Social distancing has redefined the retail experience and its relationship with technology – in many ways, it could also be the catalyst for more permanent changes. To highlight this, in June 2020, several months after the nationwide lockdowns began in March, U.S. online sales were up 76.2% year over year.  But even with the uptake and investment in expanded eCommerce, several sectors of retail remain hurt. Perhaps most notably, clothing sales have decreased 87%, 62%, and 23% year over year from April through June 2020.  Yet, over the same time frame, food & beverage (i.e. grocery) saw a 12%, 15%, and 12% increase year over year.
As we discuss the future of technology in retail, it becomes clear that eCommerce, while an enabler for growth and expanded reach, is not a replacement for the traditional retail environment. But, given the stricter health & safety guidelines and cost consciousness of cash stressed businesses, there will be continued emphasis and interest of how robots can replace and/or augment humans in retail. However, as we explore the use cases of technology in retail today, it is important to reframe technology as an enabler, not a replacement, for retail. Or more specifically, technology needs to maximize the store conversion, reduce operating costs, and improve the customer experience.
In the following report, we dive into three perspectives of where technology is enabling retail.
- Computer vision AI is creating new seamless ways to identify and monitor the customer
- Drones are reaching new heights both inside and outside the warehouse
- Robots need to enable the store associate to do what they do best – sell
Ongoing advancements in machine learning and both localized and cloud hardware capabilities have opened up new ways to utilize artificial intelligence in today’s retail environment. Specifically, they’ve enabled new ways for us to interact, understand, and utilize big data to better understand, support, and sell to the customer. Two of the burgeoning use cases include:
- Computer vision: Technology using artificial intelligence to analyze and process patterns in images and video enabling new ways to automate, understand, and predict business operations.
- Chatbots: Software applications developed on natural language processors and AI to automate and improve communication without the need of a live agent, often via either a text or voice interface.
Various advancements in camera dynamic range and resolution, computational cameras, real-time detection of moving objects, use of color information, analysis of point clouds, and cloud computing of machine vision have resulted in the increased application and capability of the computer vision industry. The AI in computer vision market expected to grow from $2.37 billion in 2017 to $25.32 billion by 2023, at a CAGR of +48%.  While the market is driven by other burgeoning industries (e.g. automotive and autonomous vehicles), retail applications can be expected to benefit from the R&D in both hardware and software in the space. Today, while only 3% of retailers use any sort of computer vision technology, 40% plan to start or finish implementing the technology within the next two years. 
Computer vision enables a variety of retail applications including:
- Facial recognition
- Cashierless stores
- Inventory visibility
- Visual search
Beyond these, we will focus on two growing fields, often turnkey solutions to build on or replace current retail infrastructure of cameras:
- Customer monitoring
- Security monitoring
As traditional in-store retail slows, there is a growing necessity to maximize conversation the existing foot traffic in stores. Computer vision enables a new sources of customer data:
- Traffic patterns and walk paths to optimize planograms
- Dwell time in front of displays, promotions, and/or shelves
- Instant recognition of loyalty programs or awards
- Demographics including age, gender, and ethnicity
- Patterns of purchases e.g. brand affinity, consumer trends, impulse buys
- Customized and personalized shopping experience for customers
A number of retailers across industries are leveraging this technology. For example, Candy retailer Lolli and Pops uses facial recognition to identify loyalty members. Not only does this ease a barrier of enrollment, it enables store associates to immediately deliver personalized service through taste profiles, identified allergies, and offer curated product recommendations. Similarly, CaliBurger uses self-ordering kiosks. Integrated cameras use facial recognition to quickly identify registered customers, their loyalty accounts, and their order preferences, including historical favorites. 
Many companies are developing AI solutions to rollout across retail – often plug and play into existing camera systems. For example, Graymatics have developed AI that use algorithms to identify specific objects, via the product logo, and customer activity & demographics. Additionally, the technology can be extended to track individual employees and measure how much time it takes for them to service clients.  Similarly, Pilot AI uses computer vision models can detect people and a broad range of objects in a retail setting while performing advanced real-time analytics to develop valuable customer insights.
Total retail shrink is the combination of theft, fraud, and other types of retail losses. In 2018, this was estimated to be about $50.6 billion, or 1.38% of total sales (vs. 1.33% in 2017).  Historically, to combat shrink, retailers have deployed cameras throughout their stores and warehouses, staffed and monitored by a security team. However, manually combing through streams of video and identifying theft is both difficult and operationally costly. Additionally, these teams are often inadequate to catch employee theft. The Retail Theft Survey found that one out of every 40 employees was apprehended for theft in 2018.
While many companies offer computer vision integration into existing camera systems, theft monitoring is often an afterthought to consumer data collection. Additionally, it will be increasingly important to think through “ethical AI” especially as it pertains to real consequences to customers. Percepta is one example of a startup developing an AI that ignores a customer’s race or gender when looking for theft.
