Adopt a customer-first strategy during the ‘A.I. Spring’

The A.I. Spring has arrived in full force, thanks to a couple of key trends coalescing at the same time.

First, prominent industry incumbents have opened up their cognitive platforms. New startups claim to base their entire platforms on machine learning capabilities. Some startups have opted to leverage new artificial intelligence (A.I.) technologies from incumbent enterprises such as Google (Google Cloud Platform), IBM (Bluemix) and Amazon (AWS). A symbiosis has formed between enterprises that own clouds and the startups that build businesses on them. These startups have tied themselves to the success of these technologies. Fortunately, enterprises won’t pull the rug from underneath them. Large companies receive valuable data and business thanks to them. So, the same enterprises offer it as a key capability to companies wishing to build A.I. services.

Secondly, the proliferation of data has forced businesses to harness A.I. as a competitive advantage. Businesses must leverage machine intelligence to gain deeper customer insights beyond traditional analytics. There are many buzzwords thrown around in the big data field (data lake is my personal favorite). Even so, anything with a microprocessor collects valuable knowledge about user behaviors. Using A.I. to understand this data helps us process it all. The next evolution of computation will process information measuring larger than the number of people on Earth.

A.I. is not a silver bullet

A technology-first strategy is inferior to a user-centered strategy. There are many examples of products that were too far ahead of their time. These products showcase great features that demo well but are never widely adopted. To be an A.I.-first company, businesses must completely reassess the needs of their customers. Technology-first companies often get lost because they lose sight of the customer experience.

Companies that embrace a user-first strategy will realize that A.I. is only one tool to create great product experiences. A.I. capabilities cannot offer silver bullets that help a business leapfrog the competition to capture market share. These APIs are so new that their true application is still not well understood. Cognitive models are not necessary when simple analytics would be acceptable. A.I.-based solutions need more development to be superior to human intuition. Unfortunately, many businesses conflate the adoption of new technologies with achieving competitive advantage. As a result, companies ship experiences that lead to customer dissatisfaction and churn.

For example, businesses hope chatbots can replace traditional apps, but these robotic agents provide terrible user experiences. Commerce companies hope they replace traditional purchase flows within their online experiences. Other firms believe chatbots can replace human interaction during a support engagement. These distinct experiences need significant UX focus to be great. Instead, most companies ship one bot that does it all. Conversations end up frustrating users, instead of providing delight to them. Instead, bots could provide initial analysis of a person’s preferences. Then, they could triage the request to a human who can provide Level 2 or 3 support.

Crafting a business strategy that benefits from A.I.

A.I. will provide its greatest service to companies when considered within a well-formed strategy. Companies should understand A.I. in the context of internal and external customers. If a business wishes to ride the cognitive wave, leaders should ask themselves the following questions:

1. Determine the as-is state: What is the current state of my business as it relates to satisfaction of employees and customers? Are customers suffering from unnecessary pain that is a direct result of my product’s usability or performance? Are employees unable to be efficient because of poor software or practices within our company?
2. Assess gaps in understanding: If there is pain internally and externally, do we have the necessary data to make improvements? How are we collecting this data? How is the data analyzed at all? Can a small fix provide big impact, without A.I.?
3. Define the end state: What is the desired experience for our customers? Do we need to improve ourselves before we can provide this experience? Or is it a matter of our products themselves? If we only need to improve our products, can we take advantage of user research to make better experiences?
4. Consider market constraints and timing: What is my competition doing to combat similar issues? Are we in the wrong business today? What technical and human resources do we have to shore up before we can capture customers? How much time do we have to adopt a new strategy?
5. Define a road map: What are the first steps to get to our end state (define the minimal viable offering)? What milestones do we set along the way to get to our end state? How much time do we have to achieve each milestone?

Adoption of A.I. will be a wasted capital expense if you haven’t deeply considered the questions above. If your end state requires new technology that an existing solution already employs, use that! You’ll gain obvious time and resource savings by applying your core competency to more important problems. It may be that the end state only requires a few changes upstream to have effects downstream. Recall the earlier commerce example: Companies can surface style suggestions using machine intelligence. Then a human takes care of the rest to close the sale. In this regard, A.I. reduces the search space of millions down to five to 10 choices. This both provides a time savings and human touch for the consumer.

A.I. adoption should be like other technologies that came before it: considered within a business context. When cloud computing first emerged, few companies could leap in because their platforms were too complex. This made it costly to move to the cloud without spending enormous amounts of capital. This initial investment can pay off, but only when executed carefully and always with the customer in mind. For example, banks are still hesitant to move to the cloud because legacy systems run on Cobol and backups are still made on tapes. Customers expect banks to have their accounts accurately tabulated, and any small error can cause major problems. It is no surprise, then, that banks have opted to adopt A.I. in niche areas of their businesses. First, they will experiment with A.I. to learn about the technology and its benefit to their customers. Businesses should balance a conservative strategy with high-pressure market forces.

Summary

The A.I. Spring has boom and bust potential. Deep learning will usher in a new period of technological experimentation. But businesses will adopt A.I. without a direct internal or external business benefit. This will lead to unnecessary costs and a loss in market position. A.I. technology will continue to mature and augment human capability in many ways. So, there is no substitute for considered strategy analysis that could lead to better execution without needing to apply A.I. In order for companies to properly embrace A.I., they should consider the end user benefit above all else.