Source Registry
Source Registry
Every figure on this site traces back to the sources below. Each entry carries the verbatim quote, publisher, access date, and verification method, with an archived copy in case the original link rots.
115 claims
Amazon has deployed its 1 millionth robot, cementing its position as the world's largest manufacturer and operator of mobile robotics.
“We've just deployed our 1 millionth robot, building on our position as the world's largest manufacturer and operator of mobile robotics.”
Amazon's milestone millionth robot was delivered to a fulfillment center in Japan, joining a global robot network spanning more than 300 facilities worldwide.
“This milestone robot was recently delivered to a fulfillment center in Japan, joining our global network that now spans more than 300 facilities worldwide.”
Amazon says DeepFleet, its generative AI foundation model, will coordinate robot movement across its fulfillment network and improve the robotic fleet's travel time by 10%.
“Called DeepFleet, this AI technology will coordinate the movement of robots across our fulfillment network, improving the travel time of our robotic fleet by 10%”
Amazon's Sequoia robotic system identifies and stores incoming inventory at fulfillment centers up to 75% faster than before.
“Sequoia allows us to identify and store inventory we receive at our fulfillment centers up to 75% faster than we can today.”
Sequoia reduces the time it takes to process an order through a fulfillment center by up to 25%.
“When orders are placed, Sequoia also reduces the time it takes to process an order through a fulfillment center by up to 25%”
As of October 2023, Amazon had over 750,000 robots working collaboratively with employees, taking on highly repetitive tasks.
“We now have over 750,000 robots working collaboratively with our employees, taking on highly repetitive tasks and freeing employees up to better deliver for our customers.”
In 2022, recordable incident rates and lost-time incident rates were 15% and 18% lower, respectively, at Amazon Robotics sites than at non-robotics sites.
“Company data shows that, in 2022, recordable incident rates and lost-time incident rates were 15% and 18% lower, respectively, at Amazon Robotics sites than non-robotics sites.”
Amazon's robotics rollout created 700 categories of new skilled job types that did not exist within the company before.
“This includes 700 categories of new job types, in skilled roles, which didn’t exist within the company beforehand.”
Robots already assist in roughly 75% of Amazon customer orders, according to the company.
“Robots already assist in roughly 75% of Amazon customer orders, according to the company, and that number is likely to rise as the technology advances further.”
75% of Amazon's global deliveries are now assisted in some way by a robot, per Wall Street Journal reporting cited by TechCrunch.
“The WSJ also reported that 75% of Amazon’s global deliveries are now assisted in some way by a robot.”
Coupang deployed more than 1,000 AGVs in its Daegu fulfillment center alone, automating shelving and picking work.
“쿠팡은 이 센터에만 1000대가 넘는 AGV를 도입해 진열·집품 작업을 자동화했다.”
According to the Daegu FC center head, automation technology reduced work intensity by 65% and eliminated human error.
“박 센터장은 “자동화 기술 덕분에 업무 강도가 65% 감소하고 '휴먼 에러'도 사라졌다”고 말했다.”
At Daegu FC, a workload that used to take a worker 3 hours can now be processed in 1 hour thanks to logistics robots.
“쿠팡 대구FC는 물류센터를 누비는 다양한 로봇 덕분에 업무량은 줄고 생산성은 늘었다. 기존 작업자가 3시간 일할 업무량을 1시간 만에 처리 가능하다.”
Thanks to AGVs eliminating worker movement, Coupang cut overall work steps by 65% and delivers items to workers within 2 minutes on average.
“사람의 이동을 없앤 AGV의 활약 덕에 쿠팡은 전체 업무 단계를 65% 줄이고, 평균 2분 안에 상품을 작업자에게 전달할 수 있게 됐다.”
Coupang has invested about KRW 6.2 trillion in its Rocket Delivery logistics network since founding, of which KRW 320 billion went into the Daegu FC.
“쿠팡은 창립 후 지금까지 로켓배송을 위한 물류망 구축에 약 6조2000억 원을 투입했고, 그중 이곳 대구 FC에 3200억 원을 쏟아부었다.”
Coupang invested KRW 320 billion in building the Daegu FC automation system.
“쿠팡은 대구FC 자동화 시스템 구축에 3200억원을 투자했다. 센터 전체가 일종의 거대한 미래형 물류 테스트베드다.”
Daegu FC created about 2,500 new jobs centered on technical roles such as automation system managers.
“무인 기술이 고용을 줄일 것이라는 통념도 깼다. 쿠팡 대구FC는 자동화 시스템 관리자 등 기술 직군 중심으로 2500여명 신규 고용을 창출했다.”
