There has been an unprecedented, and continuing, volume of data growth worldwide. By some estimates, 90 percent of all data on the internet is unstructured text.

Enterprises have an enormous opportunity to gain insights from this data, and it will become a competitive necessity to analyze it to achieve valuable and actionable insights to help shape business outcomes. In 2017, senior-level executives are increasingly looking to Artificial Intelligence (AI) innovations to help create information assets to fuel competitive advantage and transform enterprise strategy. According to the IDC Worldwide Semiannual Cognitive/Artificial Intelligence Systems Spending Guide, “identifying, understanding, and acting on the use cases, technologies, and growth opportunities for cognitive/AI systems will be a differentiating factor for most enterprises and the digital disruption caused by these technologies will be significant.”

However, there are two common questions that arise:

  • Is AI technology ready to deliver value today for my enterprise?
  • What should I focus on to drive transformation and create value?

A recent research report, “Can Artificial Intelligence Deliver for Today’s Enterprise?,” issued by Rage Frameworks based on data compiled by a survey of senior-level business executives, took a closer look at how artificial intelligence will be a key differentiator for businesses in 2017 and beyond. Here are some of the findings.

Enabling the intelligent enterprise

AI has quickly become a dominant enterprise technology as businesses strive to remain competitive. Eighty-two percent of research participants stated that their organizations would be using AI in 2017, compared to just 7 percent who said they were not expecting to use AI. The remaining 11 percent said they were strongly considering an AI deployment.

To ensure the successful adoption of AI in the enterprise, however, there are several challenges that current technologies have to overcome in order to power transformational enterprise applications. AI will help enable the intelligent enterprise by automating the detailed analysis of large volumes of structured and unstructured data using a linguistics-based approach. The research indicated that the top capabilities that are important to the AI solutions the market is looking to invest in, each identified by more than half of the respondents, include reasoning and traceability (the ability to understand why your AI solution came up with the results it did), at 55 percent, and natural language understanding (your AI solution understands context and relevance), at 53 percent.

Reasoning and traceability

Reasoning and traceability, or having the ability to comprehend the logic behind why an AI solution reached its conclusion, is essential for widespread adoption in many enterprise-level applications. However, many current AI technologies use a computational statistics approach and are essentially “black boxes,” meaning that this visibility is not apparent. When a recommendation from the AI technology is not intuitive, there is no way of tracing it back to source. It is also unknown if it is truly causal or spurious, and users therefore must go on blind faith.

For mission-critical processes such as contract review or wealth management, business users have to be able to trust that the AI engine’s reasoning is sound. Visibility of reasoning is also important in order to more easily improve the engine when there is a false positive or false negative. With a black box, the user has to find enough instances of the false positive or false negative and rebuild the engine. There is no way of knowing whether all variations or permutations of that error have been addressed or not. However, with the adoption of deep linguistic learning approaches, full and complete visibility of the reasoning can be achieved.

Natural language understanding context and relevance

Another challenge many modern AI technologies face is the inability to process language as humans do. This is because the majority of current AI approaches focus on natural language processing (NLP), largely driven by computational statistics that treat text as data rather than language — and use a pattern recognition technique. These methods do not attempt to understand text but instead simply transform it into data, and attempt to learn from the patterns in that data.

“Pattern recognition methods ultimately fail because of the inherent challenge of understanding context and relevance; context and meaning are often lost during the mechanical process of converting natural language into data. To comprehend the meaning of written text, enterprises need an approach that helps the AI technology understand text using its linguistic structure.”

This approach — transitioning from natural language processing to natural language understanding (NLU) — involves applying computational linguistics principles to reverse engineer the text back to its fundamental ideas, and realizing how the ideas were connected together to form sentences, paragraphs, and the full document. Processing the natural language text needs to be done in the right context, which can only be developed if the technology focuses on the language structure and not just on the words in the text as most current technologies do.

Understanding context is a multi­faceted challenge. First, the words in many languages can be used in multiple senses, so it is important to disambiguate word senses so their usage in a particular document can be accurately understood. Second, text documents often use domain-specific discourse models, (e.g. legal contracts, news articles, research reports, etc.). There are certain properties of such domain discourse models that should be incorporated in the AI technology in order to enhance NLU.

Many words may also be used as proxies within a document. AI technology must have a way to recognize and understand proxies like “Xerox” for “copy.” In some cases, text in a document may refer to knowledge that is not explicitly part of the text. Humans can understand this with prior knowledge. AI technologies, on the other hand, have to create a repository of global knowledge that can be retrieved to supplement the document text in order to gain full understanding of its meaning.

What we need

The market for enterprise AI technologies is growing at an unprecedented pace. However, to ensure its successful adoption, enterprise AI applications must meet the market demand for solutions that address reasoning and traceability, as well as context and relevance through NLU, to generate the actionable insights business users need to create competitive advantage.