AI In The Clothing Industry Statistics

61% of shoppers said personalization made them more likely to buy (research) [1]

80% of shoppers say they are more likely to purchase from a company that offers personalized experiences (survey) [2]

94% of shoppers said they would be willing to use a company’s personalization features if it improved their shopping experience (survey) [2]

74% of consumers feel frustrated when their shopping experiences are not personalized (research) [2]

67% of consumers expect retailers to understand their preferences and make relevant recommendations (survey) [3]

48% of consumers said they would be more likely to shop with retailers that use AI to recommend products they would like (survey) [4]

57% of consumers said they have used online recommendations to make a purchase (survey) [5]

45% of consumers said they prefer retailers that use AI to personalize (consumer research) [6]

79% of marketers reported AI tools have improved engagement (survey) [7]

23% of consumers said they had used a visual search feature to find products (survey) [8]

35% of online shoppers said they are more likely to buy when they can “see” products through augmented reality (study) [9]

61% of consumers said AR helped them make a purchase decision (consumer survey) [10]

40% of consumers said they have used AR to try on products virtually (survey) [11]

70% of respondents said they would pay more for personalized products (survey) [12]

31% of consumers said they want product recommendations based on their size (survey) [13]

63% of shoppers said they prefer personalized search results (survey) [14]

55% of consumers said they engage with chatbots to get product information (survey) [15]

20% of consumers said they trust chatbot responses as much as human agents for product recommendations (survey) [16]

33% of consumers said they used a chatbot to locate a product (survey) [17]

53% of consumers said they will use voice assistants to search for products (survey) [18]

42% of consumers said they want “smart” fitting recommendations (survey) [19]

62% of apparel shoppers said they would use a virtual stylist/AI recommendation tool (survey) [20]

25% of fashion shoppers said they have tried virtual try-on (survey) [21]

30% of consumers said they are more likely to purchase after using visual search (research) [22]

40% of fashion shoppers said AI size recommendations improved fit satisfaction (survey) [23]

2.5 billion pairs of shoes produced annually (global) [24]

1.8 million tons of textile waste generated in the US annually (estimate) [25]

Global apparel & footwear market revenue reached $2.6T in 2023 (estimate) [26]

Global fashion market expected to reach $3.0T by 2025 (forecast) [27]

Global fashion e-commerce sales reached $679B in 2023 (estimate) [28]

E-commerce share of global retail sales is projected to reach 24% by 2025 (forecast) [29]

The global retail AI market is projected to reach $XX by 2032 (forecast) [30]

The global computer vision market is projected to reach $XX by 2030 (forecast) [31]

The global AI in retail market was valued at about $3.4B in 2022 (estimate) [32]

The global AI in fashion market is expected to grow at ~35% CAGR (forecast) [33]

US retail apparel returns rate estimated around 20–30% annually (benchmark) [34]

About 60% of fashion consumers report returns as an issue (survey) [35]

5.8 million tons of textiles were landfilled in the US in 2018 (EPA) [36]

2.3 million tons of textiles were incinerated in the US in 2018 (EPA) [36]

US textile recycling rate about 15% in 2018 (EPA) [25]

In 2022, clothing and textiles were the second-largest category of waste by weight in the EU (Eurostat) [37]

EU textiles waste generated about 6.3 million tonnes in 2019 (Eurostat) [38]

Global textile waste generation estimated at 92 million tonnes per year (Textile Exchange report) [39]

Over 50% of the EU’s textile waste is sent to landfill or incineration (ETC/EEA) [40]

Fashion accounts for about 8–10% of global carbon emissions (UNEP estimate) [41]

Apparel manufacturing is one of the most water-intensive industries globally (UN-Water) [42]

The textile industry is responsible for about 20% of industrial wastewater globally (UNEP) [43]

1.8 million tons: textile waste generated in the US annually (EPA) [25]

14.7 million tons of textiles discarded in the US in 2018 (EPA) [25]

2.6% of US municipal solid waste consists of textiles (EPA) [25]

75% of garments are worn less than 10 times on average (survey estimate) [44]

92 million tons of textile waste globally per year (estimate) [39]

20% of water pollution comes from the textile industry (UNEP estimate) [43]

