Reshaping Business With Artificial Intelligence 2 10-01

Figure 5

Organizations expect to create competitive advantage from AI — but also anticipate increased competition. 

Executives simultaneously recognize that their organization will not likely be the sole beneficiary of AI in their markets. Respondents expect that both new entrants and incumbents would similarly see the potential for benefits. Three-quarters of respondents foresee new competitors using AI to enter their markets while 69% expect current competitors to adopt AI in their businesses. Furthermore, they realize that suppliers and customers in their business ecosystem will increasingly expect them to use AI.

4. Disparity in Adoption and Understanding

Despite high expectations, business adoption of AI is at a very early stage: There is a disparity between expectation and action. Although four in five executives agree that AI is a strategic opportunity for their organization, only about one in four has incorporated AI in some offerings or processes. Only one in 20 has extensively incorporated AI in their offerings or processes. (See Figure 6.)

Figure 6

Only about a quarter of all organizations have adopted AI so far.


The differences in adoption can be striking, particularly within the same industry. For example, Ping An, which employs 110 data scientists, has launched about 30 CEO-sponsored AI initiatives that support, in part, its vision “that technology will be the key driver to deliver top-line growth for the company in the years to come,” says the company’s chief innovation officer, Jonathan Larsen. Yet in sharp contrast, elsewhere in the insurance industry, other large companies’ AI initiatives are limited to “experimenting with chatbots,” as a senior executive at a large Western insurer describes his company’s AI program.

Organizations also report significant differences in their overall understanding of AI. For example, 16% of respondents strongly agreed that their organization understands the costs of developing AI-based products and services. And almost the same percentage (17%) strongly disagreed that their organization understands these costs. Similarly, while 19% of respondents strongly agreed that their organization understands the data required to train AI algorithms, 16% strongly disagreed that their organization has that understanding.

Combining survey responses to questions around AI understanding and adoption, four distinct organizational maturity clusters emerged: Pioneers, Investigators, Experimenters, and Passives.2

• Pioneers (19%): Organizations that both understand and have adopted AI. These organizations are on the leading edge of incorporating AI into both their organization’s offerings and internal processes.

• Investigators (32%): Organizations that understand AI but are not deploying it beyond the pilot stage. Their investigation into what AI may offer emphasizes looking before leaping.

• Experimenters (13%): Organizations that are piloting or adopting AI without deep understanding. These organizations are learning by doing.

• Passives (36%): Organizations with no adoption or much understanding of AI.

If expectations and sense of opportunity are so high, what prevents organizations from adopting AI? Even in industries with extensive histories of integrating new technologies and managing data, barriers to AI adoption can be difficult to overcome. In financial services, for example, Simon Smiles, chief investment officer, ultra high net worth at UBS, puts it this way: “The potential for larger-scale financial institutions to leverage technology more actively, including artificial intelligence, within their business, and to harness their data to deliver a better client experience to the end user, is huge. The question there is whether these traditional institutions will actually grab the opportunity.” Taking advantage of AI opportunities requires organizational commitment to get past the inevitable difficulties that accompany many AI initiatives.

These differences are less about technological limitations and much more about business. In aggregate, respondents ranked competing investment priorities and unclear business cases as more significant barriers to AI implementation than technology capabilities. Evans of Airbus makes the critical distinction: “Well, strictly speaking, we don’t invest in AI. We don’t invest in natural language processing. We don’t invest in image analytics. We’re always investing in a business problem.” Airbus turned to AI because it solved a business problem; it made business sense to invest in AI instead of other approaches.

Smiles at UBS notes that organizations do not all face the same challenges. With respect to incumbents and fintech startups, he says: “There is a bifurcation between the groups that have the scale needed to develop incredibly valuable platforms and those unencumbered by legacy business models and systems to arguably have the better model going forward, but don’t have the clients and accompanying data to capitalize fully on the opportunity.” Differences like these lead to differences in rates of AI adoption.