Lotte Mart says that since introducing AI-based fruit sorting, it has kept the defect rate of AI-sorted fruit within 0.01% of sales volume, an improvement of about 30% versus before adoption.
“AI 선별 도입 이후 불량률은 판매량 대비 0.01% 이내로 관리되고 있다. 도입 이전보다 30%가량 개선됐다고 롯데마트는 강조했다.”
In 2025, Lotte Mart's sales of AI-sorted fruit grew more than 7x compared with 2022, the first year of adoption.
“올해 AI 선별 과일 매출은 2022년 도입 첫해 대비 7배 이상 성장했다. AI 선별 도입 이후 불량률은 판매량 대비 0.01% 이내로 관리돼, 도입 이전 대비 30%가량 개선됐다.”
From January to August 2025, Lotte Mart's sales of high-Brix (high-sweetness) fruit roughly doubled year over year.
“차별화된 고품질 과일이 고객에게 호응을 얻으며, 올해 1~8월 롯데마트의 고당도 상품 매출은 전년 대비 약 2배 증가했다.”
Lotte Mart combined deep-learning analysis with non-destructive near-infrared Brix sorters, starting with melons in 2022 and expanding AI-sorted products to 9 fruit varieties.
“롯데마트는 비파괴 당도선별기에 딥러닝 기반 분석 기능을 결합한 AI 선별기법을 도입했다. 내부 갈라짐과 익은 정도, 수분 함량 등 세부 품질 요소까지 정밀하게 선별할 수 있다. 2022년 멜론을 시작으로 현재 9종의 과일에 AI 선별 상품을 운영 중이다.”
Lotte Mart reported that over the 12 months from June 2025 through May 2026, its high-Brix premium fruit sales rose about 20% year over year.
“롯데마트는 최근 1년(지난해 6월1일~지난 5월31일)간 고당도 프리미엄 과일 매출이 전년 동기 대비 약 20% 증가했다고 11일 밝혔다.”
Premium products' share of Lotte Mart's total fruit sales expanded more than 3x, from about 6% four years ago to about 20%.
“전체 과일 매출에서 프리미엄 상품이 차지하는 비중은 약 20%로 4년 전(약 6%)보다 3배 이상 확대했다.”
Lotte Mart's Busan Customer Fulfilment Centre (CFC), built on Ocado automation, plans to handle over 30,000 deliveries per day using AI demand forecasting and robotics, more than doubling throughput versus its existing online fulfilment centers.
“AI 기반 수요예측과 재고관리, 로봇 자동화로 하루 3만건 이상 배송을 처리하고 기존 온라인 물류센터보다 처리량을 2배 이상 높일 계획이다.”
Ocado's OSP robotic technology enables retail partners to pick a 50-item grocery order in just five minutes.
“Our robotic technology enables our partners to pick 50-item orders in just five minutes.”
An AI 'air traffic' control system orchestrates the bot fleet in Ocado CFCs, communicating with each bot 10 times per second while bots move at speeds of up to 4m/s.
“An AI ‘air traffic’ control system orchestrates the movement of the fleet, communicating with each bot 10 times per second to ensure peak performance. Moving at speeds of up to 4m/s, the bots collaborate to transport containers”
Ocado's last-mile routing system makes 100 million optimisation calculations per second to optimise deliveries in real time.
“We make 100m optimisation calculations per second to keep things running smoothly.”
Ocado says its end-to-end platform delivers food waste of just 0.4% of sales, versus 2.5%-3% for most supermarket groups.
“delivering metrics such as food waste of just 0.4% of sales versus 2.5%-3% for most supermarket groups.”
At Ocado's largest site, the $310 million CFC in Erith, more than 2,000 robots move across the grid at four metres a second, replacing humans who once retrieved groceries.
“At its largest site, the $310 million building in Erith, southeast London, more than 2,000 robots whizz around the steel and aluminium grid covering four metres a second”
The Erith CFC's productivity is already 25% higher than Ocado's first British CFC, with scope for further gains of 50-70% according to Ocado Solutions CEO Luke Jensen.
“Erith's productivity is already 25% higher than that of Ocado's first British CFC and Jensen sees scope for further gains of 50-70%.”
Ocado's machine-learning robotic picking arm could already handle 15% of the product range in 2021, with the company expecting around 50% within two years and 80% feasible longer term.
“The arm can already handle 15% of Ocado's product range, a figure the company expects to reach around 50% in two years, with 80% "absolutely feasible" in the longer term.”
Ocado states its fulfilment centres are already among the most labour-productive sites in the world, with a 50-item order picked in a matter of minutes and hundreds of orders picked simultaneously.