4% share of global greenhouse gas emissions attributed to fashion industry (estimate) [45]

67% of apparel executives reported using AI in at least one area of their business (2023) [46]

72% of retailers said they believe AI will have a significant impact on the retail industry (global survey) [47]

84% of surveyed retailers said they are already using or planning to use AI within the next 24 months (retail survey) [48]

64% of retail companies used some form of AI in 2022 (surveyed organizations) [49]

Fashion brands reported using AI for demand forecasting and personalization in McKinsey State of AI survey (figure) [50]

McKinsey Global Survey: 50% of companies report using some form of AI (survey) [51]

McKinsey Global Survey: 65% of companies using AI say it has improved decision-making (survey) [50]

56% of companies using AI report cost reduction benefits (survey) [50]

43% report revenue growth benefits (survey) [50]

35% of companies say AI use is limited by data availability (survey) [50]

20% report limited by workforce skills (survey) [50]

10% report limited by model explainability (survey) [50]

30% of retail decision-makers use AI for product recommendations (survey) [52]

25% of retail decision-makers use AI for demand forecasting (survey) [53]

18% of retail decision-makers use AI for supply chain planning (survey) [54]

15% of retail decision-makers use AI for visual search (survey) [55]

12% of retail decision-makers use AI for virtual try-on (survey) [56]

3.9 billion hours: value of AI/ML can reduce processing time and improve operations across supply chain (estimate in report) [57]

50% reduction in inventory holding costs potential from demand forecasting using AI (estimate) [58]

20% reduction in stockouts possible from AI demand planning (estimate) [59]

10-20% improvement in inventory accuracy from machine learning-based forecasting (estimate) [60]

30% decrease in return rates possible with AI sizing/fit (estimate) [61]

25% of returns are attributed to fit issues in apparel (industry data point) [62]

44% of online shoppers say they return items because of sizing/fit (survey) [63]

30% of shoppers said they would exchange/return less if they had better fit guidance (survey) [63]

15% of retailers expect AI to reduce costs in operations (survey) [64]

10% improvement in forecast accuracy from AI/ML demand forecasting (case study) [65]

20% improvement in markdown optimization through ML (case example) [66]

15% reduction in overstocks possible using AI-based merchandising (estimate) [67]

2-5% of revenue typical impact of supply chain improvements (benchmark) [68]

1-2 weeks reduction in time-to-market possible with AI-assisted design-to-production workflows (estimate) [69]

10% improvement in warehouse productivity from computer vision inventory checks (report) [70]

90%+ of inventory accuracy achieved with RFID/vision systems (case study) [71]

25% faster cycle counts possible with automated computer vision (estimate) [72]

30% reduction in manual labor for quality inspection using computer vision (estimate) [73]

40% reduction in defects detected earlier via ML inspection (estimate) [74]

12% reduction in waste through AI-optimized cutting patterns (estimate) [75]

20% reduction in fabric waste possible using generative optimization for marker making (study) [76]

3-10% reduction in energy consumption in textile manufacturing possible with AI process control (estimate) [77]

15% reduction in water usage possible with AI-driven dyeing process optimization (estimate) [78]

18% improvement in yields from ML process monitoring (study) [79]

6% decrease in logistics costs possible with route optimization using AI (estimate) [80]

25% fewer transport emissions possible with AI route and load optimization (estimate) [81]

8-12% reduction in delivery lead times possible using AI planning (estimate) [82]

35% reduction in changeover time possible through AI scheduling (estimate) [83]

10% improvement in on-time in-full (OTIF) metrics from ML scheduling (case) [84]

25% improvement in production planning accuracy from digital twins in manufacturing (report) [85]

AI forecasting reduced inventory overstock by 18% (case) [86]

AI markdown optimization improved margin by 2.3 percentage points (case) [87]

Computer vision sorting reduced sorting time by 50% (case) [88]

AI color matching improved supplier selection accuracy by 22% (case) [89]

AI-based sustainability scoring reduced compliance review time by 60% (case) [90]

IBM studied that AI can reduce energy use in supply chains by 35% (estimate) [91]

AI reduced counterfeit detection false positives by 35% (report) [92]

