Can you ever truly future-proof your business? Probably not, but more and more companies today are leveraging Artificial Intelligence (AI) to help them find and manage the “unknown unknowns” in our rapidly-changing business environment. In fact, findings from Aberdeen’s recent 2017 Big Data survey demonstrate that Best-in-Class companies are more likely to explore investment in AI technologies, including predictive analytics, data lakes, natural language processing (NLP), and real-time / streaming analytics (Figure 1).
Figure 1: Top Companies Prioritize Cutting-Edge, AI Technologies
AI in Action
Despite its relative nascence as an enterprise technology, the application of AI in a business setting is seemingly limitless. As the technologies continue to refine and evolve, it’s fair to wager that the theoretical will transform into the practical, and more companies will adopt AI in more areas of their business. The following are just a few ways in which the technologies can be (or are being) used:
Supply chain planning. With the ubiquity and expediency of data in today’s global economy, supply chain networks have become more flexible in their ability to accommodate more geographically dispersed elements, forming a broader supplier network. However, that trend only exacerbates the complexity of managing the flow of parts and goods through the supply chain in an efficient manner. By using machine learning (ML) algorithms, AI can help to better anticipate changes in lead times, predict price fluctuations, and recommend certain suppliers. Ultimately, this is making organizations more proactive than reactive in managing their supply chains.
Contact center optimization. The contact center is the nerve center of businesses. It has become a customer engagement hub where companies must seamlessly manage multiple channels to deliver consistent and personalized interactions. Contact centers using AI can automate analysis of historical customer traffic across each channel (e.g., phone, live chat and email), and optimize agent forecasting activities. This would mean minimizing understaffing and the associated customer churn, and overstaffing and the associated unnecessary labor costs. Another use case is analyzing previous phone conversations with the help of ML and speech analytics to determine keywords associated with customer churn. Using this capability, AI can help contact centers notify agents when a customer uses one of these keywords and determine customer churn risk. This, in turn, alerts the agent to take the next-best action to minimize the chances of losing the customer.
Fraud detection. Efforts to thwart fraud and cybercrime, particularly within credit card companies and other major financial institutions, have been around for decades. However, with the staggering volume of transactions taking place every minute of every day, and the array of vulnerable endpoints tied to the typical individual, traditional methods of fraud detection are being augmented with AI and deep learning (DL) algorithms. At its essence, fraud detection is about identifying anomalies and unusual behaviors hidden within a mountain of routine transactions. With the use of DL algorithms, the enormity of available data can actually work in favor of organizations looking to detect fraud. (The volume of data provides a rich foundation of transactions that can “train” these algorithms to identify what constitutes “normal” behavior, and what is considered “suspicious.”)
Customer sentiment analysis. One of the most enticing and approachable aspects of AI is the concept of natural language understanding (NLU). Sometimes used in enterprise search applications that process speech into text, or text into a data query, users can perform analytical activities simply by speaking or typing (i.e., “Show me sales by rep for Q2” or “What were our operating margins in 2015?”). Nowadays, more companies are turning to NLU to leverage text processing capabilities to develop a deeper understanding of their customers and prospects. With the volume of unstructured data being generated through social media channels, blogs, and message boards, companies can use AI to process massive quantities of customer-generated language. These efforts will help develop a clearer understanding of what customers want, and how to adapt their product roadmaps and pricing structures to accommodate those changing preferences.
When applying AI, the sky is the limit
The list of criteria required for applying AI effectively is fairly low. Any situation or business problem that involves a lot of data and / or disparate, variant data, is a good candidate. Detecting patterns, connections, and correlations in data is useful in too many areas of business to count, but companies should think about prioritizing those activities. Certain areas might be better beachheads for implementing AI, such as search / NLP within marketing, or augmenting predictive algorithms for sales forecasting. Any low-hanging-fruit applications that would require fewer resources and offer a greater short-term impact would be a good place to start.
To delve further into the emergence of AI in enterprises today, and the vast potential of AI-related technologies, read the complete research report here.