“Ocado’s fulfilment centres are already among the most labour productive sites in the world, with a 50-item order picked in a matter of minutes and hundreds of orders picked simultaneously.”
OLIVE YOUNG's in-house item-similarity language model (ItemLM), trained on user behavior data, improved recommendation click-through rate (CTR) by about 50% compared with the previous model.
“속성 미보유 상품과 카테고리에 대한 추천 커버리지를 100%로 확대 그리고 관련된 추천 화면의 실질적인 비즈니스 성과(기존 모델 대비 CTR 약 50% 개선)를 모두 달성했다는 점은 이번 프로젝트의 핵심 성과로 꼽을 수 있습니다”
The ItemLM model expanded recommendation coverage to 100% for products and categories lacking attribute data, which the previous attribute-based recommender could not cover.
“속성 미보유 상품과 카테고리에 대한 추천 커버리지를 100%로 확대 그리고 관련된 추천 화면의 실질적인 비즈니스 성과(기존 모델 대비 CTR 약 50% 개선)를 모두 달성했다는 점은 이번 프로젝트의 핵심 성과로 꼽을 수 있습니다”
OLIVE YOUNG obtained roughly 3 million training labels for ItemLM from one year of search logs, using 20% of them as validation data.
“결과적으로 1년의 검색 로그로부터 약 300만개 정도의 학습 레이블을 확보할 수 있었고, 이 중 20%를 검증용 데이터로 사용 했습니다”
OLIVE YOUNG launched an app service in which its self-developed small LLM (sLLM), trained on some 40 million customer reviews, generates new recommendation themes daily and curates matching products.
“4000만개에 달하는 소비자 리뷰를 학습한 자체 AI 모델이 매일 새로운 주제를 생성하고, 주제에 맞는 상품을 선별해 보여준다”
The sLLM extracted about 100,000 expressions from review data and, based on them, generates new themes daily to recommend up to 15 products per theme.
“현재 리뷰 데이터를 기반으로 약 10만개 표현을 추출했다. 이를 기반으로 매일 새로운 테마를 생성해 주제별 맞춤 상품을 최다 15종까지 추천한다”
The share of quick-commerce 'Oneul Dream' (same-day 3-hour delivery) orders in OLIVE YOUNG's total online-mall orders grew from 24.5% in 2021 to 41.3% in 2023.
“올리브영 전체 온라인몰 주문 건수 중 오늘드림 비중은 2021년 24.5%에서 지난해 41.3%까지 커졌다”
As of September 2024, OLIVE YOUNG's own membership had about 15 million members, well past the 10 million mark of 2021.
“올리브영 올해 9월 기준 자체 멤버십 회원은 2021년 1000만명을 훌쩍 넘어선 약 1500만명이다”
OLIVE YOUNG's online revenue, which hovered around KRW 60 billion in 2017, surpassed KRW 1 trillion in 2023.
“2017년 600억원 수준에 머물렀던 올리브영 온라인 매출은 지난해 1조원을 돌파했다”
Walmart's Self-Healing Inventory AI system, which automatically reroutes overstocked supply to stores that need it, has by itself already saved Walmart more than $55 million.
“That one system alone has already saved Walmart more than $55 million.”
Walmart cut unit costs by 20% year over year with its next-generation automated fulfillment centers compared to manual sites, per John Furner, president and CEO of Walmart US.
“Walmart cut unit costs by 20% year over year reduction with its next-generation automated fulfillment centers compared to manual sites, according to John Furner, president and CEO of Walmart US.”
By the end of 2025, Walmart expects its automated centers to drive an over 30% improvement in cost reduction across its network.
“By the end of 2025, the retailer expects automated centers will drive an over 30% improvement in cost reduction across its network”
By the end of 2025, Walmart expects 65% of its stores to receive merchandise from high-tech distribution centers.
“By the end of this year, the retailer expects 65% of its stores to receive merchandise from high-tech distribution centers, according to an October email to Supply Chain Dive”
At its 2023 investor meeting, Walmart projected that by the end of fiscal year 2026 roughly 65% of stores would be serviced by automation, about 55% of fulfillment center volume would move through automated facilities, and unit cost averages could improve by approximately 20%.
“By the end of fiscal year 2026, Walmart believes roughly 65% of stores will be serviced by automation, approximately 55% of its fulfillment center volume will move through automated facilities, and unit cost averages could improve by approximately 20%.”
Walmart plans to implement Symbotic's robotics and software automation platform in all 42 of Walmart's regional distribution centers over the coming years.
“implement Symbotic’s robotics and software automation platform in all 42 of its regional distribution centers over the coming years.”
Precedence Research estimates the global AI-in-retail market at USD 14.03 billion in 2025 and projects it to reach about USD 70.95 billion by 2035, implying a 17.60% CAGR for 2026-2035.