Today call volume at the average contact center currently makes up about 45% of interactions with customer service, but by 2022 this is expected to significantly decrease to 14% as customer service shifts from voice to digital. Now, and increasingly so, customers have the option for asynchronous, messaging-based services— a form of customer service that consumers have long preferred. In fact, surveys have found that 9 of 10 consumers would like to be able to use messaging to talk to businesses. 
The changing consumer preferences and the demand for 24/7 customer assistance in the face of pressure for lower operating costs has fueled the growth and development of self-learning chatbots to deliver more human-like conversational experiences. This amalgamates into a market that expected to grow at a CAGR of +29.7%.  While used across a variety of industries, by 2023, over 70% of all chatbots accessed will be retail based. Furthermore, chatbot ecommerce transactions will reach $112 billion annually.
Kore.ai is one of the leading startups in this space. On their website, they advertise up to a 25% increase in customer satisfaction and a 27% reduction in support costs. Thus, the value proposition continues to build for digital retailers to continue to keep customers engaged.
With chatbots, customers can:
- Instantly receive advice from a personal shopper.
- Instantly make purchases directly through the chat portal.
- Receive automated updates on orders, inventory, and pick-ups.
- Utilize flexible customer service channels, e.g. web, mobile, or on the phone.
Established retail has taken note. Back in 2016 H&M launched a chatbot on Canadian messaging app Kik. The chatbot allowed customers to see, share and purchase products from H&M’s catalogue. But more so, the chatbot enabled a personal stylist service as well using photo options and asking questions about shopper’s style. Customers were able to use the bot to create their own outfits and vote & browse outfits created by other users on top of regular ecommerce.
Whole Foods launched a Facebook messenger based chatbot for their grocery stores. Online, customers can enter a zip code or an address or share their location to find their nearest store and its details more quickly. To further engage the audience, they are using the chatbot to enhance the in-store shopping experience. Shoppers can use the chatbot to ask for recipes and to find where products are positioned in the store. For those with special dietary restrictions, the chatbot also helps find foods and meals that leave a particular ingredient out. 
Startups we are watching
The advancement in drone hardware – in terms of cost, runtime, automation, etc. – has enabled a new world of autonomous flight with direct retail applications. While drone based delivery services are often most covered in media, the indoor use of drones, outside of government regulation, is leading the way with present-day applications.
Last mile delivery
Consumers increasingly are demanding more from last-mile delivery as expectations for fast shipping become the norm. Some 20 to 25% of consumers would pick same-day or instant delivery if it were available at low prices – which is a sizeable share compared to roughly 1% of actual volumes today.  However, to get there, last-mile delivery needs to be significantly more efficient. The last-mile leg comprises 53% of the total costs of the average shipment (other estimates put this figure closer to ~30%). In a more specific example, in the case of food delivery, the cost of a delivery person depending on the market is approximately of $8/delivery. Recent estimates indicate that autonomous drones could reduce last-mile delivery costs by ~40% larger by reducing the labor component of delivery.
It’s no surprise then that most funding in logistics startups, around $11.1 billion, was raised by those offering last-mile delivery services to retailers and individuals. Of this amount, $9.9 billion went to startups that rely on unconventional delivery modes, such as crowdsourced delivery, drones, autonomous vehicles, and shipments to parcel lockers. 
The market in general, led by the explosion of ecommerce, has seen a 33% increase in B2B last-mile delivery demand in the last two years, and a 67% lift for B2C businesses over the same time period. Amazon, a large driver and pioneer here, has steadily established a moat for itself by vertically integrating its own supply chain — starting with the launch of Fulfillment by Amazon in 2006. Now, Amazon’s last-mile network handles 3.5B deliveries per year. As a result, all US businesses today that rely on physical delivery to reach their end users are fighting an uphill battle. 
Drones create a new opportunity last-mile delivery by reimagining the business model and its unit economics. For example, in food delivery a restaurant could see sub-5-miunte deliveries cost less than $1.  By operating as a hub-and-spoke system, non-reliant on a network of idle drivers, a restaurant could service its own deliveries through drones going back and forth directly between them and the consumer. The potential of autonomous last-mile drones creates a potential market of $91.5 billion by 2030, growing at a CAGR of +20%. 
Globally, regulatory bodies are slowly beginning to trial various drone delivery services. The ongoing concerns are how to deal with air pollution (both sound and physical space), interference with current air zones, damages & theft, etc. In China, for example, the Civil Aviation Administration of China (CAAC) is permitting SF Holding and JD.com to start sending packages by drone in some of the rural areas. 
