Artificial Intelligence in Practice
Page 9
It uses this to inform new product decisions – for example, the decision to launch Cherry Sprite as a bottled product in the United States was taken because the data showed that this was likely to be a winning initiative.7
Computer vision analysis and natural language processing of social media posts, as well as deep learning-driven analysis of social engagement metrics, allows Coca-Cola to produce social advertising that is more likely to resonate with customers and drive sales of its products.
Applying TensorFlow to create convolutional neural networks enabled scanners to recognize product codes from a simple photograph, increasing customer engagement with Coca-Cola's different loyalty programs around the world.
Key Challenges, Learning Points And Takeaways
If you sell hundreds of different products across multiple countries, perceptions and customer behavior can vary greatly from market to market. Understanding these differences helps tailor specific messages for different markets, rather than relying on a one-size-fits-all approach.
When you're dealing with global brands, user data from social media or generated through your own systems (such as vending machines) is vast and messy. AI provides a viable method of structuring this data and drawing out insights.
Computer vision technology such as image recognition tools can analyze millions of social media images to help a brand understand when, how and by whom its products are enjoyed.
As well as making marketing decisions, brands that are fully invested in AI are beginning to use it for designing new products and services.
Notes
1Venturebeat, Coca-Cola reveals AI-powered vending machine app: https://venturebeat.com/2017/07/11/coca-cola-reveals-ai-powered-vending-machine-app/
2Digital Food and Beverage, Coca-Cola is Using AI to Put Some Fizz in Its Vending Machines: https://foodandbeverage.wbresearch.com/coca-cola-artificial-intelligence-ai-omnichannel-strategy-ty-u
3Nastel, Social Media Analytics At Coca-Cola: Learning From The Best: https://www.nastel.com/blog/social-media-analytics-coca-cola-learning-best/
4Adweek, Coca-Cola Wants to Use AI Bots to Create Its Ads: https://www.adweek.com/digital/coca-cola-wants-to-use-ai-bots-to- create-its-ads/
5Digiday, How Coca-Cola targeted ads based on people's Facebook, Instagram photos: https://digiday.com/marketing/coca-cola-targeted-ads-based-facebook-instagram-photos/
6Google Developers Blog, How Machine Learning with TensorFlow Enabled Mobile Proof-Of-Purchase at Coca-Cola: https://developers. googleblog.com/2017/09/how-machine-learning-with-tensorflow.html
7Coca-Cola, Fountain Favorite: Sprite Cherry is First National Brand Inspired by Coca-Cola Freestyle: https://www.coca-colacompany.com/ stories/fountain-favorites-sprite-cherry-and-sprite-cherry-zero-become -first-national-brands-inspired-by-coca-cola-freestyle
13
Domino's: Using Artificial Intelligence To Serve Up Hundreds Of Thousands Of Pizzas Every Day
Domino's Pizza is the largest pizza company in the world – it sold over 300,000 pizzas every day in 2017,1 from 48,000 stores in 85 markets.
While cooking and delivering pizzas may not immediately seem like the most tech-driven business, Domino's has consistently ensured it is harnessing new technologies as they become available. Most noticeable until now has been its drive to allow customers to order pizzas from any platform – over 60% of its sales now come through digital channels2 and you can order pizzas through smart TVs, Facebook, Twitter, Amazon Echo, smart watches and numerous other methods – including by simply sending a pizza emoji via SMS.
Data and analytics have long played a key part in Domino's marketing strategy, and it collects vast amounts of data to understand who is ordering its pizzas and how it can improve its service. Now it is embracing artificial intelligence (AI) to ensure a more consistent quality and build a speedier, more environmentally friendly delivery infrastructure.
What Problems Is Artificial Intelligence Helping To Solve?
In the fast food business, customers are fickle creatures. New options are regularly becoming available as habits and food fashions change, and if pizzas are cooked or delivered that do not meet their expectations in terms of consistency and quality, customers will become dissatisfied. This means they are likely to look to new alternatives and rival businesses for their fast food fix.
And while pizza delivery may be a very convenient way of getting fed from a customer's point of view, it is an expensive exercise – in terms of both fuel and wages, which the company has to cover, and the environmental cost of making an individual journey to deliver each pizza.
How Is Artificial Intelligence Used In Practice?
Domino's has started using a system called Pizza Checker that photographs every pizza when it leaves the oven, and then uses machine learning algorithms to inspect it for quality before it reaches the hungry customer.3
The camera system checks the type of pizza against the customer's order to make sure they are getting what they paid for. It also verifies that toppings are distributed evenly and that the crust has been properly baked at the correct temperature.
The system – installed in 2,000 Domino's kitchens in seven countries in 2017 – sends users a picture of their pizza before it is delivered, and also notifies them if a quality failure meant that the pizza had to be remade. The hope here is that this will make them more understanding when there are occasionally inevitable delays!
