《麥肯錫：2023生成式AI報告-解鎖時尚的未來（英文版）（7頁）.pdf》由會員分享，可在線閱讀，更多相關《麥肯錫：2023生成式AI報告-解鎖時尚的未來（英文版）（7頁）.pdf（7頁珍藏版）》請在三個皮匠報告上搜索。

1、March 2023QuantumBlack,AI by McKinsey,Retail,and Digital PracticesGenerative AI:Unlocking the future of fashionWhile still nascent,generative AI has the potential to help fashion businesses become more productive,get to market faster,and serve customers better.The time to explore the technology is n

2、ow.by Holger Harreis,Theodora Koullias,Roger Roberts,and Kimberly TeAs this seasons fashion weeks wrap up in London,Milan,New York,and Paris,brands are working to produce and sell the designs theyve just showcased on runwaysand theyre starting next seasons collections.In the future,its entirely poss

3、ible that those designs will blend the prowess of a creative director with the power of generative artificial intelligence(AI),helping to bring clothes and accessories to market faster,selling them more efficiently,and improving the customer experience.By now,youve likely heard of OpenAIs ChatGPT,th

4、e AI chatbot that became an overnight sensation and sparked a digital race to build and release competitors.ChatGPT is only one consumer-friendly example of generative AI,a technology comprising algorithms that can be used to create new content,including audio,code,images,text,simulations,and videos

5、.Rather than simply identifying and classifying information,generative AI creates new information by leveraging foundation models,which are deep learning models that can handle multiple complex tasks at the same time.Examples include GPT-3.5 and DALL-E.(For more on generative AI and machine learning

6、,see“What is generative AI?”1 and“Generative AI is here:How tools like ChatGPT could change your business.”2)While the fashion industry has experimented with basic AI and other frontier technologiesthe metaverse,nonfungible tokens(NFTs),digital IDs,and augmented or virtual reality come to mindit has

7、 so far had little experience with generative AI.True,this nascent technology became broadly available only recently and is still rife with worrisome kinks and bugs,but all indications are that it could improve at lightning speed and become a game changer in many aspects of business.In the next thre

8、e to five years,generative AI could add$150 billion,conservatively,and up to$275 billion to the apparel,fashion,and luxury sectors operating profits,according to McKinsey analysis.From codesigning to speeding content development processes,generative AI creates new space for creativity.It can input a

9、ll forms of“unstructured”dataraw text,images,and videoand output new forms of media,ranging from fully-written scripts to 3-D designs and realistic virtual models for video campaigns.These are still early days,but some clear use cases for generative AI in fashion have already emerged.(Many of these

10、use cases also apply to the adjacent beauty and luxury sectors.)Within product innovation,marketing,and sales and customer experience in particular,the technology can have significant outcomes and may be more feasible to implement in the short term compared with other areas in the fashion value chai

11、n.In this article,we outline some of the most promising use cases and offer steps executives can take to get started,as well as risks to keep in mind when doing so.In our view,generative AI is not just automationits about augmentation and acceleration.That means giving fashion professionals and crea

12、tives the technological tools to do certain tasks dramatically faster,freeing them up to spend more of their time doing things that only humans can do.It also means creating systems to serve customers better.Heres where to begin.Understanding the use casesGenerative AI has the potential to affect th

13、e entire fashion ecosystem.Fashion companies can use the technology to help create better-selling designs,reduce marketing costs,hyperpersonalize customer communications,and speed up processes.It may also reshape supply chain and logistics,store operations,and organization and support functions(see

14、sidebar,“Generative AI use cases in fashion”).1“What is generative AI?”McKinsey,January 19,2023.2 Michael Chui,Roger Roberts,and Lareina Yee,“Generative AI is here:How tools like ChatGPT could change your business,”McKinsey,December 20,2022.2Generative AI:Unlocking the future of fashionGenerative AI

15、 use cases in fashionFoundation models and generative AI can be used across the fashion value chain.Merchandising and product:Convert sketches,mood boards,and descriptions into high-fidelity designs(for example,3-D models of furniture and jewelry).Enrich product ideation by collaborating with AI age

16、nts that generate creative options(for example,new ideas,variations)from data(for example,past product lines,inspirational imagery and style).Customize products for individual consumers at scale (for example,eyeglasses based on facial topography).Supply chain and logistics:Support negotiations with