Barriers to Adoption

The clusters of organizations demonstrate how barriers to AI differ and affect rates of adoption. (See Figure 7.) Pioneers have overcome issues related to understanding: three-quarters of these companies have identified business cases for AI. Senior executives are leading organizational AI initiatives. Their biggest hurdles are grappling with the practicalities of developing or acquiring the requisite AI talent and addressing competing priorities for AI investment. They are also much more likely to be attuned to the security concerns resulting from AI adoption. Passives, by contrast, have yet to come to grips with what AI can do for them. They have not identified solid business cases that meet their investment criteria. Leadership may not be on board. Technology is a hurdle. Many are not yet even aware of the difficulties in sourcing and deploying talent with AI expertise.

Figure 7

While AI talent limits Pioneers, Passives don’t yet discern a business case for AI.

Our clustering also reveals nuanced differences in understanding among the clusters.

• Business potential: AI may change how organizations create business value. Pioneers (91%) and Investigators (90%) are much more likely to report that their organization recognizes how AI affects business value than Experimenters (32%) and Passives (23%). Evans at Airbus reports that “there was no question of value; it was trying to address an in-service issue on one of our aircraft.”

• Workplace implications: Integrating the capabilities of humans and machines is a looming issue. AI stands to change much of the daily work environment. Pioneers and Investigators better appreciate that the presence of machines in the workplace will change behavior within the organization. Julie Shah, an associate professor of aeronautics at MIT, says, “What people don’t talk about is the integration problem. Even if you can develop the system to do very focused, individual tasks for what people are doing today, as long as you can’t entirely remove the person from the process, you have a new problem that arises — which is coordinating the work of, or even communication between, people and these AI systems. And that interaction problem is still a very difficult problem for us, and it’s currently unsolved.”

• Industry context: Organizations operate in regulatory and industry contexts; respondents from Experimenter and Passive organizations do not feel that their organization appreciates how AI may affect industry power dynamics.

5. The Need for Data, Training, and Algorithms

Perhaps the most telling difference among the four maturity clusters is in their understanding of the critical interdependence between data and AI algorithms. Compared to Passives, Pioneers are 12 times more likely to understand the process for training algorithms, 10 times more likely to understand the development costs of AI-based products and services, and 8 times more likely to understand the data that’s needed for training AI algorithms. (See Figure 8.)

Figure 8

Organizations have different levels of understanding for AI-related technology and business context.

Most organizations represented in the survey have little understanding of the need to train AI algorithms on their data so they can recognize the sort of problem patterns that Airbus’s AI application revealed. Less than half of respondents said their organization understands the processes required to train algorithms or the data needs of algorithms.

Generating business value from AI is directly connected to effective training of AI algorithms. Many current AI applications start with one or more “naked” algorithms that become intelligent only upon being trained (predominantly on company-specific data). Successful training depends on having well-developed information systems that can pull together relevant training data. Many Pioneers already have robust data and analytics infrastructures along with a broad understanding of what it takes to develop the data for training AI algorithms. Investigators and Experimenters, by contrast, struggle because they have little analytics expertise and keep their data largely in silos, where it is difficult to integrate. While over half of Pioneer organizations invest significantly in data and training,
organizations from the other maturity clusters invest substantially less. For example, only one-quarter of Investigators have made significant investments in AI technology, the data required to train AI algorithms, and processes to support that training.

Misunderstandings About Data for AI

Our research revealed several data-related misconceptions. One misunderstanding is that sophisticated AI algorithms alone can provide valuable business solutions without sufficient data. Jacob Spoelstra, director of data science at Microsoft, observes:

I think there’s still a pretty low maturity level in terms of people’s understanding of what can be done through machine learning. A mistake we often see is that organizations don’t have the historical data required for the algorithms to extract patterns for robust predictions. For example, they’ll bring us in to build a predictive maintenance solution for them, and then we’ll find out that there are very few, if any, recorded failures. They expect AI to predict when there will be a failure, even though there are no examples to learn from.