Google’s Wing Aviation, rather than Amazon, was the first drone delivery company to gain air carrier certification from the FAA in the United States. Wing is initially restricted to areas of southwest Virginia and operates during the day when the weather is clear enough so that the drones can be seen.  In Reno, Nevada, Flirtey has been working with local first-responders to deliver defibrillators in life-or-death situations.
Beyond tech companies, parcel delivery companies, e.g. UPS, are looking for their own innovations. Specifically, they’ve looked into portable hubs where the drone autonomously delivers a package to a home and then returns to the vehicle while the delivery driver continues along a route to make a separate delivery.
COVID-19 and the recent stay at home orders have help accelerate the demand for autonomous deserves. On March 30th, Virginia Gov. Ralph Northam enacted a stay-at-home order. After the lockdown went into effect, orders through Wing’s delivery app for the town shot up. Additionally, Wing brought on new local businesses, including Brugh Coffee and Mockingbird Cafe, to offer more products to people stuck at home. In their first weekend of drone deliveries, Mockingbird sold 50% more pastries and Brugh twice as much cold brew as they would over a regular weekend. Across all its global delivery locations, Wing saw a roughly 350% month-on-month increase in the number of sign-ups for its service. 
Another startup Zipline, which uses a proprietary fixed-wing drone to deliver medical supplies over massive distances, has been operating in Africa since 2016, primarily in Rwanda and Ghana. Its drones deliver blood and other life-saving medical supplies to community hospitals that are cut off from the medical access found in major cities. Additionally, newer players such as HubVery are experimenting with peer-to-peer business models to make sure drone CAPEX is more palatable and utilized.
While progress is made, FAA and other global regulations remain a hinderance to rapid deployment. Jonathan Rupprecht, an aviation lawyer who focuses on the drone industry, said that it would likely be another “year or two” before the FAA would have figured out its plans for allowing drone operators to fly over people, let alone allowing drones to fly themselves places. 
Flying warehouse workers
While last-mile delivery faces scrutiny from the FAA and other global regulators for growth, indoor drone use is completely fair game – and in warehouses especially. Indoors, companies can take advantage of the autonomous flexibilities of drones to rapidly understand their inventory. More specifically, indoor drones enable solutions for:
- Lost or stolen inventory (via tracking, counting, searching, etc.)
- Labor cost of managing inventories
- Safety, especially as warehouses evolve into smaller, more dense units
They answer two questions: What items do we still have in inventory? And where are they?
With customers demanding shorter and reliable delivery dates, new technologies are essential for transforming traditional warehouses to keep up. For example, in a warehouse with 50,000 storage bays, on average three employees are continually occupied with locating goods. “Drones could reduce the costs of this manual stocktaking by as much as 70 percent,” says Höfle.  In another example, at a Ryder customer warehouse, drones successfully scanned pallets and locations in 20 minutes, compared to a manual scan which took 90 minutes. Additionally, a cycle count on the entire warehouse took just three hours versus two days. With these improvements, drone enable warehouse workers to focus on jobs that are more helpful and productive. 
Startups are helping lead the industry to embrace the new technologies. Ryder has partnered with Fetch Robotics to provide the technologies and power the warehouse-of-the-future. “Our top priority is to help our customers achieve on-demand automation with a fast, flexible and scalable cloud-based platform that minimizes strenuous manual labor and streamlines inventory processes. We’re seeing this come to life successfully through the smart warehouses Ryder has developed.”  Additionally, partnering with established hardware partners has allowed for startups to specialize their software and scale quickly. Ware, for example, has focused on their software solution and partnered with hardware company Skydio. “We partnered with Skydio because, frankly, they make the smartest drone on the planet. Nothing else even comes close to matching it on being able to navigate and follow our instructions indoors, while maintaining flexibility and safety.” 
Startups we are watching
Despite the growing prevalence and importance of ecommerce, the vast majority of retail sales, almost three quarters, occur in-store.  So while technology continues to enable a new digital channel, it also needs to revolutionize our in-store experiences. More specifically, we need to rethink what tasks from our store associates today can be automated so that they can spend more time with the customer.
The effect of store associates on the in-store retail experience has been shown in a variety of surveys. Grail Research found that shoppers are 43% more likely to make a purchase and spend 81% more if they interact with an associate. Conversely, 40% of shoppers report being unable to find an associate when they needed help and leaving because of it. According to PwC, customers will pay up to 16% more for a better customer experience and 75% want more human interaction. Conversely, consumers would stop doing business with a company due to unfriendly service (60%) or unknowledgeable employees (46%).
Even as alternative ways of shopping as changing the retail store, knowledgeable associates and good customer service are essential to maximize the conversion the in-store traffic. With so many different ways to receive a product, consumers are less forgiving to a bad experience, especially millennials. Research from Morning Consult found that “Poor customer service is the easiest way to lose brand loyalty” among Gen Y consumers. The study found 74% of millennials would be less likely to purchase from a brand they’re loyal to if customer service wasn’t up to snuff.