“The global artificial intelligence (AI) in retail market size is calculated at USD 14.03 billion in 2025 and is predicted to increase from USD 16.64 billion in 2026 to approximately USD 70.95 billion by 2035, expanding at a CAGR of 17.60% from 2026 to 2035.”
88% of organizations report regular AI use in at least one business function, up from 78% a year earlier (McKinsey Global Survey on the State of AI, Nov 2025).
“88 percent report regular AI use in at least one business function, compared with 78 percent a year ago.”
Despite widespread AI use, most organizations remain in experimenting or piloting stages, with only about one-third having begun to scale AI programs (McKinsey, Nov 2025).
“But at the enterprise level, the majority are still in the experimenting or piloting stages (Exhibit 1), with approximately one-third reporting that their companies have begun to scale their AI programs.”
In BCG's October 2024 survey of 1,000 CxOs and senior executives, 74% of companies had yet to show tangible value from their use of AI.
“Seventy-four percent of companies have yet to show tangible value from their use of AI.”
95% of enterprise generative AI pilots are failing to deliver measurable P&L impact, per MIT NANDA's 'The GenAI Divide: State of AI in Business 2025' report.
“The 95% failure rate for enterprise AI solutions represents the clearest manifestation of the GenAI Divide,” the report states.”
In a Feb. 17-19, 2025 survey of employees at 396 Korean trading companies cited in KITA's June 2025 report, 78.0% said AI adoption was necessary for efficiency, while 16.9% said they used AI to improve productivity or perform work actively.
“조사 기업의 78.0%는 효율성 향상을 위해 AI 도입이 필요하다고 답했다. 그러나 실제 AI를 활용해 생산성을 향상시키거나 능동적으로 업무 수행을 하는 기업은 16.9%에 그쳤다.”
MIT NANDA reports that despite $30-40 billion in enterprise GenAI investment, 95% of organizations are getting zero return (MIT NANDA, 2025).
“Despite $30–40 billion in enterprise investment into GenAI, this report uncovers a surprising result in that 95% of organizations are getting zero return.”
For enterprise-grade AI systems, 60% of organizations evaluated such tools, only 20% reached pilot stage, and just 5% reached production (MIT NANDA, 2025).
“Sixty percent of organizations evaluated such tools, but only 20 percent reached pilot stage and just 5 percent reached production.”
Gartner predicts at least 30% of generative AI projects will be abandoned after proof of concept by end of 2025, due to poor data quality, inadequate risk controls, escalating costs or unclear business value.
“At least 30% of generative AI (GenAI) projects will be abandoned after proof of concept by the end of 2025, due to poor data quality, inadequate risk controls, escalating costs or unclear business value, according to Gartner, Inc.”
Gartner: GenAI deployment approaches for business model transformation come with significant costs ranging from $5 million to $20 million.
“However, these deployment approaches come with significant costs, ranging from $5 million to $20 million (see Figure 1).”
Only 26% of companies have developed the capabilities needed to move beyond proofs of concept and generate tangible AI value; 74% have yet to show tangible value (BCG, Oct 2024, survey of 1,000 CxOs and senior executives across 59 countries).
“only 26% of companies have developed the necessary set of capabilities to move beyond proofs of concept and generate tangible value, according to new research by Boston Consulting Group (BCG).”
AI maturity distribution: just 4% of companies have cutting-edge AI capabilities across functions consistently generating significant value, and an additional 22% are beginning to realize substantial gains (BCG, 2024).
“While just 4% of companies have developed cutting-edge AI capabilities across functions and consistently generate significant value, an additional 22% have implemented an AI strategy, built advanced capabilities, and are beginning to realize substantial gains.”
AI leaders follow the 10-20-70 rule: 10% of resources into algorithms, 20% into technology and data, and 70% into people and processes (BCG, 2024).
“Leaders follow the rule of putting 10% of their resources into algorithms, 20% into technology and data, and 70% in people and processes.”
Over the past three years, AI leaders achieved 1.5x higher revenue growth, 1.6x greater shareholder returns, and 1.4x higher returns on invested capital than other companies (BCG, 2024).
“Over the past three years, AI leaders have achieved 1.5 times higher revenue growth, 1.6 times greater shareholder returns, and 1.4 times higher returns on invested capital.”
88% of organizations report regular AI use in at least one business function (up from 78% a year ago), but at the enterprise level the majority are still in experimenting or piloting stages (McKinsey State of AI, Nov 2025).
