje ključne uvide o svakom dobavljaču, uključujući njihov profil, istoriju kupovine ELC-a i druge relevantne detalje.

The Estée Lauder Companies structures its GPT creation process like a sprint, focusing on quick prototyping and testing to identify success predictors and build for scale.

“We look at the value the GPT would bring to the organization, and we also look at the effort that is needed,” explains Kingsuk Chakrabarty, Director of Enterprise Architecture, AI and R&D at ELC. “Then we prioritize the GPTs which have high value and can be built quickly.”

Within the GPT Lab, “teams” of people—a business user, a subject matter expert (SME), and a technical lead—are responsible for ensuring each idea is based on impact and feasibility:

1. Design: the business user defines the purpose, scope, and audience for the GPT in a two-page Use Case Brief to gain clarity before building.
2. Prepare: the SME shapes the use case by gathering and preparing relevant data, ensuring best practices for GPT development.
3. Build & Test: the tech lead builds the GPT with sets of data and rigorously tests to assess accuracy and consistency.
4. Launch: the full team deploys the GPT and a user guide for teams to use.
5. Pivot & Scale: the full team uses feedback loops to iterate and optimize the outputs based on GPT performance.

“Designing the right use cases means asking the right questions,” shares Pek. “Why do we want to build this GPT? What is the problem that we're trying to solve? What impact will it have?”

By working in iterative cycles and sharing insights, the GPT Lab was able to develop AI-powered prototypes in a number of weeks.

Not only has ChatGPT helped employees work faster, it’s expanded their creative capacity by reducing manual tasks:

• Time savings: Across R&D and marketing teams, ChatGPT has improved response time by more than 90%. What once took teams several hours to research—like finding claims about product efficacy—now takes minutes. 
• Speed to market: By accelerating data analysis, ELC can launch products faster, ensuring they remain responsive to fast-changing consumer trends. 
• Internal adoption: There’s been enthusiastic adoption of GPT across the organization, with more teams asking to integrate AI into their workflows for creative solutions.

“Using ChatGPT Enterprise has done two things,” says Khan. “One, it allows our incredible human talent to spend more time on tasks that require creativity. And two, it actually increases capability to deliver on what matters most to our consumers.” 

ELC plans to apply top GPTs across its portfolio of brands, empowering more teams to take advantage of AI-driven insights and creativity. 

“AI enables us to deliver market-leading products on a larger scale, and better. With OpenAI, we’re reducing low-value work for our employees and giving our teams the opportunity to create on a whole new scale,” says Lauder.