This blog post has been co-authored by Ken Toombs, Managing Partner, Infosys Consulting and Roberto Busin, Partner and Manufacturing Segment Head, Infosys Consulting.
In part one of this blog, we reviewed how artificial intelligence (AI) is shifting customer expectations through personalized experiences, especially in industries like retail, finance and travel, and how it is actively helping enterprises work with unstructured data. Part II of this blog post looks at the increasing adoption of AI and the need to stay competitive in a fierce business environment.
How will AI transform the way businesses are run?
This is a question that is on the minds of many decision-makers, and they can expect their operations to be impacted in three important ways due to AI, namely, in automation of tasks, creation of more efficient systems, and increase in proactive decision-making.
1. Automation of tasks
Much has been written about the adoption of robotics to automate repetitive tasks. We have, for example, observed the rise of tools such as optical character recognition (OCR), which successfully navigated the journey from cutting-edge technology to 21st century table stakes.
But as we move into an epoch where AI is embedded into technology infrastructure, robotics is able to access the core application data required to travel beyond low-level, repetitive tasks into the realm of high-value, sophisticated tasks.
This shift can be observed through the development of models. Historically, this has been the domain of highly paid quants. Each week, these math czars would sift through sets of data, identify disparate elements with the promise of correlation and build models to explain their relationships. These models would then be plugged into algorithms, which would power things like trading strategies for hedge funds.
But in a world where machine learning meets big data, the task of creating models is shifting to software that is guided by engineers to identify meaningful business correlations that unlock valuable competitive insights. A process that used to generate two models per quant per week now generates thousands of models each day.The benefits of these efforts are now appearing across several industries. As an example, Google has built an entire health care division within DeepMind, which is effectively bridging the gap between the two massive data sets of medical literature and an individual's human genome in a way no human could.
Taking it one step further, tech giants like Facebook are now building AI that builds AI. Essentially, because AI is hard and there is a dearth of people who do it well, programs are being built that can generate AI algorithms in alternate environments. This technology could enable companies that can't hire their own AI developers to generate native AI programs for their businesses.
2. Creation of more efficient systems
The creation of models doesn't just apply to customer-facing solutions. In July'16, Google announced that DeepMind researchers had developed energy efficiency strategies allowing them to reduce energy consumption by multiple percentage points. Given Google's massive consumption of energy to power data centers, this AI-generated insight has yielded massive annual savings.
3. Increase in proactive decision-making
Cybersecurity has rapidly shifted from an afterthought to a top CIO priority in a few short years. Much is made of this threat from external hackers, and for good reason, but according to Fortune magazine, "27% of electronic attacks on organizations - public and private - come from within."
The tricky part about internal breaches is that it's traditionally been hard for companies to proactively identify employees at the greatest risk of committing breaches.
A new software program, Scout, which was developed by cybersecurity firm Stroz Friedberg, is now leveraging AI to change that. By leveraging Scout's algorithms, companies can "autonomously detect risk indicators in employee communications, enabling companies to detect, assess, and respond to threats before they cause harm to the organization, its assets, or its people."
Scout also scans large bodies of text to identify unnaturally high concentrations of unconscious, negative sentiments - such as feelings of victimization - by specific employees. The program can then generate an actual list of the 10 employees at any given time who are most at risk of committing a breach.
This doesn't mean a company needs to mimic Minority Report and penalize workers for breaches they are likely to commit in the future. Instead, it enables firms to address disgruntled workers before negative feelings escalate into an actual breach, which is ultimately a much more efficient, inexpensive and humane outcome.
Dating firm eHarmony, which has spent an inordinate amount of time evaluating interpersonal compatibility, saw the similarity between choosing the right person in a professional setting as romantic and realized their AI algorithms could help.
As Dan Erickson of eHarmony recently told Christopher Steiner of Forbes, "We are not allowing the HR department or C-suite to just say what the culture is - we're using current employees to get a real reading."
To accomplish this, eHarmony gives personality tests to a wide group of a company's employees. They then aggregate the data into a true company culture, which is evaluated against potential applicants' readings. The end result is ever-improving employee fit, which leads to happier offices and more productive employees.
How should a company respond to AI, and what will the implications be if they fail?
The explosion of structured data from the Internet of Things and the cataloguing of web data is providing AI the fuel it needs to create ever more sophisticated algorithms. As the volume of this data continues to accelerate, so will the power and reach of AI.