17、suppliers by compiling research.Augment robotic automation for warehouse operations and inventory management through real-time analytics(for example,insights enabled by augmented reality,or AR).Tailor product return offers based on individual consumers.Marketing:Identify and predict trends to improv

18、e targeted marketing from unstructured data(for example,consumer sentiment,in-store consumer behavior,omnichannel data).Automate consumer segmentation at scale to tailor marketing initiatives.Generate personalized marketing content based on unstructured data from consumer profiles and community insi

19、ghts.Collaborate with AI agents to accelerate content development and reduce creative blocks for in-house marketing teams.Digital commerce and consumer experience:Structure and generate sales descriptions based on past successful sales posts.Personalize online consumer journey and offers(for example

20、,web pages,product descriptions)based on individual consumer profiles.Tailor virtual product try-on and demos to individual consumers(for example,clothing try-on,styling recommendations).Enhance intelligent AI agents (for example,conversational chatbots,virtual assistants)and self-service to address

21、 advanced consumer inquiries(for example,multilingual support).Store operations:Optimize store layout planning by generating and testing layout plans under different parameters(for example,foot traffic,local consumer audience,size).Optimize in-store labor to avoid bottlenecks such as gaps in staff a

22、llocation and theft detection through real-time monitoring of video data.Support AR-assisted devices to better inform workforce in real time on product(for example,condition,assortment,inventory,recommendations).Organization and support functions:Coach sales associates to sustain successful“clientel

23、ing”relationships via real-time recommendations,feedback reports,and high-value consumer profiles.Develop individualized training content for employees based on role and performance.Enable self-serve and automate support tasks(for example,HR tickets,accounting for large documents,review of legal doc

24、uments).3Generative AI:Unlocking the future of fashionProduct development and innovationInstead of relying on trend reports and market analysis alone to inform designs for next seasons collection,both mass-market fashion retailers and luxury brands creative directors can use generative AI to analyze

25、 in real time various types of unstructured data.Generative AI can,for example,quickly aggregate and perform sentiment analysis from videos on social media or model trends from multiple sources of consumer data.Creative directors and their teams could input sketches and desired detailssuch as fabric

26、s,color palettes,and patternsinto a platform powered by generative AI that automatically creates an array of designs,thus allowing designers to play with an enormous variety of styles and looks.A team might then design new items based on these outputs,putting a fashion houses signature touch on each

27、 of the looks.This opens the door to creating innovative,limited-edition product drops that may also be collaborations between two brands.Products such as eyeglasses could be designed for individuals by using facial-recognition technology powered by generative AI to scan facial topography and adjust

28、 for a customers size and style preferences.This scenario became reality in December 2022,when a group of Hong Kongbased fashion designers from the Laboratory for Artificial Intelligence in Design(AiDLab)held a fashion show featuring generative-AI-supported designs.3 Using tools from tech companies

29、such as Cala,Designovel,and Fashable,fashion designers are already tapping into the power of generative AI to spark new ideas,try myriad design variations without having to produce expensive samples,and vastly accelerate their processes.(For beauty businesses,generative AI also provides an opportuni

30、ty for brands to identify new product formulations,potentially helping to reduce lab testing costs.)MarketingMarketing executives and agencies can use generative AI to brainstorm campaign strategies,product campaign content,and even virtual avatars for every marketing channeland do it fast.Striking

31、marketing gold can often be a numbers game.Consider TikTok:theres no single winning formula for going viral on the platform.Instead,the more you produce,the higher your chances are of becoming a trending topic and boosting brand awareness and sales.Prompting a generative-AI-powered video platform to

32、 create short-form videos for TikTok or other social-media platforms can help save time and costs associated with pumping out social-media content.Generative AI can recognize patterns and trends in viral content and create new content that also follows specifications from the marketer.These exercise

33、s can help in-house marketing teams manage their workloads while reducing their reliance on outsourcing work to creative agencies.Marketers will want to be careful with this approach,however:trying to reach consumers by replicating what other brands have done can counteract the unique identity and v

34、alue proposition that a brand spends years building.Generative AI could also be applied to personalized customer communications.Companies that excel at personalization increase revenues by 40 percent compared with companies that dont leverage personalization,according to McKinsey research.4Several s

