Artificial Intelligence in Practice
Page 10
4Webtrends, https://www.webtrends.com/about-us/client-success/ kimberly-clark/
5Nielsen, Machine Learning Powered Marketing Personalization Innovation: https://www.nielsen.com/us/en/press-room/2016/machine- learning-powered-marketing-personalization-innovation-unveiled.html
6Tableau, How Kimberly-Clark saved $250k with a platform powered by Tableau, Amazon Redshift, and Panoply: https://www.tableau.com/ about/blog/2018/2/how-kimberly-clark-saved-250k-platform-powered- tableau-amazon-redshift-and-panoply
7KC Lab, http://kcdlab.com/
8Webtrends, https://www.webtrends.com/about-us/client-success/ kimberly-clark/
15
McDonald's: Using Robots And Artificial Intelligence To Automate Processes
McDonald's employs 400,000 staff across its 36,000 restaurants spread over 120 countries.1
At the start of 2017, it unveiled a new strategy2 focused on using digital technology, including artificial intelligence (AI), to drive growth.
Self-service kiosks are increasingly prominent in its outlets, and are often cited as one of the more visible signs of a trend towards replacing human workers with machines.
However, in the company's own words, its main drive is to use technology for “re-shaping our interactions with the customer – whether they eat in, take out, drive thru or order delivery”.
What Problem Is Artificial Intelligence Helping To Solve?
Restaurant chains have to carefully manage their inventories to avoid unsold food going to waste or items not being in stock when customers want them.
Traditional, static menu displays don't offer the flexibility to allow stores to promote items most likely to appeal at different times of the day unless they are manually updated.
On top of that, McDonald's restaurants often attract large volumes of customers at peak times during the day. While this is largely what has made the brand the success it is, dealing with large numbers of people all expecting to be served very quickly and accurately is challenging in its own right.
How Is Artificial Intelligence Used In Practice?
Rather than just being straightforward terminals, those McDonald's self-service kiosks that have sprung up around the world in recent years, as well as the digital menu boards used in stores, are equipped with smart, analytic technology.
This means that decisions can be made autonomously at restaurant level about what to promote and sell to customers. This decision is based on factors including local ordering trends, the weather and what inventory the restaurant needs to shift before it goes out of date.3
The menus can intelligently adapt too, for example, begin to offer more hearty, warming comfort food items when the temperature drops. While on sunny days, salads and ice creams will be given prominence on the displays.
McDonald's also offers an app that lets users create an account and place orders directly from their phone. This gives it access to data that is used to offer exclusive deals to customers it predicts will be interested.
It also allows it to build aggregated datasets showing what the most popular purchases are across different locations among different segmented customer groups.
What Technology, Tools And Data Were Used?
McDonald's self-service kiosks are already installed at all of its restaurants in the United Kingdom and Canada. In the United States it is installing 1,000 self-service kiosks per quarter and they will be in every restaurant by 2020.4
The company has said very little about the exact data that is collected and used, but Intel states that the terminals help the company “collect data and learn about customer behavior”.5
It is likely that it uses this data to predict what customers are likely to buy, based on other orders that fit a similar pattern, both in terms of what was bought, and external factors such as the time of day and the restaurant's location.
McDonald's is far from the only fast food chain switching over to self-service kiosks and it is very likely to be keeping a keen eye on what its competitors are doing with the technology.
For example, KFC in Beijing has partnered with Baidu to fit kiosks with facial recognition technology, which uses AI to predict what menu items a customer is likely to be interested in, based not only on their age and sex, but also their mood. Items that match those ordered by people with similar profiles are displayed most prominently on the screen.6 The chicken chain plans to roll these screens out to 5,000 of their restaurants, and if results are positive McDonald's will be keen to follow in their footsteps.
And it isn't just front-of-house jobs that are being automated in the fast food industry. US chain Caliburger is trialling a robotic kitchen assistant called Flippy, which uses cloud-based AI along with thermal vision that enables it to “see” how well a burger is cooked on the inside.7
What Were The Results?
Customers get to skip lines when they arrive to pick up their order, and benefit from spending less time browsing menus to find the items they want.
In return, McDonald's collects detailed information about their behavior, which it can use to predict future buying patterns.
In Canada where stores were first fitted with the digital, predictive menu boards, sales increased overall by 3 to 3.5% in their first year.8
And in Japan, users of the mobile phone app spend on average 35% more per transaction, which McDonald's attributes to presentation of individually chosen promotions.9
So, what about the social implications? Are their terminals and apps the visible signs of a move by big businesses to replace humans with machines?
Well, if it leads to a sustainable reduction in costs (and it hasn't yet been proven that it will), the answer is unfortunately likely to be yes – in the long term, at least.
It's possible that politicians could step in and stop this from happening or mitigate against its effects. Some seem to have the issue in their sights, for example, Jeremy Corbyn, leader of the UK Labour Party, who has called for a “robot tax”.10
His argument is that corporations benefiting from AI should be obliged to share some of the wealth they create with society, and to absorb some of the cost of the human unemployment they seem set to create.
