Artificial Intelligence (AI) and Machine Learning (ML) are rapidly evolving to help companies with both digital transformation and innovation. There has been a lot of hype and discussion about these topics, and, in a very short time, we have moved the conversation from “AI is cool” to “AI can drive specific business outcomes.”
Experience has allowed companies to clarify the economics of AI and reduce time to value. In addition, the technology has continually improved, with advancements including:
- extracting unstructured data for improved insights and processes
- moving from simple chatbots to more sophisticated conversational assistants that are smart, use more natural language interaction (text and/or voice), and enable the initiation of transactions
- integrating operational and knowledge management systems
As our clients start to understand more about AI/ML, there are 2 key questions typically asked:
Question 1: “How do AI and Machine Learning differ from traditional programmed systems?”
There are three different capabilities that AI/ML typically extends and enhances into existing applications. One aspect is understanding – enabling systems to understand language, other unstructured data, including pictures, just like humans do. A second aspect is reasoning – enabling systems to grasp underlying concepts, form hypotheses, and infer and extract ideas, similar to humans. The third aspect is focused on learning – improving over time and avoiding repeated mistakes. These capabilities, often referred to as U-R-L (understand, reason, learn), are easily integrated into existing applications as consumable APIs, can reduce time to value.
Question 2: “Where is the value from Artificial Intelligence/Machine Learning?”
As we look across client environments, most clients are very comfortable with structured data that is within their firewall – things like transaction systems, customer records, and even predictive models. Many analysts estimate that 80% of data created today is unstructured, which requires clients to expand their current perspectives in 2 dimensions:
- Structured to unstructured, such as documents, transcripts, social media, weather, images, IoT and sensor data, and news
- Within the firewall to outside the firewall, including public data, licensed private data, and new types/sources being created every day
From a value perspective, the enticing aspect of AI and machine learning is connecting these structured and unstructured data types within and outside these walls. This enables richer, more, and unexpected insights, new business processes, and improved workflows at dramatically reduced cost levels.
As AI and machine learning mature, three use cases patterns have emerged around customer care, human capital management and the rethinking/reimagining of processes thanks to insights gained from unstructured data and natural language capabilities.
We’ll expand on these use cases and how we are applying AI/ML for our clients in our next blog post.
