Best uses of AI and machine learning in business

Sky

Sky

Sky has been building machine learning models that can make personalised recommendations based on a viewer's mood at the time by matching their state of mind to genres and content tags.

The algorithm analyses keywords to create a weighted set of semantic representations of moods for an individual programme or film. Spy comedy Johnny English, for example, might be assigned a score of 5.2 for "funny", 3.8 for "adventurous" and 1.4 for "exciting". Users then select keywords to receive a list of movies related to their moods.

Sky also used machine learning technology from Amazon Web Services (AWS) to automatically label famous guests at the Royal Wedding between Prince Harry and Megan Markle. The names were tagged when they arrived at George's Chapel in Windsor Castle, and added to a list with a biography and information on their connection to the couple. Viewers could then search the footage on demand for their favourite guests.

"We think it's a really interesting method of storytelling," Hugh Westbrook, senior product owner at Sky, told Computerworld UK.

"We certainly want to explore what machine learning and automatic recognition of objects within video can do for us, because that's a very interesting opportunity for us to analyse live data and tell people different things. I think we've definitely opened a theme where we can go with this."

The Wall Street Journal

The Wall Street Journal

The Wall Street Journal has been using machine learning to adjust the price of its paywall for different readers based on the probability of them subscribing.

The newspaper scores each reader across more than 60 dimensions of behaviour  including their location, reading habits and method of accessing the site. These are then combined into a total score from 0 to 100 that reflects the likelihood of them paying. The system can then predict the optimal moment to surface the paywall, quickly shutting out readers with high scores while giving those with low scores a larger number of articles before the paywall is enforced.

"We want to get to the point, through these exercises, that we're actually segmenting our audience into these tiny micro segments," WSJ data scientist John Wiley explained at the 2018 O'Reilly Artificial Intelligence conference.

"If you were to partition off different internet traffic, you'd actually be able to say that people with a score of a hundred were generally converting at about three times the rate of someone with a score of one."

Tinder
© Tinder

Tinder

Tinder is using the deep learning-powered Amazon Web Services (AWS) Rekognition service to find the best matches for premium users of the dating app.

The system automatically tags the 10 billion photos that users upload daily with personality markers based on what's in the image, such as labelling them as "creative" if they're playing music or "adventurous" if they're rock climbing. The tags are combined with Tinder's other data to determine the most significant aspects of a profile, whether the user has supplied a text profile or not.

"It provides not only cloud scalability that can handle the billions of images we have but also powerful features that our experts and data scientists can leverage to create sophisticated models to help solve Tinder's complex problems at scale,"  Tinder VP of engineering Tom Jacques explained at AWS re:Invent 2018.

"Privacy is also important to us and Rekognition gives us separate APIs to provide control and allow us to access only the features we want. By building on top of Rekognition we are able to more than double the tag coverage."

Ocado
© Ocado

Ocado

Ocado has announced what it says is the world's first AI-based fraud detection system for online grocery purchases.

The system identifies orders that are delivered but not paid for and determines whether it is the result of malicious intent by analysing data from past orders.

Ocado engineers implemented a deep neural network using Google's open sourced TensorFlow software library and uploaded the fraud detection system to a data lake in Google Cloud. The online supermarket claims that the system has improved Ocado's precision of detecting fraud by a factor of 15.

The company is using a combination of TensorFlow machine learning tools and cloud APIs to support a number of internal AI projects. One such initiative focuses on automating management of the deluge of customer service-related emails the supermarket receives.

Ocado is also building a computer vision system in an effort to replace barcode scanning in its warehouses, which it hopes will help both within its warehouse and delivery processes.

Read next: Ocado to replace barcode scanning with AI 'vision' to speed packing processes

Ocado is using TensorFlow for everything from routing algorithms for its robots to move around warehouses to improving existing features such as demand forecasting and suggesting items to add to shopping basket based on past habits.

Virgin Holidays

Virgin Holidays

Tour operator Virgin Holidays is using AI-powered software from UK startup Phrasee to automate and optimise the writing of its marketing email subject lines, with the machines outperforming humans by up to 10 percent when it comes to all-important open rates.

Read next: Virgin Holidays turns to AI for email marketing subject lines

Virgin fed Phrasee its brand guidelines and three years of email subject lines so that it could pick up the brand's tone of voice. Phrasee now feeds back optimised subject lines for the marketing team to use.

Once Virgin turned over the keys to Phrasee open rates jumped by two percent – accounting for several million pounds in revenue.