But overall it doesn't seem to be an issue that is inspiring the strength of public opinion it often takes to get politicians to take up a cause. At least not yet … it's possible that this is a problem the public will be more aware of over time.
Of course, it's only fair to point out that McDonald's own position on the matter is that it is emphatically not replacing human staff with robots.
It says that cashiers whose jobs are made redundant by the machines will be transferred into other roles, primarily front-of-house customer service.11 In theory, this should make their work more rewarding as well as teach them new skills, which will be useful in their own career development.
But then, they would say that, wouldn't they?
Key Challenges, Learning Points And Takeaways
Businesses that are engaging in automation and AI are keen to imply that technology will assist workers rather than replace them.
However, the long-term impact that the spread of intelligent automation into all business sectors will have on human job markets is still unknown.
Notes
1Investopedia, McDonald's vs. Burger King: Comparing Business Models: https://www.investopedia.com/articles/markets/111015/mcdonalds-vs-burger-king-comparing-business-models.asp
2McDonald's, https://corporate.mcdonalds.com/corpmcd/about-us/our-growth-strategy.html
3Food Business News, McDonald's finds flexibility with digital menu boards: https://www.foodbusinessnews.net/articles/7624-mcdonald-s-finds-flexibility-with-digital-menu-boards
4USA Today, McDonald's: You buy more from touch-screen kiosks than a person. So expect more kiosks: https://eu.usatoday.com/story/ money/nation-now/2018/06/07/mcdonalds-add-kiosks-citing-better- sales-over-face-face-orders/681196002/
5Intel, McDonald's And Predictive Analytics: They're Lovin’ It:
https://www.intel.co.uk/content/www/uk/en/it-managers/mcdonalds-predictive-analytics.html
6The Guardian, KFC China is using facial recognition tech to serve customers – but are they buying it?: https://www.theguardian.com/ technology/2017/jan/11/china-beijing-first-smart-restaurant-kfc- facial-recognition
7Live Science, Humans Couldn't Keep Up with This Burger-Flipping Robot, So They Fired It: https://www.livescience.com/61994-flippy-burger-flipping-robot-flops.html
8Food Business News, McDonald's finds flexibility with digital menu boards: https://www.foodbusinessnews.net/articles/7624-mcdonald-s-finds-flexibility-with-digital-menu-boards
9Techemergence, Fast Food Robots, Kiosks, and AI Use Cases from 6 Restaurant Chain Giants: https://www.techemergence.com/fast-food-robots-kiosks-and-ai-use-cases/
10ZDnet, Jeremy Corbyn wants to tax robots and their greedy overlords: https://www.zdnet.com/article/jeremy-corbyn-wants-to-tax-robots-and-their-greedy-overlords/
11Business Insider, McDonald's shoots down fears it is planning to replace cashiers with kiosks: http://uk.businessinsider.com/what-self-serve-kiosks-at-mcdonalds-mean-for-cashiers-2017-6
16
Samsung: Automating The Home And Workplace With Artificial Intelligence
Samsung is the world's largest consumer electronics company by revenue, manufacturing and selling 500 million connected devices every year.
In 2018 it announced that within two years, artificial intelligence (AI) will be baked into every single one of them.1 From the latest phones to fridges, TVs and washing machines, Samsung is betting that the added convenience will drive consumer spending towards their products.
Samsung also make industrial technology, and this year the company unveiled Saram (Korean for “Person”) – an AI-equipped robot that initially will be used for heavy lifting, but could end up performing surgery.
What Problems Is Artificial Intelligence Helping To Solve?
In our homes, connected devices are increasingly a part of our lives – modern cars, homes, phones and appliances are packed with data capture and sharing capabilities. The applications are endless, from fitness trackers that collect data that we can analyse on our smartphones, to electricity meters that analyse our use of power and tell us where we could cut waste.
The problem is, it's still very early days for much of this technology, and there's a messy tangle of different standards and protocols vying to become our favourite. When we have to rely on a bunch of different apps and interfaces to get at this data, our ability to correlate it and make useful sense of it can be severely hampered by information overload.
At work, too, intelligent technology and robotics are becoming commonplace. In the past they were capable of carrying out routine and boring tasks if they were diligently programmed to do so. However, abilities to carry out tasks involving even a tiny degree of deviation from the norm were severely limited until the arrival of machine learning.
How Is Artificial Intelligence Used In Practice?
Currently, Samsung has an AI personal assistant, much like Apple's Siri and Amazon's Alexa, that goes by the name of Bixby.
With so many voice-enabled virtual assistants becoming available, the differences show where companies are being innovative and where they are just going with the flow. One of Bixby's distinguishing characteristics is that it is capable of integrating throughout the entirety of an application it interfaces with.2 With other voice systems, specific keywords are hard-linked to specific functions of the application, such as starting or stopping media playback on a music app.
It would be a good bet to say that this technology will end up underpinning Samsung's plans for all of its devices to be artificially intelligent by 2020.