“88 percent report regular AI use in at least one business function, compared with 78 percent a year ago. But at the enterprise level, the majority are still in the experimenting or piloting stages”
Nearly two-thirds of organizations have not yet begun scaling AI across the enterprise (McKinsey State of AI, Nov 2025).
“Most organizations are still in the experimentation or piloting phase: Nearly two-thirds of respondents say their organizations have not yet begun scaling AI across the enterprise.”
Just 39% of respondents report enterprise-level EBIT impact from AI, even though 64% say AI is enabling their innovation (McKinsey State of AI, Nov 2025).
“Respondents report use-case level cost and revenue benefits, and 64 percent say that AI is enabling their innovation. However, just 39 percent report EBIT impact at the enterprise level.”
More than two-thirds of leaders say 30% or fewer of their GenAI experiments will be fully scaled in the next 3-6 months (Deloitte State of Generative AI in the Enterprise Q4, survey of 2,773 leaders, Jan 2025).
“More than two-thirds of respondents say that 30% or fewer of their experiments will be fully scaled in the next three to six months, revealing that no matter how quickly the technology advances—or how hard the companies producing GenAI technology push—organizational change in an enterprise can only happen so fast.”
69% of leaders say fully implementing an AI governance strategy will take over a year to resolve (Deloitte State of Generative AI in the Enterprise Q4, Jan 2025).
“Sixty-nine percent of respondents say fully implementing a governance strategy will take over a year to resolve, underscoring the need for perseverance and a strategic approach to setting the appropriate governance foundation.”
Korea International Trade Association survey: while consensus on the need for AI transformation is spreading among Korean exporters, only 17.9% of export manufacturers actually use AI on the shop floor (KITA, Apr 2026).
“무역협회 조사에 따르면 수출기업 전반에서 AI 전환 필요성에 대한 공감대는 확산하고 있으나, 실제 제조 현장에서 AI를 활용하는 기업은 17.9%에 그쳤다.”
In a KITA-reported survey, companies cited difficulty identifying AI adoption areas and processes (41.6%) and lack of basic data infrastructure for AI adoption (39.1%) as reasons AI adoption was delayed (Apr 2026).
“기업들은 AI 도입이 지연되는 이유로 ▲도입 영역과 공정 파악의 어려움(41.6%) ▲데이터 인프라 미비(39.1%) 등을 꼽았다.”
In a KITA survey of 396 Korean export companies, 8 in 10 recognized the need to adopt AI, but only about 17% actively use AI in their work (Sept 2025).
“수출기업 396개 사를 대상으로 진행한 설문조사 결과 10곳 가운데 8곳이 AI 도입 필요성을 인지하고 있으나, 응답기업의 17% 정도만이 AI를 활용해 능동적으로 업무를 수행하고 있는 상황이라고 전했다.”
The AI Voucher program provides demand-side companies vouchers of up to KRW 200 million to purchase AI solutions and pays suppliers for the AI solutions sold.
“수요기업에게 최대 2억원의 바우처를 지급하여 원하는 AI 솔루션을 구매할 수 있도록 지원하고, 공급기업에게는 판매한 AI 솔루션에 대한 대금을 지급하는 사업입니다.”
Amazon's DeepFleet generative AI foundation model improves the travel time of its robotic fleet by 10%.
“Called DeepFleet, this AI technology will coordinate the movement of robots across our fulfillment network, improving the travel time of our robotic fleet by 10%”
DHL Supply Chain's AI pick-routing software IDEA reduced warehouse employee travel distance by up to 50% and increased productivity by up to 30% in first deployments.
“In its first deployments, the software's modeling has reduced the distance traveled by warehouse employees up to 50% while increasing productivity up to 30%.”
Korea Customs Service says its AI-based HS-code item classification currently has an accuracy of 80% or higher.
“관세청 관계자는 "현재 AI 품목분류 정확도는 80% 이상이다. 80%인 것은 수입이 적은 물품은 데이터도 적기 때문에 정확도가 떨어질 수밖에 없는 것"이라고 말했다.”
CJ Logistics' AI-powered Cargo Tracking system raised maritime cargo arrival prediction accuracy from 40% to 85%.
“카고 트레킹 시스템은 항해 정보, 경로, 날씨를 비롯하여 화물선의 경로상 위성위치확인시스템(GPS) 정보의 유무, 화물선 이동거리 등 변수들을 적용해 분석, 도착 일시를 예측해 줍니다. 이를 통해 해상운송의 특성상 정확도 40%에 불과했던 예측이 85%까지 향상됐습니다.”
CJ Logistics stated that lowering safety stock by 30-40% versus before through improved arrival prediction made it possible to expect lower storage-related logistics costs, manufacturing schedule planning, and prevention of overproduction.