35、tart-upsCopyAI,Jasper AI,and Writesonic,to name just a feware helping pioneer personalized marketing at scale through generative AI.Using these tools,a marketers daily tasks might start to look like this:they could choose the type of content they want to create,whether its an email,a long-form blog

36、post,or something 3“In Hong Kong,designers try out new assistant:AI fashion maven AiDA,”Reuters,December 28,2022.4“The value of getting personalization rightor wrongis multiplying,”McKinsey,November 12,2021.4Generative AI:Unlocking the future of fashionelse;add a prompt describing what they are look

37、ing for;and include the targeted audience and other parameters,such as tone,that help create marketing communications that are in line with the brand.The AI tool then offers several options from which the marketer can choose.These tools are most helpful when applied to lower-funnel marketing channel

38、s(those that are mostly used to encourage sales conversions)as opposed to more prestigious brand-building communications.Marketers are still required to prompt and edit the work.Sales and consumer experienceTodays generative-AI-powered chats,which use stronger natural-language processing to better u

39、nderstand and interact with humans,are already a measurable improvement over existing AI chats.That said,there isnt(yet)a foolproof generative-AI chatbot for businessescurrent chatbots and other text-generating tools still occasionally make errors that could cause serious customer service disasters.

40、Eventually,though,this technology could help customer support agents outsource complex inquiriesfor example,using chatbots to help provide personalized responses in numerous languages.Today,there are services that assign a generative AI“representative”to a brand to handle customer service queries ac

41、ross email,chat,text,and a brands own platforms.These services help to reduce customer service wait times and improve response times.Generative-AI agents can also serve luxury brands,particularly when it comes to“clienteling,”a retail strategy whereby sales associates develop long-term relationships

42、 with a brands highest-spending customers to encourage purchases and improve brand loyalty.(High-end brands can hit a sales conversion rate of 60 to 70 percent in luxury boutiques,through appointment-only shopping,for example.5)That process has remained somewhat analog and manual,relying on brands s

43、ales associates to reach out to customers through a variety of messaging platforms or texts,and is limited to only when those associates are working.Generative-AI-powered tools can keep the conversation going or make styling recommendations after a shopper leaves the store,coach sales associates on

44、how to engage with customers,personalize communications for specific customers,and analyze consumer profiles and online real-time interaction.In July 2022,apparel retailer Stitch Fix said it was experimenting with GPT-3 and DALL-E 2,the text-to-image AI generator,to boost sales and improve customer

45、satisfaction with better styling services.These generative models are being tested to help stylists quickly and accurately interpret reams of customer feedback and curate products that customers would be likelier to purchase.For example,the AI tool could analyze all of a customers feedback,which cou

46、ld include hundreds of text comments,email requests,product ratings,and online posts.If a customer regularly comments on,say,the“great fit”and“fun color”of a certain style of pants,DALL-E could generate images of similar pants that the customer would likely want to purchase.The stylist could then fi

47、nd similar items in Stitch Fixs inventory and recommend them to that customer.Virtual try-ons are yet another example of how generative AI can improve sales and consumer experience.Paris-based Veesual enables virtual try-on integration for e-commerce fashion brands,meaning customers can choose their

48、 model and pick clothes to try on.How to get startedAs exciting as generative-AI technology might be,companies will still want to tread cautiously before entrusting any of their core tasks entirely 5 Miles Socha,“What to watch:Clienteling is the new black for Europes luxury giants,”Womens Wear Daily

49、,January 4,2023.5Generative AI:Unlocking the future of fashionto generative AI.But neglecting to explore the possibilities that this technology offers could be just as risky,given the pace at which it is evolving and the explosive growth of the user base.Executives can start thinking now about how t

50、heir businesses could use generative AI.There are a few steps leaders can take to begin.Make value your North StarFashion leaders should outline where generative AI can offer the greatest value to their business.Start by noting which areascreative design,merchandising,runway campaigns,or clienteling

51、could benefit the most from generative AI.Leaders can then prioritize the generative AI use cases they should pursue based on the level of impact the use cases may have on their business.Some measures of impact include improving customer satisfaction scores and reducing customer service wait times.O

52、nce the value is identified,use cases should also be prioritized according to how feasible they are to implement;determining how seamlessly generative AI can be used will depend on things like a teams technical skills.Afterward,teams should build a short-term road map to test and validate these use

53、cases.At the same time,they can also consider what long-term goals might include,such as how to build a generative-design platform that can be updated and used by designers for every season.It may be tempting to have a bit of fun with generative AI,but harnessing its power will take extra diligence.