What Samsung is trying to do is use AI to break down the final barriers in the way of our communication with machines. Previously, machines had to be laboriously programmed in technical programming languages. Now we are seeing a shift towards communicating in simple spoken human languages.
This revolution is taking place in industrial settings as well as homes. Samsung is already using an artificially intelligent robotic arm in manufacturing operations.
Its Saram robot AI was unveiled in 2018 and will initially be used in factories and manufacturing plants to carry out heavy lifting duties. It will be integrated into robotic arms of the type often used in high-tech manufacturing operations. However, the addition of AI to the mechanical arms will mean they are capable of working smarter, for example, spotting if parts are damaged or faulty before assembling them into a finished product.
In the future it could become part of a bipedal robot, which Samsung is reportedly also developing.3
Robotic factory workers could be used to carry out tasks that are dangerous or uncomfortable for humans to carry out for prolonged lengths of time.
Samsung has also invested in the Israeli robotics manufacturer Intuition Robotics.4 Intuition has developed the “social companion” robot ElliQ, a virtual assistant designed to help and engage the elderly in their homes. This shows that Samsung is committed to its plan of rolling out its AI and robotics into people's homes.
What Technology, Tools And Data Were Used?
Samsung's Bixby voice assistant uses natural language processing to interpret what it should do based on what a person says. Like other voice assistants, it uses neural nets and deep learning to become better at understanding how we talk and what we mean. This gets over the barriers to communication caused by humans with their inferior brains having to memorise a list of exact commands and phrases the computer will understand.
This year it also launched its SmartThings app, which is a platform for unifying all of its different intelligent devices. It hopes mainstream buyers will start installing smart refrigerators, kettles and coffee makers in their homes once the process of managing them is convenient enough.
Very little has been made public knowledge about the Saram robot system, apart from a spokesman stating that it uses AI technology. Speculation based on patent and trademark applications has suggested they are working towards a human-like, bipedal robot for industrial applications. The problems inherent in creating such a robot are quickly being solved as AI brings us closer to computer-controlled motors, which can work accurately enough to ambulate a humanoid body. The latest humanoid robots can keep their balance if struck by other moving objects without warning.5
What Were The Results?
Bixby has enabled Samsung to compete with Amazon, Apple and Microsoft by equipping its phones with a natural language virtual assistant. Users have a more frictionless interaction with their device and ultimately can get things done more quickly and easily.
It accounted for 6.2% of the market for AI-enabled voice assistant devices sold in May 2018 – behind market leaders Siri (Apple) at 45.6% and Google Assistant at 28.7%.6 It's worth noting that unlike Google Assistant and Siri, Bixby is only available on phones and isn't a stand-alone assistant device. But this is likely to change if Samsung's plan for all their devices to be equipped with AI by 2020 comes to fruition.
By developing the SmartThings platform, Samsung is in with a chance of becoming one of the first companies to successfully market the concept of the “connected home” to the non-techie public. With homes becoming increasingly computerised, whoever develops the operating system that becomes the standard for connecting everything together is likely to win big.
Key Challenges, Learning Points And Takeaways
Samsung has shown it is firmly committed to the idea that AI will dominate consumer electronics with its 2020 promise.
Samsung is also betting heavily that autonomous, mobile robots will have a huge impact in our homes and workplaces.
Services will let us use AI to tie together all of the data we gather through disparate “smart” devices and present it to us in a way we can use to take action.
Smart, connected homes are still the preserve of the tech-savvy slice of the population, but that's likely to change as more people b
ecome aware of the efficiencies on offer, and solutions are presented in a more consumer-friendly way.
Notes
1Financial Review, CES 2018: Samsung vows to add artificial intelligence to everything it does: https://www.afr.com/technology/ces-2018-samsung-vows-to-add-artificial-intelligence-to-everything-it-does-20180108-h0fdtd
2Forbes, How Is Samsung's Bixby Different From Other Voice First Systems?: https://www.forbes.com/sites/quora/2017/03/23/how-is- samsungs-bixby-different-from-other-voice-first-systems/#6ea3d30445f3
3Business Korea, Samsung Electronics to Make Artificial Intelligence Robot “Saram”: http://www.businesskorea.co.kr/news/articleView.html? idxno=20610
4Venturebeat, Intuition Robotics nabs Samsung as investor, launches U.S. beta trial of ElliQ companion robot: https://venturebeat.com/2018/ 01/09/intuition-robotics-nabs-samsung-as-investor-and-launches-beta- trial-of-elliq-social-companion-robot/
5The Week, Bipedal humanoid robot masters human balancing act: https://www.theweek.in/news/sci-tech/2018/10/03/Bipedal-humanoid-robot-masters-human-balancing-act.html
6Business Insider, Siri owns 46% of the mobile voice assistant market – one and half times Google Assistant's share of the market: http://uk. businessinsider.com/siri-google-assistant-voice-market-share-charts- 2018-6
17
Starbucks: Using Artificial Intelligence To Sell Millions Of Coffees Every Day
Starbucks has close to 30,0001 outlets around the world and serves up around 4 billion cups of coffee annually.