No amount of algorithmic sophistication will overcome a lack of data. This is particularly relevant as organizations work to use AI to advance the frontiers of their performance.

Some forms of data scarcity go unrecognized: Positive results alone may not be enough for training AI. Citrine Informatics, a materials-aware AI platform helping to accelerate product development, uses data from both published experiments (which are biased toward successful experiments) and unpublished experiments (which include failed experiments) through a large network of relationships with research institutions. “Negative data is almost never published, but the corpus of negative results is critical for building an unbiased database,” says Bryce Meredig, Citrine’s cofounder and chief science officer. This approach has allowed Citrine to cut R&D time in half for specific applications. W.L. Gore & Associates, Inc., developer of Gore-Tex waterproof fabric, similarly records both successful and unsuccessful results in its push to innovate; knowing what does not work helps it to know where to explore next.

Sophisticated algorithms can sometimes overcome limited data if its quality is high, but bad data is simply paralyzing. Data collection and preparation are typically the most time-consuming activities in developing an AI-based application, much more so than selecting and tuning a model. As Airbus’ Evans says:

For every new project that we build, there’s an investment in combining the data. There’s an investment sometimes in bringing in new sources to the data platform. But we’re also able to reuse all of the work that we’ve done in the past, because we can manage those business objects effectively. Each and every project becomes faster. The upfront costs, the nonrecurring costs, of development are lower. And we’re able to, with each project, add more value and more business content to that data lake.

Pioneer organizations understand the value of their data infrastructure to fuel AI algorithms.
Additionally, companies sometimes erroneously believe that they already have access to the data they need to exploit AI. Data ownership is a vexing problem for managers across all industries. Some data is proprietary, and the organizations that own it may have little incentive to make it available to others. Other data is fragmented across data sources, requiring consolidation and agreements with multiple other organizations in order to get more complete information for training AI systems. In other cases, ownership of important data may be uncertain or contested. Getting business value from AI may be theoretically possible but pragmatically difficult.

Even if the organization owns the data it needs, fragmentation across multiple systems can hinder the process of training AI algorithms. Agus Sudjianto, executive vice president of corporate model risk at Wells Fargo & Co., puts it this way:

A big component of what we do is dealing with unstructured data, such as text mining, and analyzing enormous quantities of transaction data, looking at patterns. We work on continuously improving our customer experience as well as decision-making in terms of customer prospecting, credit approval, and financial crime detection. In all these fields, there are significant opportunities to apply AI, but in a very large organization, data is often fragmented. This is the core issue of the large corporation — dealing with data strategically.

Make Versus Buy

The need to train AI algorithms with appropriate data has wide-ranging implications for the traditional make-versus-buy decision that companies typically face with new technology investments. Generating value from AI is more complex than simply making or buying AI for a business process. Training AI algorithms involves a variety of skills, including understanding how to build algorithms, how to collect and integrate the relevant data for training purposes, and how to supervise the training of the algorithm. “We have to bring in people from different disciplines. And then, of course, we need the machine learning and AI people,” says Sudjianto. “Somebody who can lead that type of team holistically is very important.”

Pioneers rely heavily on developing internal skills through training or hiring. Organizations with less experience and understanding of AI put more emphasis on gaining access to outsourced AI-related skills, but this triggers some problems. (See Figure 9.)

Figure 9

Pioneers build AI-related skills through training and hiring, while Passives more heavily rely on external resources. 