Because so much of the shopping experience depends on the store associates, it is essential they are best prepared and enabled. Surveys from Zebra Technologies indicate that retail store associates are frustrated about being overworked and unable to help customers as effectively as they’d like, with 55% saying they agree that their company is understaffed. More so, 83% of retail decision makers and 74% of store associates believe that customers would have a better shopping experience if employees were equipped with technology — especially handheld mobile computers and barcode scanners. The objective of robots here shouldn’t be to replace the store associate, it needs to be to free the store associate to do what they do best – sales and customer experience.
Where robots should be used
Today, we’ve already implemented robots in our stores. Self-checkout has been a growing trend in retail, especially grocery, and has only been accelerated with social distancing measures. Self-checkout improves convenience and cuts down on wait times for shoppers but most importantly, allows for associates to spend less time with the mindless task of scanning items and taking payment. Associates can now repurpose this time by assisting customers still debating their decision and who might need that last little confidence boost or nugget of information to convince them to buy.  Several brands such as Apple and Nordstrom have taken this a step further. By enabling associates with mobile POS devices, they can immediately make the sale to the customer – both more efficient for the associate and the customer.
New technology is enabling even more tasks. Simbe Robotics’ Tally is an example of how some of the same computer vision technology that was utilized for drones can enable in-store robots. In one case study, Schnuck Markets went live with Tally, the autonomous shelf-scanning robot. The robots traverse the store three times per day, scanning approximately 35,000 products per traversal. On an average day, the Tally robots scans millions of products, giving Schnucks more accurate, frequent, and comprehensive insights into product flow and in-store operations. “Teammates are spending several hours a week completing tasks like inventory scanning and price tag auditing,” said Dave Steck, VP of IT, infrastructure and application development, Schnucks. “Since we implemented Tally, the robot now completes those mundane tasks, allowing teammates to focus on customer service. 
Empowering the store associate
Parallel to freeing up store associates, new software technology is empowering the store associate to provide better service to the customer. A number of startups are reimagining the associate & customer experience.
“Frontline employees…want a digital relationship with their employer as they do in every other relationship,” Steven Kramer, CEO of Workjam says. But, “there’s been no innovation in HR and systems related to frontline employee management. For example, there are still binders in break rooms full of printed materials for training retail store employees.” With Workjam, employees can use their mobile devices for clocking in/out, managing schedules, picking up shifts through an open shift marketplace, training, communication and task management—automating what have always been manual processes. Another workforce management software started off in the beauty industry – AllWork. They utilize many similar end-to-end software solutions to better manage an hourly workforce but also include integrated payroll and compliance.
Technology can also improve the traditional associate-customer store experience. Valtech, a digital marketing agency, worked with The Container Store to create The Organization Studio, an interactive design tool and digital experience that helps customers organize their space.
Originally the tool didn’t involve store associates but, rather, was completely AI driven. Now the store associate can come up with a personalized organization solution for the customer – deepening the customer engagement and customer experience. 
Startups we are watching
Two notable themes emerge from the perspectives above: (i) startups that have entered these spaces over the last few years have grown incredibly and (ii) the asset-heavy nature (e.g. robotics) generally inflates the round sizes. For new ventures, however, the key will be sourcing niche applications, especially for software & asset-lite business models.
Computer vision AI: The market has seen some saturation of various companies promising better customer tracking through existing CCTV systems. The future opportunity lies in (i) niche applications of computer vision such as security monitoring or compliance or (ii) SaaS products that are geared around transforming data into actionable recommendations for retailers.
Drone inside and out: Ware is an example of a software-based solution that can increasingly be competitive in this space. The hardware side of drones is increasingly saturated by incumbents; however, this also means that unique applications are more feasible through software focused companies. Additionally, ventures should also look at business models that are asset-lite, e.g. where the retailers own their own drones and utilize a SaaS product for inventory scanning or delivery.
Enabling store associates: Reimaging the recurring tasks of a store associate aligns with RevTech’s mission for new retail concepts. We have discussed already how new technologies like robots and drones are limiting mundane tasks like counting inventory from the associates. Now, the next wave of new ventures can rethink how technology will enable associates to provide better customer care and sales in the store. Ventures should invest in companies looking to increase the efficiency and level of offering in the store similar to how Apple redefined tech retail, e.g. through POS equipped associates for effortless checkout.
 Markets and Markets
 RIS’ 29th annual retail technology study
 National Retail Federation
 Multi Channel Merchant
 Markets and Markets
 Retail Insight Network
 Dynamo VC
 Dynamo VC
 Dynamo VC
 Research and markets
 Vice President Portfolio & Innovation Management at Körber Supply Chain Automation
 Melonee Wise, CEO of Fetch Robotics
 Ware co-founder and CEO Ian Smith