The Zoological Society of London
iStock

The Zoological Society of London

The Zoological Society of London (ZSL) is using Google's new Cloud AutoML platform to track wildlife by automatically analysing images captured by cameras in the wild.

ZSL uses these cameras to capture the motion of animals and humans and identify poaching threats. The images need to be tagged in order to analyse the data, which ZSL conservationists previously had to do manually.

ZSL provided Google with around 1.5 million tagged images to help advance the Cloud AutoML platform, which was announced in January. The charity is now training a custom model by feeding it data on conservation details such as region, environment and species that will create a complete auto-tagging tool.

Tesco

Tesco

Tesco is using machine learning algorithms across its business, from internal applications such as driver routing, to customer facing apps like integration with Google's home assistant device.

Read next: How Tesco is using AI to stock shelves, route drivers and order groceries with Google Assistant

For customers, Tesco's Labs division has worked with IFTTT to open up its APIs and create 'recipes' so that online shoppers can start to personalise their shopping, get automatic price drop alerts for certain items and order groceries though AI-powered home assistants like Google Assistant.

EDF Energy

EDF Energy

EDF Energy is looking to use AI to help make its nuclear power station's more efficient and to reduce customer's home energy consumption.

Read next: How EDF wants AI to optimise its nuclear power stations and the smart home

On the customer-facing side, EDF has been testing AI to perform character recognition to pick out and process the figures on meter readings sent in by energy customers. EDF can then use machine learning to do pattern recognition to spot trends in the usage data being collected.

EDF is also looking into using AI for real time condition monitoring to provide advice to power station operators. So the AI would monitor the various systems within the power station and provide either real time advice, prescribing what they should do, or, in the future, automating the power station.

Bloomberg
© Bloomberg

Bloomberg

Financial data specialist Bloomberg employs hundreds of data scientists to keep users hooked on its ubiquitous Terminals - the keyboards and monitors that give financial staff access to reams of market information.

Read next: How Bloomberg is using machine learning and data science to keep users hooked to its terminals

The overall aim is to use techniques like computer vision and natural language processing to improve the breadth of financial information available through the terminal. This should allow users to increasingly make queries on the terminal using natural language instead of specialised commands.

Fukoku Mutual Life Insurance

Fukoku Mutual Life Insurance

Japanese insurance firm Fukoku Mutual Life Insurance is replacing its 34-strong workforce with IBM’s Watson Explorer AI.

The replacement artificial intelligence system will calculate insurance policy payouts, which the firm believes will increase productivity by 30 percent and save around 140 million yen (£977,000) a year in salaries.

While the savings are significant, the initial cost of setting up IBM Watson is estimated to be 200 million yen (£1.4 million), and annual maintenance is expected to be about 15 million yen (£105 million).

The Watson-based system will be able to analyse medical certificates, surgery and procedure data and hospital stays and calculate the relevant payout.

NHS helpline

NHS helpline

The NHS is to trial an AI-powered chatbot on the 111 non-emergency helpline.

Being trialled in North London, its 1.2 million residents can opt for a chatbot rather than talking to a person on the 111 helpline.

The chatbot, created by Babylon Health, encourages people to enter their symptoms into the app, it will then consult a large medical database and users will receive tailored responses based on the information they've entered.

According to the Telegraph, this six-month trial aims to "reduce pressure on the NHS during the winter and beyond."

Uber

Uber

The car hailing service Uber has a core team providing pre-packaged machine learning algorithms 'as-a-service' to its team of mobile app developers, map experts and autonomous driving teams.

Read next: How Uber pre-packages machine learning algorithms for the whole organisation to use 'as a service'

Head of machine learning at Uber, Danny Lange told Computerworld UK: "We have really had machine learning for a while but it is something that can be really hard for software engineers to get. So we have created machine learning-as-a-service inside the company as a cloud service.”

Uber uses these capabilities to better predict your travelling habits within its core mobile application, improves its maps using computer vision and create algorithms for its autonomous vehicles.

Auto Trader

Auto Trader

The online car marketplace Auto Trader has a data science and insights team of thirty people, and is currently building a cluster with open-source database specialists MongoDB to store derivate data - for example, the year a specific model of car was manufactured. Having a database capable of recognising this ensures customers are being delivered accurate valuations.