AI is also used when processing customer orders received over the telephone. It has developed its own “virtual assistant” technology in the style of Apple's Siri, which can communicate with customers by voice when they call to place an order. The first iteration of this technology was launched in 2014 and is known as Dom. Last year, a new version – called DRU (Domino's Robotic Unit)4 – was rolled out, featuring more sophisticated natural language processing technology.
This means it can respond to more complex queries and understand the sometimes very different speech patterns and mannerisms used by different callers.
DRU is actually the name used for the company's autonomous technology across all of its units, including the pizza checkers, virtual ordering assistants and its autonomous delivery vehicles.
Yes, Domino's is also actively working to make self-driving pizza delivery a reality. Thanks to a partnership with Ford, pizza lovers in Ann Arbor5 and Miami,6 USA, have been able to take part in an experiment where their pizzas are delivered by a fleet of autonomous vehicles equipped with ovens to keep the pizzas warm during their journey. In their initial trial, the vehicles had to be accompanied by humans.
More recently, the company has been launching completely autonomous delivery pilots in Germany and the Netherlands, where it is partnering with Starship Technologies.7 Don Meij, Domino's Group CEO and Managing Director, says: “We are a global company and we are eager to progress innovative technology in all of the countries in which we operate – we are very excited to be partnering with Starship as it brings regular deliveries by robot one step closer to commercial operations.”
What Technology, Tools And Data Were Used?
Domino's Pizza Checker has been developed by Dragontail Systems and uses Google image recognition algorithms to identify the type and placement of toppings on a pizza, as well as the temperature that the pizza was cooked at. It uses data from the customers’ orders to ensure the right pizza is being prepared. This is done using computer vision technology, which allows AI algorithms to be trained to “see” – and recognise objects in the same way humans do.
As for autonomous delivery – Domino's has partnered with Starship Technologies to use delivery robots that are capable of guiding themselves to customers’ homes. The six-wheeled Starship bots use GPS, radar, ultrasonic sensors, as well as cameras to navigate autonomously. The customers will then be able to access the compartments where their food is kept hot, and the drinks are kept cold, by using a smartphone app.
What Were The Results?
Domino's hopes that the computer v
ision system will lead to fewer pizzas being rejected because they don't meet customers’ quality expectations. In worst case scenarios, this can lead to customers who otherwise have high lifetime value expectations to the chain taking their custom elsewhere.
This year the business announced that it will invest a further $1 million in fitting its kitchens out with the automated camera systems provided by Dragontail.8
And although Domino's has said that the DRU delivery vehicle won't be appearing at customers’ doorsteps “tomorrow”, it fully anticipates that autonomous delivery will become a reality in the near future.
Key Challenges, Learning Points And Takeaways
When you have huge numbers of outlets serving millions of customers, those customers will come to expect a level of consistency of the quality of products and service. AI can help maintain that level.
Companies like Domino's believe that natural language technology is at a stage where it can provide the same level of customer service as a human telephone operator.
As well as moving people around, autonomous vehicles have applications for moving goods from place to place, and even delivering them straight to customers’ doors. As well as saving the operators’ money (which can be passed onto the customers with lower prices), this could have positive environmental benefits as robots will be able to route themselves more efficiently than humans.
Notes
1The Times, Pizza guzzlers give Domino's a slice of success: https://www. thetimes.co.uk/article/pizza-guzzlers-give-dominos-a-slice-of-success- dzftlldtn
2PR Newswire, Domino's on Quest for Digital Dominance Using Artificial Intelligence: https://www.prnewswire.com/news-releases/dominos-on- quest-for-digital-dominance-using-artificial-intelligence-300633827 .html
3Interesting Engineering, Domino's Will Use AI to Make Sure Every Pizza They Serve Is Perfect: https://interestingengineering.com/dominos-will-use-ai-to-make-sure-every-pizza-they-serve-is-perfect
4ZDNet, Domino's partners with Nuance for DRU artificial intelligence: https://www.zdnet.com/article/dominos-partners-with-nuance-for-dru-artificial-intelligence/
5Tech Radar, Ford and Domino's are filling self-driving cars with pizza to see how we feel about it: https://www.techradar.com/news/ford-and-dominos-are-filling-self-driving-cars-with-pizza-to-see-how-we-feel-about-it
6Tech Radar, Ford and Domino's demonstrate self-driving deliveries with – what else – pizza: https://www.techradar.com/news/ford-and-dominos-demonstrate-self-driving-deliveries-with-what-else-pizza
7Starship, https://www.starship.xyz/press_releases/starship-technologies-launches-pilot-program-with-dominos-pizza-enterprises/
8Domino's Pizza moves forward with Dragontail Systems AI: https://www. finnewsnetwork.com.au/archives/finance_news_network190563.html
14
Kimberly-Clark: Using AI To Make Sense Of Customer Data
Kimberly-Clark produces some of the world's best-known personal care brands, including Huggies, Kleenex and Scott, which are sold in 175 countries. In fact, one in four of the world's population uses its products on a regular basis.1
While its products may not be the most glamorous, they are an essential part of day-to-day life, meaning huge amounts of resources are spent to ensure they are effectively produced, sold and distributed.