“그 결과 도착이 늦어질 경우 대비한 안전재고를 기존 보다 30~40% 낮춤으로써 보관 관련 물류비 절감, 생산 공장의 제조 일정 수립, 과잉 생산 방지 효과를 기대할 수 있게 되었습니다.”
RPA software robots replacing repetitive manual tasks across all business units saved 5,600 hours, with about 30,000 additional hours expected.
“소프트웨어 로봇이라 불리는 RPA가 사람이 하던 단순 반복 업무를 대신 수행한 결과, 무려 5,600시간을 단축시켰습니다. 전 사업 부분을 대상으로 RPA 추가적용 대상 업무를 선정 중이며, 이후 약 3만 시간을 더 절감할 것으로 예상됩니다.”
System integration time between online shopping malls and the logistics provider was cut to one-tenth.
“이 외 온라인 쇼핑몰과 물류사 간의 시스템 연동 시간도 10분의 1로 단축시켰으며, 부분 출고, 백오더 등의 기능도 도입해 편의성을 높였습니다.”
CJ Logistics deployed 'PackCheck', an AI-based excessive-packaging diagnosis solution developed by its TES Logistics Technology Institute, across 26 logistics centers nationwide.
“CJ대한통운은 TES물류기술연구소가 개발한 AI 기반 과대포장 진단 솔루션 '팩체크(PackCheck)'를 전국 26개 물류센터에 도입해 운영 중이라고 9일 밝혔다.”
CJ Logistics' 3D-vision AI depalletizer robot arm recognizes irregularly stacked boxes in real time and processes 700 boxes per hour, lifting boxes up to 20kg.
“특히 3D 카메라 비전 기술을 탑재한 '로봇 팔(디팔레타이저)'는 팔레트에 제각각 쌓인 상자를 실시간으로 인식해 시간당 700상자를 처리한다. 최대 20kg 무게의 상자도 특수 패드와 진공흡착 기술로 안정적으로 들어 올린다.”
In its first commercial deployments, DHL Supply Chain's AI-based IDEA algorithm reduced the distance traveled by warehouse employees by up to 50%.
“In its first commercial deployments, the solution has reduced distance traveled by warehouse employees by up to 50% and increased productivity of individual DHL locations by up to 30%.”
The IDEA algorithm increased productivity of individual DHL warehouse locations by up to 30% in its first commercial deployments.
“In its first commercial deployments, the solution has reduced distance traveled by warehouse employees by up to 50% and increased productivity of individual DHL locations by up to 30%.”
DHL's international Vision Picking (AR smart glasses) trials showed an average productivity improvement of 15 percent, with higher accuracy rates.
“The international trials have shown an average improvement of productivity by 15 percent and higher accuracy rates. The user-friendly and intuitive solution has also halved onboarding and training times.”
DHL Supply Chain's completed international AR smart glasses trials cut onboarding and training times by 50%.
“The international AR smart glasses trials completed by DHL Supply Chain in addition to showing increased productivity accuracy, cut onboarding and training times by 50%.”
Smart glasses were successfully piloted at DHL Supply Chain in early 2015 before becoming a standard in DHL's warehouse operations.
“Smart Glasses have been very successfully piloted at DHL Supply Chain in early 2015 before making them a standard in today's warehouse operations with DHL.”
Independent trade press corroboration: DHL's IDEA pick-routing software reduced warehouse employee travel distance up to 50% and increased productivity up to 30%, with results varying by location depending on existing WMS capabilities.
“In its first deployments, the software’s modeling has reduced the distance traveled by warehouse employees up to 50% while increasing productivity up to 30%. The percentages will vary from location to location depending on existing WMS capabilities.”
KOTRA's Nairobi office used TriBIG's overseas-buyer import-probability prediction model to market to 1,100 buyers with a 70%+ import probability; of 87 interested buyers given focused export support, 5 local buyers signed export deals totaling USD 1.45 million.
“수입확률이 70% 이상인 1100개의 바이어 대상으로 온·오프라인 전방위 마케팅을 진행한 것이다. 이 중 우리기업 제품에 관심을 보인 87개에 집중 수출 지원을 한 결과 1차로 현지 바이어 5개사로부터 총 145만달러의 수출 성약을 내는 쾌거를 이뤘다.”
KOTRA analyzed 7.7 million import/export records across 21 countries and used AI/machine learning to predict buyers with a 70%+ probability of importing, then matched them with Korean companies.
“우선 자체 확보한 실거래 데이터와 관련해서는 21개국 770만 건의 수출입 데이터를 분석해 해외바이어의 수입 수요를 예측했다. AI를 활용해 경쟁국으로부터 유사 품목을 수입하고 있는 바이어 중 수입 확률이 70% 이상 되는 바이어를 예측했고 이를 국내기업과 매칭시키는 방식이다.”
