Artificial Intelligence (AI) is predicted to be one of the next big digital disruptions. According to a PWC report, the big question is how to secure the talent, technology, and access to data to make the most of this disruption.
Let’s look at the growth of AI using some facts. In a discussion paper on AI published by McKinsey Global Institute, it has been highlighted that in 2016, companies invested $26 billion to $39 billion in artificial intelligence with tech giants and startups leading the pack. While 20% of the AI firms say they’re adopters, 41% of the firms are not very sure about the benefits of AI. The sectors that are strong adopters are high tech, automotive, and financial services sectors. These are followed by retail, media, and CPG, which are medium adopters. Finally, education, healthcare and tourism are sectors in which the adoption is pretty low.
AI implementation can directly impact the value chain of a company by impacting various functions such as R&D, production, marketing, and user experience. AI promises benefits but also poses urgent challenges that cut across firms, developers, governments, and workers. Let’s analyse the scenario of the practical implementation of artificial intelligence across industries and how companies can strike a fair balance between technology and human interaction.
Are businesses ready for AI?
The history of AI proliferates with multiple boom and bust cycles, rosy promises of a better future, and irksome disappointments. So, what is it this time? Is AI going to revamp the way business processes are executed or will it be another disaster?
Research suggests that AI is finally ready to deliver real-life business results. With more than 74% of customers making their purchasing decisions online, algorithms getting more precise and sophisticated, and the availability of enormous amounts of data, the right recipe is almost here.
Under such circumstances, it’s important to understand how users are adopting AI technologies. To delve deep into the analysis, McKinsey conducted thorough research that offers a snapshot of the current state of the rapidly changing AI industry, looking through the lens of both suppliers and users to come up with a more robust view of the economic potential of AI and how it will unfold.
Upheaval of AI
The idea of artificial intelligence dates back to 1950, when Alan Turing, an English mathematician proposed the prodigal Turing test to analyze whether a computer can communicate well enough to convince a human that, it too, is a human. This was swiftly followed by the development of the first AI program at the Carnegie Mellon University. From 1970 onwards, the growth slumped due to a lack of practical applications of AI. However, it again picked up in the 2000s thanks to faster computers and deep learning algorithms. Advances in the field of Graphics Processing Units have facilitated faster training of deep learning algorithms and have boosted the functional capabilities of AI, like image and pattern recognition.
Growth and lag
Though tech giants are the pioneers in investing in AI technologies, commercial adoption is still lagging. Indeed, internal investment by large corporations dominates – it amounted to approximately $18 billion to $27 billion in 2016. Despite all this, the overall adoption of AI has been pretty low so far. One primary reason is that most of the R&D work is largely focused on increasing the internal performance of the firm. According to a McKinsey survey, most business leaders are unsure about how AI can benefit the processes of their organization, where to obtain AI-powered applications, how to integrate them into their companies, and how to assess the RoI in the technology.
The bigger companies, on the other hand, are playing larger cards. Toyota, for example, has allocated $1 billion for new research. Similarly, IBM has pledged to invest $3 billion to make its Watson cognitive computing service a leader in the internet of things.
A significant challenge organizations face is the dearth of true experts in the field of AI. Companies are actively buying AI startups and using M&As to acquire not only promising technologies but also rare resources. This trend has also encouraged more venture capitals and private equity funds to invest in AI startups. However, investors are yet to recognize their RoI as only 10% of the AI based startups are generating good revenue.
Ready to adopt but not to act
Only 20% of businesses have adopted one or more AI technologies into their core business processes, while only 12% of the industries have moved past the experimental stage. Like in every new wave of technology, AI has some early and late adopters. Coincidentally, the companies that were the early adopters in the previous digital wave are maintaining the same pace in this one as well. Typical characteristics of early adopters are:
- they belong to sectors like cloud and big data and are already ahead in the digitization race
- they’re primarily larger companies that want to empower their employees and increase productivity using AI
- their executive leadership is well-informed about the effects of AI.
Also, early AI adopters tend to become serial adopters going forward.
Impact
Adoption is indeed a challenge when it comes to AI in business. Most companies have their own apprehensions in this regard. But what about the firms that have already adopted AI in their core business processes? Let’s have a look at how AI can shape different functions across the whole value chain and in diverse sectors.
AI implemented at scale provides diverse returns
AI can help organizations gain a competitive advantage by helping them estimate demand. It allows businesses to optimize their supply chain and design better offerings by finding meaning in disparate information and adjusting to random variations to discern trends that can be acted upon.
If we talk about the retail sector, in some cases, AI-based approaches to demand forecasting are expected to reduce forecasting errors by 30 to 50% from conventional approaches. Lost sales due to unavailability of a product can be reduced up to 65% and there can be a reduction of 5-10% and 25-40% respectively in costs related to transport & warehousing, and supply chain administration.