When evaluating how a company should respond to the changes being ushered in by AI, start with your customers.
Ask yourself what the implicit assumptions are that drive engagement with your customers? How would that engagement change if your company knew more about them and were able to effectively predict what their needs were before they realized it themselves? These answers will illuminate where forces of AI will push your particular industry. From there, develop a strategy to evolve your offerings and accommodate customer expectations.
When considering AI's impact on customers, companies need to redirect their focus internally. Identify the knowledge tasks with rates of production limited by the uniqueness of their offering. These activities - which conventional wisdom traditionally argued were beyond the reach of outsourcing or automation - are precisely the things next generation AI is targeting.
Paradoxically, this elimination of human activities may be a good thing for those affected. According to a 2015 Gallup Workforce Survey, only 32% of the U.S. workforce is considered engaged. The majority (50.8%) of employees were "not engaged," while another 17.2% were "actively disengaged", which suggests that today the biggest waste in an enterprise relates to human potential - raising the question of using AI to redesign work for top enterprise roles to amplify engagement, performance and innovation (and automate more routine, task-oriented roles.)
After looking at their workforce, organizations need to examine the way they consume resources, placing particular focus on the largest areas of spend. Are these consumption cycles fully optimized? If these line items could be reduced by 5%, what would that do to the overall company performance?
If the competition gets there first, a company will be stuck with inferior margins while fighting to retain customers. But if AI can be leveraged to transform the way business operates, a company can generate an enduring cost advantage while concurrently increasing customer responsiveness.
While pondering on the way forward, it's instructive to remember Victor Hugo's prescient words, "You can resist an invading army; you cannot resist an idea whose time has come." The age of AI is here. It's up to companies to leverage it as a tool for a better world and superior company performance.
This blog post has been co-authored by Ken Toombs, Managing Partner, Infosys Consulting and Roberto Busin, Partner and Manufacturing Segment Head, Infosys Consulting.
Artificial intelligence (AI) is polarizing. Elon Musk has called it "our greatest existential threat," and tweeted that it is "potentially more devastating than nukes".
At the same time, renowned AI expert and Google/DeepMind Director of Engineering Ray Kurzweil has said, "...biological humans will not be outpaced by the AIs because they will enhance themselves with AI. It will not be us versus the machines ... but rather, we will enhance our own capacity by merging with our intelligent creations."
The debate is not about to be settled, but what we do know is that the age of AI is upon us. For evidence, we again turn to Elon Musk - the man intent on being AI's conscience - who has now embraced it through Tesla's commitment to self-driving cars.
Indeed, in this two part blog post, we see examples of AI making their way into industries as diverse as medical services, financial engineering and travel. The changes underway represent tectonic shifts that will play a key role in determining who wins and who loses in the coming decades and how AI impacts our everyday life.How will AI impact business?
While the implications of AI on business can be vast, it can be distilled into three key questions, which will be addressed in these blog posts:
How will AI shift the expectations of customers?
AI is fundamentally altering customer experiences in two important ways. One, it is personalizing customer experiences in ways that used to be reserved only for premium clients, and second, it is moving beyond the delivery of tools into the delivery of solutions that can work with unstructured data.
Let me explain. Historically, when a customer engaged with a business, the exchange would start with establishing a dialogue with the customer, learning about their needs, and then prescribing solutions. In a world with AI however, the conversation starts with businesses harnessing contextual information, comparing it with millions of historical patterns and determining what customers need, even if customers aren't able to articulate it themselves. This enables businesses to respond to customers in a highly contextual manner, providing a far superior experience. To demonstrate, let's look at a few real world examples.
Most customers find calling the customer service frustrating. The customer spends time educating the representative about what they need and then often receive solutions to the wrong problem.
Yseop Smart Machine observed this situation and developed an AI program that acts as a 'smart coach' for customer service teams by providing contextual knowledge based on both pattern recognition and CRM data. Instead of relying on training and experience to suggest solutions, the smart coach aggregates meta data about a client and guides the customer service representative through a dialogue that's being informed by all of the conversations that came before it. Even better, the system evolves as time passes, adjusting to changing customer preferences as markets change and more permutations are recorded.
