Skip to main content

Six design principles for artificial intelligence in digital business

(Image credit: Image Credit: John Williams RUS / Shutterstock)

Artificial intelligence (AI) is capable of augmenting and automating decisions or tasks currently performed by humans, which makes it an indispensable tool for digital business transformation. With the help of AI, organisations can hope to reduce labour costs, generate novel business models, and generally streamline processes and improve standards of customer service. Nevertheless, it is important to stay pragmatic in approach, since the vast majority of AI technologies remain in their infancy.

To navigate this issue of AI technology immaturity, CIOs should ensure that applications intended to serve a strategic business purpose — such as maximising revenue or scaling certain services — are designed for strategic effect

Gartner has identified and outlined six design principles to help CIOs evaluate each AI application offering with strategic intent. These are applications intended to help achieve business results, not just operational improvements. It is not necessary for applications to adhere to all six principles. However, designs which show fewer than two principles should be reconsidered.

Design principle no. 1: Anticipate the future

When applied to digital business, AI generates insights which lead directly to business execution. Strategic AI solutions are capable of providing granular insights which can suggest how particular customers or markets will behave in specific situations in the future, and what the business can do to influence their decisions. If an application can deliver proven, trustworthy insights, it will reap the reward of being adopted and relied on by more enterprises to guide future execution systems.

AI can produce more granular insights that are better tailored to individual situations than conventional analytics applications. Therefore, AI applications can reduce false readings — the more trustworthy the insights, the more enterprises are likely to rely on them to guide execution systems.

Design principle no. 2: Act autonomously

The value of AI applications lies in its the automation of manual processes. However, the technology can also enable the autonomous operation of a business. A strategic AI application that acts autonomously does not require human direction. This autonomy, in turn, produces considerable productivity gains by complementing the work completed by humans and freeing up the workforce to carry out more personalised tasks.

Those responsible for designing AI applications for autonomous operations should ensure that the applications are located in proximity to the work being carried out, have close-to-real-time comprehension of processes and their condition, and possess the capability to make decisions on the spot.

Design principle no. 3: Connect to the customer

Intimate knowledge of markets and customers is integral to the success of digital business. To aid digital business initiatives, AI applications should aim to get as close to customers as possible. CIOs would be well-advised to follow the lead of digital behemoths whose popular technologies — run by AI — get between companies and customers.

Amazon’s Alexa and Apple’s Siri are pertinent examples. Consumers use these devices powered by these technologies as intermediaries in order to access the capabilities of external, third-party platforms. Consequently, Amazon and Apple are able to gather better data about customers than the companies whose services they are accessing. In a similar vein, CIOs may want to consider strategic AI applications that enable their business to capture critical information, with a view to building more intimate customer relationships in the long run.

Design principle no. 4: Elevate the physical

The advent of autonomous mines promises to transform the economics of the mining industry. Robotic surgery aims to deliver improved results for patients. Similarly, developers of strategic AI applications should strive to make a difference in the physical world. AI can effect physical change by working in tandem with other advanced technologies, or by facilitating new ways in which things can interact and collaborate. Increasingly sophisticated 3D printing is a prime example: GE Aviation now creates fan blades for its jet engines using 3D printing technology. Introducing AI into the mix can open up further possibilities and more complex use cases for 3D printing, such as adjusting the printing process to accommodate manufacturing where a number of variables need to be controlled simultaneously.  This is especially important when conditions are unpredictable: severe weather conditions, battlefield operations, conditions at sea, etc.

Design principle no. 5: Detect the invisible

AI can manage operations to a standard that human ability cannot match and, frankly, cannot even see or detect in a meaningful timeframe when a rapidly breaking event happens in cyberspace. Strategic AI applications should exploit the advantages of greater resolution and speed. Digital technology can enable an organisation to control things, events and outcomes with incredibly high precision. High-speed trading applications are already able to move large sums of money in a matter of nanoseconds. These applications are powered by algorithms that simultaneously consider stock prices, weather and political developments. This intelligence enables traders to execute millions of orders in mere seconds, giving their organisations the upper hand.  In many cases, these capabilities will become the minimum required to enter into a business and will drive newer capabilities to become competitive.

The digital world is increasingly granular. With 1 trillion connected things forecast by 2050, AI is the only possible solution to scaling decision making to such a level and to dealing with so much complexity.

Design principle no. 6: Manage risk

Security, risk and privacy issues constitute the biggest obstacles when it comes to developing AI applications. These become even greater issues when AI applications serve a strategic business purpose. A mistake doesn’t merely disrupt operations, it harms the core of a brand or enterprise.  In some cases, the AI applications themselves, capable of learning, develop their own recipe for completing their mission that may endanger others. As a result, CIOs should devise explicit plans to identify and mitigate risk in AI application designs, and define behavioural limits.

These six design principles should be used to evaluate all proposed AI applications. CIOs are advised to prioritise design principles differently, depending on the specific needs and circumstances of an individual enterprise. For example, cautious enterprises should focus on managing risk, while companies involved with the Internet of Things and blockchain should take greater heed of “elevating the physical” and “detecting the invisible.” With a combination of tailored principles, sufficient funding for data quality improvement, and executives at the helm who have a grasp of business strategies, organisations will be better equipped to overcome the often-daunting barriers to digital success.

Jorge Lopez, distinguished vice president, Gartner (opens in new tab)
Image Credit: John Williams RUS / Shutterstock

Jorge Lopez is a distinguished vice president at Gartner.