Technology & Innovation

The CIO’s guide to artificial intelligence

By Whit Andrews, vice president and distinguished analyst at Gartner
Technology & Innovation
Published: 13 April 2018

I never knew what a whismo was before my client said it had a “whismo problem,” and then asked if artificial intelligence could help.

The company explained: About 30% of its inbound calls were from customers asking about order status. Thus, “whismo,” or “Where is my order?” The calls are expensive and unproductive.

Leadership wanted to know if artificial intelligence (AI) would be able to help manage the interactions. The short answer was yes, a virtual customer assistant could answer questions ranging from “Where is my order” to “How long will I have to wait?” But I said that a bigger question was if AI could help the company in even more impactful ways.

For example, I said, a whismo call is a call you want to get rid of when there’s a person answering the question, because fast resolution means a fulfilled ticket and the next call starting. But firms with AI-powered service answering have an infinite number of agents. Organisations can look at how they are using technology today during critical interactions with customers and consider how the value of that moment could be increased.

AI allows organisations to employ data from a wide variety of places and apply self-improving analysis that can take action. When combining information with other data about a given customer (for instance, X orders of Y products every Z weeks, with detail on when within those weeks the orders are placed), the company can use AI to further enrich the relationship beyond that interaction.

During future interactions, the data might enable the seller to ask questions specific to the customer, such as “We know you are frequently waiting on delivery. Would you like to subscribe to this product or order larger quantities?”

Key insights for CIOs
Savvy CIOs are experimenting jointly with business peers to discover top use cases for AI, to evaluate its potential to disrupt markets and remake existing business models. Here are three key insights for CIOs to know before they start a successful AI journey.

  1. Digital business is accelerating interest in AI at a pace that has left many CIOs hurrying to build an AI strategy and investment plan appropriate for their enterprise.
    Over the past few years, the pace of innovation in AI technologies has been staggering, predominantly coming from small vendors. CIOs are in the perfect position to educate their company’s CEO and board about recent AI developments and show how AI might influence their business and competitive landscape. By following this approach, CIOs can potentially flip the traditional engagement model between IT and the business, influencing business strategy from the beginning, rather than simply developing implementation projects that follow up on the executive team’s decisions. One-third of AI projects today are driven by IT, a higher proportion than any other business group.
  2. Deep learning, natural-language processing (NLP) and computer vision are leading areas of rapid technology advancement, and are the areas where CIOs need to build knowledge, expertise and skills.
    Recent breakthroughs in machine learning, big data, computer vision, and speech recognition are increasing AI’s commercial potential. But it requires organisations to adopt new skills and a new way of thinking about problems. CIOs must ensure that IT owns the strategy and governance of AI solutions. Although pilot AI experiments can start with a small investment, for full production rollout, the biggest area of investment is building and retaining the necessary talent. These skills include technical knowledge in specific AI technologies, data science, maintaining quality data, problem domain expertise, and skills to monitor, maintain and govern the environment.
  3. Market conditions for commercial success with AI technology are well-aligned, making AI safe enough for CIOs to investigate, experiment with and strategise about potential application use cases.
    Capabilities like voice recognition, NLP and image processing benefit from advances in big data processing and advanced analytical methods, such as machine learning and deep learning. Leading-edge AI technologies will play an increasingly important role in the top three business objectives often cited by CEOs — greater customer intimacy, increasing competitive advantage and improving efficiency. CIOs should look for cloud SaaS applications that apply AI to these areas. Greater experience with AI solutions will help CIOs to build business cases and identify the limitations in current-generation technologies to understand skills needed to fill talent gaps.

Businesses must not underestimate the importance of where AI is used. If they are interested in exploring AI, the all-important first step is to pursue something that is critical to the organisation. However, they need to do their due diligence and avoid the hype around AI.

Avoiding the hype
Hype isn’t always a bad thing. Within limits, it fosters attention, and triggers innovation and potential investment. However, too much hype may lead to false hopes and misguided planning assumptions. Although AI offers exciting possibilities, the huge increase in start-ups and established vendors claiming to offer AI products without any real differentiation has confused potential buyers and obscured the value of more straightforward, proven approaches.

As AI accelerates up the Hype Cycle with the promise to change aspects of business forever, CIOs have to distinguish between faux and real AI offerings. One way of doing this is by asking a vendor to describe the analytical model used in its AI solution and, from there, deduce how well the solution might perform in a given situation. Ask how the system learns and listen closely for indicators of self-learning, necessary training by humans or just fancy “rules” that have to be manually changed.

There are three key questions that should be put to vendors:

  1. What AI learning method is it proposing to use in its solution?
  2. What specific skills and level of experience are needed to be successful?
  3. How much training data is needed to “prime” the solution, and how often will it need to be retrained?

The answers to these questions go well beyond a traditional “demo.” Companies must understand how a vendor’s product uses AI and whether it would work well with the data and processes that already exist.

Another factor to consider is the justification for having AI in a product, as it introduces risks, complexity and costs.

Consequently, any vendor claiming that its product includes AI should also be able to explain how it will benefit the end user more than versions without the technology. Go beyond verifying that AI makes the product better, and get a sense of how a vendor’s AI-enabled product is superior to others in the market.

When comparing different AI products, CIOs should ask vendors how they manage risk with their AI products, and how it’s superior to their competitors’ means of doing so. This is vital, as many vendors do not understand the risks involved in using AI.

AI systems are not static and require vendors to be fully invested in improving their flexibility and resilience. Find out what vendors are doing to improve their offerings, whether by collaborating with independent data scientists or being active players in the industry. Cloud SaaS deployment facilitates continuous innovation from a vendor, and potentially other participants in the shared environment.

Moving forward with AI
Keep these considerations in mind as you adopt AI for critical business priorities:

  1. Look for ideas and possibilities in areas you couldn’t approach before because you didn’t have or couldn’t attract enough talented people.
  2. Learn the lessons that are unique to your organisation and minimise those that are more mainstream in nature.
  3. Survey and engage your highest-value workers about mundane aspects of their roles that can be addressed through AI.