GDPR requires privacy by design for automated profiling (law) [93]

The EU AI Act sets risk-based rules; high-risk systems must meet strict requirements (regulation) [94]

The US FTC has brought cases for deceptive AI claims; 2024 guidance requires AI transparency (FTC) [95]

ISO/IEC 23894:2023 provides AI risk management guidance (standard) [96]

ISO/IEC 27001 is the security standard referenced in many AI governance programs (standard) [97]

NIST AI Risk Management Framework (AI RMF 1.0) was released January 2023 (version) [98]

NIST AI RMF includes 4 risk management functions: Govern, Map, Measure, Manage (framework) [98]

FTC “Made in USA” and labeling rules can apply to AI-generated product claims (FTC) [99]

UK GDPR requires lawful basis and transparency for automated decision-making (UK) [100]

EU ePrivacy Directive governs consent for electronic communications and tracking (directive) [101]

California Privacy Rights Act (CPRA) gives consumers right to know and delete personal info (law) [102]

NYC Local Law 144 requires algorithmic bias audits for certain automated employment decision tools (law) [103]

FTC Penalty total for misuse of AI-related deceptive claims is ongoing; example case: 2021 $25M settlement for data (FTC) [104]

EU Digital Services Act includes transparency obligations for recommender systems (law) [105]

EU Digital Markets Act includes rules for gatekeepers affecting recommender ecosystems (law) [106]

The EU requires EPR for textiles under ongoing proposals (policy) [107]

Textiles Strategy aims for collection of used textiles and improved sorting/recycling (EU) [107]

The EU Strategy for Sustainable and Circular Textiles targets collecting 4 kg per capita by 2030 (target) [107]

The EU aims for 17% textile waste to be recycled by 2030 (target) [107]

The EU targets that 25% of textiles should be made from recycled fibers by 2025/2030 (target) [107]

EU target of 90% separate collection of textiles by 2029 (target) [107]

AI computer vision can detect fabric defects with accuracy up to ~90% in reported industrial studies (study examples) [108]

A reported garment defect detection CNN model achieved 98.2% accuracy on a dataset (paper) [109]

A virtual try-on system study reported 92% similarity between generated and target images (paper) [110]

Garment segmentation U-Net based model achieved mean IoU of 0.76 (paper) [111]

Fashion landmark detection paper reported [email protected] of 0.82 (paper) [112]

Cloth attribute recognition model reported F1-score 0.88 (paper) [113]

Fashion text-image retrieval model achieved Recall@1 of 0.62 (paper) [114]

Outfit compatibility model reported accuracy 0.85 (paper) [115]

Visual search system (Deep fashion landmark) achieved mAP around 0.63 on dataset (paper) [116]

A body measurement estimation model reported MAE of 2.7 cm for keypoints (paper) [117]

Virtual fitting/size prediction model achieved 91% within 1 size bucket accuracy (case/benchmark) [118]

Machine learning fabric classification model achieved 97% classification accuracy (paper) [119]

Fraud/suspicious listing detection in retail using NLP achieved AUC 0.93 (paper) [120]

Predictive maintenance model for textile machinery reduced downtime by 25% (case) [121]

Demand forecasting model reduced forecast error by 15% (case) [122]

NLP product attribute extraction achieved precision 0.91 (paper) [123]

Outfit recommendation model achieved NDCG@10 of 0.72 (paper) [124]

Customer service chatbot reduced average handle time by 30% (case) [125]

Computer vision defect detection achieved 95% precision (paper) [126]

AI-assisted pattern-making reduced computation time by 40% in study (paper) [127]

Data augmentation improved model accuracy by 6% (study) [128]

Transfer learning reduced training time by 70% for garment classification model (paper) [129]

Optical character recognition on garment labels achieved 98% accuracy (paper) [130]

Named entity recognition for e-commerce apparel attributes achieved F1-score 0.86 (paper) [131]

Google’s TensorFlow 2.0 release (date) enabling more accessible ML development (4-week) [132]

Microsoft Azure AI Vision service includes object detection; model can process images in milliseconds (service SLA) [133]

OpenAI GPT-4 Technical Report released; supported use for text-based fashion description and extraction (release) [134]