Through the AI pilot, KOTRA's Nairobi office recommended a buyer that had never imported Korean products to polypropylene exporter K, winning an export contract worth about USD 900,000.
“시범사업에 적극 참여한 케냐 나이로비무역관에서는 폴리프로필렌을 수출하는 국내 K사에 한국 제품을 한 번도 수입해 본 적 없는 바이어를 추천해 90만 달러 상당의 수출계약을 따냈다.”
KOTRA's self-collected overseas company database exceeds 10 million records, and the agency plans to keep expanding data collection and analysis through external data linkage.
“현재 KOTRA가 자체 확보한 해외기업 데이터는 1000여만건에 이르며 앞으로도 외부 데이터와 연계하는 등 데이터 수집과 분석의 범위를 더욱 넓혀 나갈 예정이다.”
Paint maker A, stalled in Latin American market entry, used TriBIG to find a Peruvian buyer and achieved a USD 30,000 initial export.
“A사는 중남미 시장개척이 지지부진하던 상황 속에 트라이빅 덕분에 페루 바이어를 찾고 3만달러 초도수출에 성공한 경우로, 내년 1월에도 추가 수출 논의를 위해 현지를 방문할 예정이다.”
TriBIG is a trade-investment big data platform that analyzes roughly 150 million data records collected from 128 overseas KOTRA offices across 83 countries.
“'트라이빅'은 국제무역통계 데이터, 해외시장 정보 등 전 세계 83개국의 해외무역관 128개소에서 수집한 약 1억 5,000만 건의 빅데이터를 분석해 서비스하는 무역투자 빅데이터 플랫폼이다.”
Metal forging manufacturer M used TriBIG to identify the U.S. as a promising market and, after steady correspondence with a potential buyer found there, signed an export contract worth USD 83,415.
“금속 단조제품을 생산하는 M사는 트라이빅을 활용해 자사에 적합한 유망시장으로 미국을 추가했고, 여기서 찾은 잠재 바이어와 꾸준한 교신 끝에 최근 83,415달러의 수출계약을 체결했다.”
Maersk Line said remote container management (RCM) devices — GPS unit, modem and SIM card — would be on all 270,000 of its refrigerated containers.
“RCM includes a GPS positioning unit, a modem and a SIM card. These will be on all 270,000 of Maersk’s refrigerated containers. This enables location, temperature, humidity and power status readings to be continuously collected and stored.”
About 400 Maersk Line vessels were outfitted with VSAT domes to relay reefer sensor data to satellite, alongside 270,000 reefers equipped with remote container devices.
“Already, the carrier has equipped 270,000 reefer containers with remote container devices that utilize a 3G SIM card, a GPS unit and an antenna. In addition, about 400 vessels have been outfitted with a VSAT dome, which transmits data gathered from the reefers to a satellite.”
In a single year, Maersk Line eliminated 225,000 pre-trip inspections (PTIs) by using RCM data to determine container condition on arrival, sending reefers for inspection only when needed.
“Last year alone, Maersk Line eliminated 225,000 PTIs, which helped provide faster turnaround of its reefers for customers.”
Time spent per pre-trip inspection fell from up to six hours to about 12 minutes, because RCM data was available before the reefer arrived at the terminal.
“According to Catja Rasmussen, Maersk Line’s project head, the time devoted to conducting PTIs has also been reduced from up to six hours to about 12 minutes because so much data was available to the carrier in advance of the reefer container arriving at the terminal.”
By December 2019, over 3,600 companies had signed up for RCM since its September 2017 customer launch, covering data from over 380,000 refrigerated containers of the combined Maersk and Hamburg Süd fleet.
“Since its launch to customers in September 2017, over 3,600 companies have signed up for RCM and the transparency on information from the over 380,000 refrigerated containers of the combined fleet of Maersk and Hamburg Süd it provides.”
Over the course of 2022, more than 80% of Maersk's reefer fleet was upgraded to deliver hourly data to the Captain Peter platform, compared to a daily update previously.
“Over the course of 2022 more than 80% of the reefer fleet was upgraded to deliver hourly data to Captain Peter - as compared to a daily update previously - enabling even more visibility and reducing the risk of damaged goods and food waste.”
Each year, 13% of global fresh produce is lost in the cold supply chain between harvest and retailer.
“Each year, 13% of the global fresh produce is lost in the cold supply chain, between harvest and retailer, which equivalents millions of Euros in revenue.”
After Samsung SDS applied ML/generative-AI risk sensing to Cello Square, the time to formulate a logistics-risk response plan was cut from one day to two hours.