The chief information officer of a large pharma company describes the products and services that AI vendors provide as “very young children.” The AI tech suppliers “require us to give them tons of information to allow them to learn,” he says, reflecting his frustration. “The amount of effort it takes to get the AI-based service to age 17, or 18, or 21 does not appear worth it yet. We believe the juice is not worth the squeeze.”
To be sure, for some support functions, such as IT management and payroll support, companies might choose to outsource the entire process (and pass along all of their data). Even if companies expect to rely largely on external support, they need their own people who know how to structure the problem, handle the data, and stay aware of evolving opportunities. “Five years ago, we would have leveraged labor arbitrage arrangements with large outsourcers to access lower cost human labor to do that work,” the pharma company CIO says. “What the vendors have done in the meantime is begin to automate those processes, oftentimes on our systems using our infrastructure, but using their technology. And I would not want it to be characterized as just rule-based. They actually have quite a bit more sophisticated heuristics to automate those things.” But such an approach is clearly not suited for companies’ distinctive offerings or core processes.
Eric Horvitz, director of Microsoft Research, argues that the tech sector is quickly catching up with the new model of offering technology tools to use with proprietary data, or “providing industry with toolsets, computation, and storage that helps to democratize AI.” Many AI algorithms and tools are already in the public domain, including Google’s TensorFlow, GitHub, and application programming interfaces from tech vendors. According to Horvitz:

Because this is a competitive space now in itself, the tools are getting easier to use and people that are there to help sell, market, and use these tools are becoming more efficacious in their abilities. That doesn’t mean that people don’t need to have their own in-house expertise and experts. While the tools and services are out there and that will make things easier, it is still going to be important for organizations to have their own experts in machine learning and AI more generally.

Privacy and Regulation

The data and the algorithms constituting AI cannot simply be accurate and high performing; they also need to satisfy privacy concerns and meet regulatory requirements. Yet only half the respondents in our survey agree that their industries have established data privacy rules.

Ensuring data privacy depends on having strong data governance practices. Pioneers (73%) are far more likely to have good data governance practices than the Experimenters (34%) and Passives (30%). (See Figure 10.) This large gap represents another barrier for companies that are behind in developing their AI capabilities.

Figure 10

Pioneers rate their companies higher across general management and leadership dimensions.


The data issues can be pronounced in heavily regulated industries such as insurance, which is shifting from a historic model based on risk pooling toward an approach that incorporates elements that predict specific risks. But some attributes are off limits. For example, while sex and religion factors could be used to predict some risks, they are unacceptable to regulators in some applications and jurisdictions.

Regulators in other financial markets also have stringent transparency requirements. As Wells Fargo’s Sudjianto says: “Models have to be very, very transparent and checked by the regulators all the time. When we choose not to use machine learning as the final model, it’s because regulatory requirements oftentimes demand solutions be less ‘black box’ and something the regulator can see very clearly. But we use machine learning algorithms to assess the model’s non-linear construction, variables and features entered, and as a benchmark for how well the traditional model performs.”

As technology races ahead of consumer expectations and preferences, companies and the public sector tread an increasingly thin line between their AI initiatives, privacy protections, and customer service. Some financial services providers are using voice-recognition technology to identify customers on the phone to save time verifying identity. Customers welcome rather than balk at this experience, in part because they value the service and trust the company not to misuse the capability or the data that enables it. Likewise, a technology vendor offers an AI-based service to help call center operators recognize when customers are getting frustrated, using real-time sentiment analysis of voice data. Less welcome applications may be on the horizon, however. In a few years, any of the 170 million installed cameras in China or the 50 million cameras in the U.S. will be able to recognize faces. In fact, jaywalkers in Shanghai can already be fined (or shamed) based on such images.

6. Beyond Technology, Management challenges.

AI requires more than data mastery. Companies also face many managerial challenges in introducing AI into their organizations.

Unsurprisingly, respondents at Pioneer organizations rate their companies higher in several general management and leadership areas: vision and leadership, openness and ability to change, long-term thinking, close alignment between business and technology strategy, and effective collaboration. As with other technology-driven transformations, these are essential general capabilities for high-performing companies.

