In a recent blog post, I highlighted the positive economic affect AI could have in the next ten years. To realize that benefit, businesses will need to affectively leverage data as a strategic asset to enable AI applications. Assuming a business has access to the right data assets; the question becomes how does it gain the most benefit from what it has? To this end, there are three key themes that companies should consider when shaping their approach: 1.) How can they meet customers’ needs more effectively, 2.) How can they better optimize internal operations, and 3.) How can they directly monetize the data assets at their disposal?
Almost every business survives on its ability to attract customers willing to pay for its products or services. Therefore, leveraging data to better understand customer behavior, expectations, needs, and service points of failures is paramount. In all of these cases there are existing, non-AI applications that may be adequate for meeting the needs of a large number of organizations. In fact, any company of scale has probably invested millions of dollars in technologies like CRM, decision support, and business intelligence. Just as likely, the data related to these systems is siloed; making it difficult for the company to gain a holistic understanding of the customer. This is where an AI-enabled, data-centric approach can truly provide value. For example, using tools like a data lake enables organizations to collect disparate information – structured to unstructured – into a centralized repository. Once gathered, the data can be utilized flexibly by a variety of machine learning-enabled tools that are designed to generate insights from data sets that are not fully cleansed and normalized. It’s not magic – appropriate skills and tools are required to achieve results – but this approach helps remove application-driven data barriers that prevent businesses from identifying patterns when dealing with their customers. In addition, it allows businesses to more easily inject new data sets that are not core customer information, but may be important to understanding their behavior. For instance, buying habits may be influenced by seasonal changes, so being able to inject weather forecasts into store inventory decisions may produce increased revenue or savings for a retailer.
In order to effectively act on customer behavior insights, companies will likely have to fine-tune their core operations and determine where AI may be applicable, and what data should be leveraged to enable it. In many ways this activity may be more straightforward than applying data for customer behavior analytics. Companies will have to make specific ROI judgments to determine where best to start, but there are potentially numerous quick wins for applying AI to owned data assets. In financial services, for example, reconciling transaction data breaks (e.g., missing fields, incorrectly entered data) is a key reason for the growth of back office operations staff over time. The industry is already actively adopting AI-enabled tools to identify breaks, flag them for resolution by human staff, and self-correct future breaks automatically based on learning from previous human intervention. Similarly, RPA and IPA tools are being used across multiple industries to both augment humans and automate processes (e.g., extracting and validating data, managing process hand-offs, cognitive agent support). Almost any operational function that deals with high volumes of information (e.g., accounts receivable/payable, compliance, call center) can benefit from AI-enabled tools to detect patterns and flag anomalies for resolution, saving significant amounts of time and money.
In addition to using data to improve the customer experience and business operations, companies should not discount the intrinsic value of their unique data assets. A company’s data assets have the potential for real value just like hard assets such as restaurant kitchen appliances or assembly line robots. For the most part, unless a company’s business is aggregating and selling data, this value is not recognized unless under unusual circumstances like a bankruptcy. However, given the world’s increasing adoption of AI-enabled tools requiring larger volumes of quality data, this may not be the case for much longer. The hunger for quality data will prompt the development of an ecosystem to support the exchange of differentiated data sets to help enable AI-powered applications. This already happens in a bilateral way (e.g., individual hedge funds using alternative data assets like weather forecasts, satellite imagery, and infrared readings of plants), but we can expect multiparty infrastructure like data trusts and exchanges to become commonplace in the near future. This infrastructure will enable companies to more easily monetize their data in more sophisticated ways, turning it into a tangible, balance sheet asset. Companies should be cognizant of this future and have a strategy for how they might reap these potential benefits while ensuring the core business is protected.
Companies using a data-centric, AI-enabled approach will benefit far more than their peers that hesitate to make this shift. All businesses need to consider how they can best leverage accessible data to drive better customer experiences and optimize their efficiency.
The next and final blog in this series will discuss how to properly maintain data assets, and why that’s important for your business.