This means, like many other large businesses, Kimberly-Clark has found that the most productive way forward has been to essentially become a technology company. As well as producing the everyday goods it has built its name on, it licenses over 150 of its self-built technologies to other businesses, ranging from start-ups to Fortune 100 companies.2
What Problems Is Artificial Intelligence Helping To Solve?
In Kimberly-Clark's market, businesses need to build intimate understandings of their customers’ lives to understand how they interact with their products.
This means understanding how life-changing events like pregnancy and weddings will affect their buying habits and patterns.
To do this, they have to make as much sense as they can out of the explosion of data customers leave behind as they browse their product websites and social channels or make purchases in stores. Even with computers this was a monumental challenge, before artificially intelligent software solutions started to become available in recent years.
As well as this, Kimberly-Clark faces a challenge common to all businesses looking to undergo digital transformation. The one ingredient that is still essential to staying ahead of the tech curve is people. Machine learning is great, but it hasn't quite reached the stage where it's capable of the blue sky/out-of-the-box (insert your own cliché!) thinking necessary for coming up with brilliant new ideas.
The people who are good at it are in very high demand – global demand for trained data scientists was said to outstrip supply by 50% in 2018,3 and this is predicted to grow. Considering their options, working for a company primarily known for producing toilet paper and diapers might not be the most obvious career path for these talented people.
How Is Artificial Intelligence Used In Practice?
Kimberly-Clark uses artificial intelligence (AI) to make sense of all of the data it gathers from customers and its business operations.
This data is used to build detailed models of who its customers are. Actual customers can then be “segmented” according to which model they fit best, to give the business clues about what they might want to buy.
One key success has been the increase in accuracy with predicting when customers would fall pregnant. Research showed that introducing them to the Huggies brand as early as possible to their finding out they were pregnant was key to converting them into a customer for the duration of their pregnancy. Once identified as likely to be shopping for pregnancy-related items, customers could be encouraged to sign up for loyalty schemes, and could be offered discount coupons, as well as useful content such as baby care advice.4
What Technology, Tools And Data Were Used?
Worldwide over 15,000 Kimberly-Clark products are sold every second of every day. Each one of those sales generates data points, ranging from time and place of sale to customer information from loyalty schemes or online shopping profiles, as well as external data from sources such as Nielsen ratings.
With so much data, traditional business intelligence and customer relations technology can't operate quickly enough to produce insights before the data becomes obsolete.
AI and the technology, which has built up to support it – big data platforms like Hadoop and Internet of Things frameworks – make it possible to extract meaning from the madness.
Kimberly-Clarke has partnered with Nielsen to use its Marketing Cloud platform and RevTrax software, as well as solutions from Webtrends, which use machine learning to offer promotions and provide tailored customer experiences.5 It involves using predictive analytics to understand what segment customers fit into, and offer them relevant products (such as diapers in the Huggies example).
Other partnerships it sees working are Tableau, Amazon and Panopoly to store and sort through the mountains of data it collects.6
In this example of targeted marketing, data would be collected by showing different social promotions and content marketing material to differently segmented audience samples. This is what marketers have always traditionally done with focus groups. But a machine learning solution working across social media can test against different target audiences far more quickly than a human-organized focus group ever could. This means Kimberly-Clark can break customers down into precisely defined groups, and build more accurate models of what potential customers look like, for particular products, at any given time.
Kimberly-Clark also hosts the K-Challenge to persuade rising stars in the world of data science and AI tech to consider building their career in the personal hygiene products arena.
The competitive event encourages innovators to submit their ideas for consumer goods technology. Kimberly-Clarke offers support for the chosen winners with research and design, mark
eting and making the idea into a reality.7
What Were The Results?
Kimberly-Clark's move towards advanced analytics with Webtrends resulted in increased sign-up rates of 17%.
Another campaign, to optimize targeting of customers for their Depend brand, saw a 24% increase in conversions.8
This was done by producing content that more closely aligned with the customer profiles that the analytics predicted would be responsive. These customers are also more likely to go on to be long-term repeat buyers, as well as to make positive recommendations to friends and family.
Key Challenges, Learning Points And Takeaways
Today, market-leading companies in every industry are transitioning into tech companies – it's essential if they want to stay ahead of the pack.
AI-driven analytics is far more powerful than traditional business intelligence solutions for customer segmentation and targeting when dealing with truly big data.
Businesses must earn their reputations as tech champions and pioneers to attract the necessary human talent. Until computers are clever enough to start running corporations on their own, of course.
Notes
1Kimberly-Clark, https://www.kimberly-clark.com/en-us/brands/our- brands
2Kimberly-Clark, https://www.kimberly-clark.com/en-us/company/ technology-licensing
3Inside Big Data, https://insidebigdata.com/2018/08/19/infographic-data-scientist-shortage/