54、Fashion executives must be intentional in building tools that can deliver value rather than experiment with existing tools indiscriminately.Know risks and plan to mitigate themIn a previous article,we listed some of the risks of using generative AI.One is that the legal parameters around generative

55、AIs use are still being ironed out.Designers are sometimes criticized for creating derivative works and copycat designs.Determining who owns the intellectual property and creative rights to AI-generated works,which could be based on multimodal data sources such as other designers past collections,wi

56、ll be decided on a case-by-case basis until there is a strong legal precedent.(Although it doesnt involve generative AI,the high-profile battle between Herms and artist Mason Rothschild surrounding MetaBirkin NFTs,in which a judge ruled that the NFTs infringed on Hermss trademark,shows how fashion b

57、rands can become embroiled in legal conundrums when new technologies emerge.)Another risk is bias and fairness in generative-AI systems,particularly around biased data sets,which may present reputational challenges for brands that rely on the technology.For example,if an image-generating tool produc

58、es an advertising campaign with inappropriate or offensive images that are then shared globally,a brands reputation could be hurt.And pointing fingers at the company AI in an attempt at damage control may do little to calm consumer ire.There is also the risk that employees who use generative AI are

59、not fully aware of its shortcomings and may fail to check for errors introduced by the technology.In this case,businesses must regularly train employees and provide them with the resources they need to understand how to use the technology.While risks are unavoidable,executives can mitigate their pot

60、ential impact by establishing a process to address risk,ethics,and quality assurance.Upskill your current workforceGenerative-AI tools could add value to a host of different areas of a business,so it will be important to educate and train employeesincluding designers,marketers,sales associates,and c

61、ustomer service representativeson the use of the technology.Some businesses have already introduced AI-focused training.Levi Strauss,for one,launched a machine learning boot camp in 2021 to train nontech employees on how to use machine learning 6Generative AI:Unlocking the future of fashionin the co

62、mpanys design process.Employees who complete the program create new AI tools that are relevant to their work.6 One of Levis goals with the program is to increase the diversity of employees who have tech knowledge so that the company can uncover problems that employees who come from traditional techn

63、ology backgrounds might otherwise miss.The program also helps teams with different specializationssuch as design teams and engineering teamscommunicate better and find common ground.Furthermore,Levis has found that the program helps improves employee retention.7With an AI-savvy workforce,collaborati

64、on will take on a new meaning.Leaders should consider:How do we define responsibilities and operate collectively between technical and nontechnical roles?Design and software engineering teams can set up weekly leadership meetings to strategize quarterly road maps and working sessions among teams.Des

65、ign leads can share their needs for certain insights and tools(a tool that generates design variations from a sketch,perhaps),while engineering teams deliver those tools.Partner with the right tech supportFashion businesses will no doubt have to invest in their workforce when it comes to leveraging

66、generative AI,but they wont have to build out applications or foundation models themselves.Instead,fashion leaders can partner with generative-AI businesses and experts to move quickly.A fashion executive might partner with a business(such as Microsoft or OpenAI)that provides new technology or a par

67、tner that provides support capabilities(such as cloud computing or APIs).While the potential use cases for generative AI are coming to light quickly,the future of this technology in the apparel and luxury industries is still being stitched together.But experimenting with new tools today means openin

68、g infinite possibilities tomorrow.6 Maghan McDowell,“Inside the new look digital clienteling tool,”Vogue Business,January 11,2022.7 Maghan McDowell,“Exclusive:Lessons from Levis data science bootcamp,”Vogue Business,November 8,2021.Scan Download PersonalizeFind more content like this on the McKinsey

69、 Insights AppDesigned by McKinsey Global PublishingCopyright 2023 McKinsey&Company.All rights reserved.Holger Harreis is a senior partner in McKinseys Dsseldorf office,and Theodora Koullias and Kimberly Te are consultants in the Bay Area office,where Roger Roberts is a partner.7Generative AI:Unlocking the future of fashion