“방안 수립에 걸리던 시간도 기존 하루에서 2시간으로 단축되어 리스크 대응 속도를 크게 높였다.”
Samsung SDS uses machine learning to automatically extract logistics risks from more than 60,000 global news items collected every day.
“매일 수집한 6만 건 이상의 글로벌 뉴스에서 머신러닝을 활용해 물류 리스크를 자동 추출한다.”
Generative AI grades each extracted risk into three severity levels, using a model built on about 20,000 past global logistics risk cases.
“이렇게 추출된 리스크를 생성형 AI를 활용해 위험도를 3단계로 구분하여 산정한다. 삼성SDS는 과거 약 2만 건의 글로벌 물류 리스크 사례로 위험도를 판단하는 모델을 개발했고, 생성형 AI가 위험도 판단을 할 수 있도록 학습시켰다.”
Cello Square performs generative-AI three-level risk grading on about 70 machine-learning-extracted logistics risks per day, drawn from roughly 60,000 daily global news items.
“글로벌 뉴스 약 6만건을 매일 받아 데이터를 수집하고, 머신러닝 기반 물류 리스크를 추출해 생성형 AI로 3단계 위험도를 판단하는 작업을 하루 70건씩 수행한다”
Cello Square posted Q4 2025 revenue of KRW 330.7 billion, up 27% quarter-on-quarter, with subscribed customers reaching about 24,600 companies and shipping customers up 12%.
“디지털 물류 플랫폼인 첼로스퀘어의 지난해 4분기 매출이 3307억원을 기록하며 전분기 대비 27% 성장했다"며 "디지털 서비스를 창고 및 내륙 운송까지 확대한 효과로 가입 고객사는 2만4600여개사로 늘어났고 선적 고객 수도 12% 증가했다”
Cello Square subscribed customers grew 27% year-on-year to 24,625 companies actively using the service as of end-2025.
“디지털 물류 플랫폼 첼로스퀘어(Cello Square)는 전년 대비 가입 고객이 27% 증가해 2025년 말 기준 24,625개 회사가 서비스를 활발히 이용하고 있다.”
Cello Square provides monitoring services covering 36 countries and 56 hubs worldwide.
“현재 삼성SDS '첼로스퀘어'는 전세계 36개국, 56개 거점에 대한 모니터링 서비스를 제공하고 있다”
UPS's ORION route optimization system saved the company $320 million cumulatively by the end of 2015.
“According to UPS, ORION saved the company $320 million by the end of 2015, but the biggest savings are going forward.”
Once fully deployed, ORION was projected to deliver an estimated $300-400 million in annual savings and cost avoidance.
“Once fully deployed, ORION is expected to account for an estimated $300 million to $400 million in annual savings and cost avoidance”
ORION's projected annual impact includes 100 million fewer miles driven, 10 million gallons of fuel saved, and a 100,000 metric-ton reduction in CO2 emissions.
“thanks to 100 million miles less traveled, 10 million gallons of fuel not consumed and a reduction of 100,000 metric tons in CO2 emissions per year”
ORION cost more than $295 million to build and deploy, against an expected $300-400 million in annual savings (per the UPS-authored, Edelman-verified Interfaces paper).
“Costing more than $295 million to build and deploy, ORION is expected to save UPS $300–$400 million annually.”
UPS's verified full-deployment estimate is an annual reduction of 100 million miles driven, with driver-cost avoidance and fuel savings of $300-400 million per year.
“The current estimate (at full deployment) is an annual reduction of 100 million miles driven with driver-cost avoidance and fuel savings of between $300 million to $400 million annually.”
A UPS driver reported his daily stops rose from 110 to 115-118 in the same working time after adopting ORION routes.
“he used to do 110 stops a day and now does 115 to 118 stops in the same amount of time”
As of December 2015, ORION was in use by more than 35,000 of UPS's 55,000 U.S. drivers.
“As of December 2015, ORION is being used by more than 35,000 of 55,000 U.S. drivers.”
ORION crunches more than 250 million address data points to generate optimized delivery routes for UPS drivers.
“more than 250 million address data points to provide UPS drivers with optimized delivery routes”
In 2021, UPS said the 'dynamic ORION' upgrade enabled an average 2-4-mile route reduction per driver, on top of the original ORION's 8-mile-per-driver reduction.
“reduce driver routes by an average of two to four miles per driver, Perez said, adding that the original ORION implementation resulted in a reduction of eight miles per driver”
UPS's in-house UPSNav navigation runs on proprietary ORION maps containing 250 million locations.
“It uses the UPS data infrastructure and proprietary ORION maps, which include 250 million locations.”