Adding another layer of value, Mattersight offers a tool that uses AI to classify callers based on personality traits and match them with people who communicate in a complementary style. Building a comprehensive emotional profile of people based on their language structure and word choice, Mattersight determines which call center employee is best suited to have an immediate emotional connection with the caller.
Combining Yseop and Mattersight yields an experience where customers are connected with someone they innately relate to, who already know about them and can seamlessly leverage big data to suggest exactly what they need.
Personalization of service is also happening in industries that require highly sophisticated customer interactions, such as financial services. The outsized role of machines in equity markets has been well documented. According to Thompson Reuters, algorithms now account for 75% of all financial market volume, a number that continues to grow.
However, private individuals want a customized, human interaction that matches sophisticated analysis with their unique financial goals. Historically, this has been too expensive to scale and has been reserved for the ultra-wealthy, leaving the investing masses with a disparate set of generic, complex tools.
Startup WealthArc is capitalizing on this customer pain point by "leveraging data analytics and artificial intelligence support systems to empower wealth managers to transform the way they share relevant and understandable information with clients."
Essentially, they're using AI to take a service - in this case heavily customized financial recommendations - that has historically been reserved for 1 percent of investors and extend it to all investors cost-effectively, thus creating a game-changing offering that can be matched only by the power of algorithms.
The CEO of WealthArc sums up this transformation nicely, "I believe that the future of private wealth management will be like Mr. Spock from Star Trek - thinking like a machine but with a human mother."
Moving beyond tools to enhance customer experience
In addition to customer intimacy, the nature of the problems being solved by companies is also changing. Historically, web applications provided users tools to solve problems: banks gave apps to deposit money, retailers offer sites with clothes to search through and buy, and so on.
But with the advent of AI, companies are taking unstructured customer problems - "I want a fun dress for my company holiday party" - and transforming them into complete solutions. A case in point is travel services company, WayBlazer. Speaking with the Harvard Business Review, Terry Jones, founder had this to say about the company:
"I started as a travel agent, and people would come in, and I'd send them a letter in a couple of weeks with a plan for their trip. The Sabre reservation system made the process better by automating the channel between travel agents and travel providers. Then with Travelocity we connected travelers directly with travel providers through the Internet. Then with Kayak we moved up the chain again, providing offers across travel systems. Now with WayBlazer we have a system that deals with words. Nobody has helped people with a tool for dreaming and planning their travel. Our mission is to make it easy and give people several personalized answers to a complicated trip, rather than the millions of clues that search provides today. This new technology can take data out of all the silos and dark wells that companies don't even know they have and use it to provide personalized service."
WayBlazer has fundamentally shifted where the thinking takes place to plan a trip. Instead of a customer saying to themselves, "I want to plan a romantic weekend with my partner to a quiet island" and then using a set of tools to run an exhaustive search for hotels and flights, they simply articulate their problem statement and the program solves the problem for them in a way that is informed by all of the customer experiences that have come before the program.
Stay tuned for the second part of the blog coming up next week.
In the late 1950s and early 1960s, at the dawn of the jet age, nothing was more glamorous than traveling by air. Frank Sinatra even had a hit song - "Come Fly with Me" - about those who were then known as the 'jet set' crowd.
How times have changed. Today flying, especially for the infrequent flier - not the business ones- can be rather unsettling. Security lines at airports are so long that travelers often need to get there a few hours before their flight is scheduled to take off. I say 'scheduled' because how many times has your plane left on time? You would think that with the array of digital tools, big data, and predictive analytics, flying would be as easy as shopping online. But this is not the case - weather, turnaround times, readiness of ground staff, and inter connected flights adds a whole set of hidden complexities to airport management. Unfortunately, airports and several players in the ecosystem have not adequately leveraged digital tools, and so consumers are yet to have the seamless experience that has become an integral part of other industries.
Is it time then for the airline industry to usher in change? I should think so. Growth in revenue passenger miles (RPM), a metric used by the transportation industry to determine the number of miles traveled by paying passengers, are slated to slow down over the next decade in the United States. RPM is calculated by multiplying the number of paying passengers by the distance traveled by them. The Federal Aviation Administration's forecast for the next decade, projects RPM to increase by a paltry 2.6 percent a year. Domestic RPMs in America are forecast to grow at an even more underwhelming 2.1 percent, and international RPMs are forecast to grow at 3.5 percent a year. Technology could play an important role in making airports digital, simplifying baggage management, increasing service offerings and addressing security concerns. This would improve the flying experience and bringing the passengers back.Digital airports
Majority of the projected growth in RPM is expected to be in large and medium airports, which are already crowded. To ease the pressure on passengers, airports of the future can simplify a number of processes. The basic function of any airport and airline relates to supporting passengers in activities such as check-in, check-in of bags, customs check, and boarding the airplane. Technology through automation can make it easier for travelers at each of these touch points. Technology can also play a role in several other complex functions that an airport has to manage such as security, manpower management, ground transportation, air traffic control, and transportation security administration, besides other things.