However, there are also some specific challenges: Executives may still need to (1) learn more about AI; (2) deepen their perspective on how to organize their business around AI; and (3) develop a more expansive view of the competitive landscape in which their business operates.

Challenge 1: Develop an Intuitive Understanding of AI

The notion that executives and other managers need at least a basic understanding of AI is echoed by executives and academics. J.D. Elliott, director of enterprise data management at TIAA, a Fortune 100 financial services organization with nearly $1 trillion in assets under management, adds, “I don’t think that every frontline manager needs to understand the difference between deep and shallow learning within a neural network. But I think a basic understanding that — through the use of analytics and by leveraging data — we do have techniques that will produce better and more accurate results and decisions than gut instinct is important.” Avi Goldfarb, professor of marketing at the University of Toronto's Rotman School of Management, notes, “You worry that the unsophisticated manager might see one prediction work once and think that it’s always good, or see one prediction that was bad and think it’s always bad.” Joi Ito, head of the MIT Media Lab, contends that “every manager has to develop an intuitive understanding of AI.”

To develop their understanding of digital, many executives have taken trips to Silicon Valley to experience digital natives, design-thinking approaches, fail-fast cultures, and more. While these are all core to building digital businesses, such trips are not particularly rewarding to learn about AI. For those who have already been exposed to the marvels of robots, self-driving vehicles, or poker-playing machines, there is little new to experience at AI companies. Instead, managers should take some time to learn the basics, possibly starting with simple online courses or online tools. They should understand how programs learn from data, maybe the most important facet of understanding how AI can benefit a particular business.

Challenge 2: Organize for AI

Adopting AI broadly across the enterprise will likely place a premium on soft skills and organizational flexibility that enable new forms of collaboration, including project teams composed of humans and machines.

Our survey finds companies exploring many approaches to developing AI capabilities. Pioneers are relatively evenly split among centralized, distributed, and hybrid organizational models. Investigators and Experimenters also pursue a mix of approaches, but almost 30% of both clusters have not yet set clear responsibility for AI in their organization. Some 70% of Passives also have not even started to lay out clear responsibilities for AI initiatives, perhaps (in part) because fewer than 50% of Passives see AI having a large effect on their processes and offerings in the next five years.
Ultimately, a hybrid model may make the most sense since many companies need AI resources both centrally and locally. TIAA, for example, has an analytics center of excellence and a number of decentralized groups. “The center of excellence is not intended to be the group that will provide all analytics for the entire organization. It provides expertise, guidance, and direction to other internal teams that are working to deploy AI and analytics,” says TIAA’s Elliott.

While companies in all four clusters rate cultural resistance to AI approaches relatively low on the list of barriers, only about half said that their company understands the required changes of knowledge and skills for future AI needs. Jessica Tan, group executive vice president, group chief operating officer, and chief information officer of Ping An, says the biggest challenges at her company have been getting units to work together; acknowledging the fact that “humans don’t want to train algorithms”; establishing centralized and decentralized technology teams; and finding the right people. It's looking in particular for three types of people: technical people who have the means to try different ways of working, technical people who understand specific business domains, and people with consulting or project management skills who are able to network and bring them all together.

Challenge 3: Re-think the Competitive Landscape

More than 60% of respondents say that a strategy for Al is urgent for their organizations, but only half of those say their organizations have a strategy in place. (See Figure 11.) In terms of company size, the largest companies (those with more than 100,000 employees) are the most likely to have an AI strategy, but only half (56%) have one.                         

FigurFigure 11

While the majority of organizations see developing an AI strategy as urgent, only half already have one.

Amy Hoe, chief technology and operations officer of insurer FWD Group, says that she sees access to data as key for competitive advantage for her company. FWD aims to secure a wide range of data sources, including partnerships with other companies, such as telecommunications companies and ride-hailing services, its customer base, agencies, social media, the public domain, and external data analysis providers. As the volume of data doubles every few years, gaining privileged access Gto data is nonstop work.

Go to Page  1  3