When it comes to matching supply and demand, electric utilities are a special case – they need to make predictions on a real-time basis. National Grid in the United Kingdom in collaboration with DeepMind, an AI startup bought by Google in 2014, is predicting supply and demand variations based on weather-related variables. The ultimate objective is to increase the use of renewable power and simultaneously decrease energy consumption by 10%.
Another crucial manner in which AI can impact firms is in the R&D departments, partly helping researchers quickly assess whether a prototype would be likely to succeed or fail in the market, and why AI-based approaches can increase the time to market by almost 10-15% and can further result in productivity gains of 10%.
The second area in which AI can help create value is manufacturing and production in both products and services. This can be done by bringing in automation to compliment teams of people. For instance, in Ocado, the U.S online supermarket, AI and robotics are embedded in its core operations. Robots take the units to and from the conveyor plants and deliver them to the human packers just in time to put the products into shopping bags. With computer vision, object recognition, and semantic segmentation, robots are able to recognize the materials that they are dealing with, making them more flexible and autonomous. This is a commendable shift from the current scenario where robots are not able to recognize the shape and pattern of objects given dynamic environment. With the help of deep learning, new AI-enhanced robots will be able to handle objects without the objects being in predefined positions. This is because they will be equipped with unsupervised learning engines. This capability leads to more precise makeovers and overall improved robustness of processes.
Charge right, deliver right
AI can be used to precisely map prices with current demand and supply. For example, prices of airlines and hotels fluctuate during the peak seasons when compared to regular ones. Today, customers are highly price-sensitive, primarily because they have multiple options and platforms to compare prices of individual offerings. Therefore, the demand for intelligent pricing is high. There are many factors that define the optimal price of a product, like the day of the week, the month, competitor’s prices, channels, and devices. AI can take the cumulative score of all the attributes and set up automatic product prices according to demand.
Give customers what they want
Today, companies don’t have the resources or manpower to provide customized attention to every customer. Personalized attention is given to only a few that are likely to contribute significantly to the bottom line, primarily because of the huge back-end cost. AI technologies such as computer vision and machine learning can open a scaled-down version of the experience to many more people. For example, if you’re in a shopping mart and you pick up a pack of brown bread, cameras could relay the information to an AI application. Based on his previous purchases, the customer can be recommended a diet butter to go with the bread.
Amazon Go is another similar example. It is a retail shop in Seattle where consumers can swipe into the store, pick up their goods, and walk out without having to stand in a queue for billing. The mechanism works like this: when customers swipe in, their face is recorded. As they pick up their products, they’re billed for the same. When they leave the store, the amount is debited from their account and the invoice is emailed. The whole process is simple and, at the same time, provides customers with an amazing shopping experience.
In sectors such as healthcare and education, personalized customer experiences have huge value. For instance, treatment for diseases like cancer can be provided based on a patient’s medical history and genetic makeup. In education, adaptive learning has been a growing trend across schools in Europe and North America. This attempts to overpower the limitations of classroom training by preparing a personalized schedule for each child based on his prior knowledge and understanding of the subject.
An overview of ongoing challenges and probable solutions
Encourage active participation and prepare for the future
Presently, AI is being adopted mostly by tech giants. Broader adoption of AI by smaller firms will drive growth, increase productivity, and lead to better wages. Though most industries claim that the percentage of employees who will lose their jobs due to automation will not be huge, there might still be people whose skills and capabilities are mismatched to the work that needs to be done. Under such circumstances, governments may have to rethink models of social support.
Resolve ethical, legal, and regulatory issues
The greatest challenge is to establish algorithmic transparency and accountability. AI algorithms will be trained by real-world data – the world that is racist, sexist, and biased in many ways. So, there is a high probability that the training data will be biased. When AI algorithms learn from biased training data, they internalize the biases, exacerbating those problems.
Another concern is privacy – who will take responsibility for sensitive data from sectors like banking and healthcare? Also, since data is the fuel on which AI programs run, regulations are needed to ensure that relevant data is readily available.
Way ahead
Do you remember the huge hue and cry from the time when computers first emerged into the market? The present situation is similar. While few occupations are fully automatable, 60% of all occupations have at least 30% technically automatable activities. Not all AI applications focus on replacing labour. In fact, about 80% of AI adopters aim to improve capital efficiency through automation. The indication is clear. Employees need to update their skills to collaborate with machines, not work against them. They must get more human by increasing their emotional quotient and filling in those gaps which, perhaps, a machine can’t.
After decades of hopes and disappointment, AI is back and could drive profound changes in the global economy. The change is here to stay. It depends on us as individuals and as a society on how fast we adapt to that change.