Read next: Auto Trader uses MongoDB and machine learning to improve car valuation accuracy

Mohsin Patel, principal database administrator at Auto Trader told Computerworld UK: "We have to teach our systems depending on what features a car might have, or what derivate of a car the customer is viewing as the price could differ, and that's where the [machine] learning comes in."

Expedia

Expedia

Flight and hotel search and booking site Expedia told Computerworld UK that its core services were “built on machine learning”.

Read next: How Expedia.com was built on machine learning

David Fleischman, VP of global product said delivering quality flight search results is an "unbounded computer science problem" because flight itineraries and schedules are constantly changing. This is why Expedia's proprietary 'best fare search' (BFS) has to 'learn' and adapt all the time.

Expedia also uses its significant in-house machine learning resource – 700 data scientists and counting – to create algorithms for detecting fraud. Next it wants to allow customers to make travel queries using natural language. "The main goal is to answer a traveller's question and we use machine learning to solve that discovery problem," Flesichman said.

Royal Bank of Scotland
iStock

Royal Bank of Scotland

In 2016, Royal Bank of Scotland announced the launch of Luvo - a natural language processing AI bot which will answer RBS, Natwest and Ulster bank customer questions and perform simple banking tasks like money transfers. If Luvo is unable to find the answer it will pass a customer over to a member of staff.

While RBS is the first retail bank in the UK to launch such a service, others such as Sweden's SwedBank and Spain's BBVA have created similar virtual assistants.

RBS’ head of digital, Chris Popple says that the aim is to “make digital customer support as powerful as face-to-face”.

Read next: RBS says it's 'twelve months ahead' of other UK banks with Luvo AI customer assistant

Royal Free Hospital
iStock

Royal Free Hospital

The Royal Free Hospital in North London announced a partnership with Google’s DeepMind in February to build a mobile app that helps hospital staff monitor patients with acute kidney injuries.

According to Thomas Davies, Director of Northern, Eastern & Central Europe, Google for Work: “DeepMind is a general tool, we have just decided to go deeper into this one particular area. We hope to find some significant frontline benefits to the doctors and nurses.”

“It is very much around putting data content applications into the hands of frontline staff to do their job better.”

Engie
iStock

Engie

Marc Florette, chief digital officer at French multinational energy firm, Engie, says that the company is using a combination of drone and AI image processing technology to inspect its infrastructure.

"We have two or three main areas we can use AI and will focus on assets,” he says. “We have industry assets of high value like gas turbines or wind turbines and we have predictive maintenance, so you can increase the efficiency and profitability of the asset."

“For security we are using drones and image processing to inspect the high pressure network to avoid damages that can be very harmful.”

Moo.com
© Moo.com

Moo.com

Like RBS, business card website Moo.com uses artificial intelligence software to improve customer interactions.

Director of customer services, Dan Moross, says: "We have adopted a fairly basic AI and ML tool called AnswerDash.

"This is a contextual self-service support tool which uses our data to give contextual support in that online journey. This gave us not just savings with the call centre but additional revenue with customer conversion."

Government of Catalonia
iStock

Government of Catalonia

Jordi Escale, CIO of the Government of Catalonia, says that while AI plans are at currently in their early stages - and is consulting with IBM on potential projects - there are a number of ways the technology can be used for public sector firms.

"We start thinking about the impact in the services in the government like traffic control and autonomous vehicles,” he says.

This includes real-time facial image recognition and number plate recognition for the police. "Then there is health, supporting the doctor for knowledge by tapping all of this unstructured data,” he adds.

Las Vegas Sands

Las Vegas Sands

Jonathan Catling, director global data architecture at Las Vegas Sands Corporation, says the business - which owns hotels such as the Venetian and Palazzo - has designed a virtual assistant chatbot for guests.

“We came up with AI concierge," he says. "We started with a chatbot as a means of communication that they feel comfortable with.

"We don’t want to send an SMS or an email but 'chat' [with our guests]. So any time, anywhere you can take one step into the customer's comfort zone."

He adds that guests can engage with the chatbot through their Facebook account whenever they want. "We want you to get to the point where you can order room service faster. Not ringing down but using a tablet in the room, using your smartphone, anywhere.

"The conversation is the key to the customer, they’re not interested in the technology but the chat.”

Unibet
iStock

Unibet

Unibet parent company Kindred Group is turning to machine learning to recognise patterns of problem gambling across its suite of betting and casino websites.

Read next: Kindred Group reviews machine learning to address problem gambling

Copyright © 2019 IDG Communications, Inc.