When airport authorities are looking to introduce technology they could, for example, use digital solutions to dig deeper into passenger demographics, preferences, and details that could help then better understand and plan service requirements, such as to decide the number of agents with special skills that would be required for a particular flight. Airport authorities can also provide passengers a virtual reality view of the airport and its facilities much in advance, so as to familiarize them with the airport rather than using outdated airport maps on paper.Re-thinking Baggage
Plane cabins are getting cramped and what aggravates the situation is large bags that are stuffed into overhead compartments, leaving no room for regular-sized bags, briefcases, and backpacks. Passengers who board later -those that are not frequent fliers or do not have a high priced ticket- often find that they do not have enough space to store their bags and have to rely on the flight attendants' creativity to find a solution, or check-in their baggage instead. All the while, the rest of the passengers are left watching and wondering if this could have been handled better. The key reason for this cramped cabin is that baggage fees have steadily increased, and travelers are allowed to haul large bags with them. If you are wondering what could help resolve this issue, well it is a smart baggage arrangement.
Imagine a baggage concierge that can be accessed from a residence or hotel. Not only will this make the check-in process faster but easier as well. If passengers do not provide information on the size and weight of the baggage they are carrying in advance, the airlines could decide where the bags are stored e.g. in overhead bins or in the cargo hold, so as to better plan weight and balance of aircraft, and thus hasten the loading and turnaround process of the aircraft. Sensors on the bags could allow passengers to track the movement of their bags, too. In case of connecting flights, passengers can be assured that their bags have made it (or not).
With beacons, sensors, and tracking technology, airlines could offer a 'bag drop' service to the home or hotel so that passengers need not wait at the terminal to claim their luggage. Sensor readers at airport baggage sorting facilities can vastly enhance and expedite their work and reduce manual intervention. In fact, the Industrial Internet Consortium recently gave its approval to the Smart Airline Baggage Management test bed co-developed by Infosys to do just this.
The right technology can not only offer an intelligent baggage solution for passengers and airport authorities, but reduce cost of operations for the airline, and free up airport staff to spend more time improving the flying experience of their passengers. Airlines, airports, and technology companies have an opportunity to work together to make this goal mainstream.Simplifying access to airport services
Often passengers wish to access certain merchandise or services from airport shops, but run out of time while searching for the sales counter. As a solution, wouldn't it be good if passengers could simply order from anywhere, anytime through an airport app or website, and have the product delivered right at the gate before they board their flight or at the airport of arrival? This would offer airports an opportunity to build a positive relationship with their customers. The app could be used to order food, book a car, order that fast track/expedited security lane pass, a bouquet of flowers or a gift.Overcoming the security hurdle
The security check at the airport is a hurdle for most passengers, especially in America. For those with a Global Entry/ TSA (Transportation Security Administration) PreCheck, there is some ease of not having to stand in serpentine queues, taking out laptops, and removing shoes and jackets. To simplify security checks, airports can create a dedicated fast track security line for those who wish to clear the security process in a shorter duration. Of course, passengers would still have to go through the same process but the queue would be much shorter. Airlines could do more to work with airports and provide this perk to their passengers, especially when loyalty points are being utilized or as promotional offers to elite members.
Digital airports can simplify and improve navigation as well, such as by enabling cars to book a parking spot at the right terminal based on the flight details streamed to the car's GPS system. This means an autonomous car can take a passenger from home to the designated parking spot without spending precious time having to scout for a location to park.Seizing the technology opportunities
Technology has tremendous potential to improve passenger service, and airports can leverage highly scalable and intelligent platforms that seamlessly interact with heterogeneous systems to facilitate communication between service providers and passengers, thus making travel pleasant. Because soon it may not just be places of business or leisure that we are flying to, destinations outside our planet itself!