How Many Features Does Your AI Understand?

We talk a lot about features at NexLP, but we don’t just mean technical features like interface elements. Of course, we think those are important, but what we really like to focus on are features of communication and how artificial intelligence technology can make better sense of them.

Next-Generation AI Understands Context

Context-aware artificial intelligence understands content, emotion and behavior as well as their relationship to each other. This gives you a full representation of the communications universe within your data.

Context-aware artificial intelligence understands content, emotion and behavior as well as their relationship to each other. This gives you a full representation of the communications universe within your data.

Understanding communication is difficult because there are dozens of elements within even the shortest communication. Think about the following sentence:

“I want you to move forward with your proposal.”

Just in this short sentence, we can see the following features:

  • The meaning of the words
  • There are two people involved
  • There is an implied intention to act
  • It is low pressure communication

Although you can process all of that while barely thinking about it, it’s a challenge for most technology, even when you consider just the meaning of the words.

For example, “moving forward” has a little nuance to it. In the example above, it means to follow through on something that had been proposed, but “moving forward” can also mean physical movement. Basic text mining software misses the nuance, and it also misses important emotional cues. 

For instance, a simple word change could shift the above sentence from low pressure to high pressure:

“I need you to move forward with your proposal.”

Using the word "need" instead of "want" creates more urgency. It's no longer just a desire for someone to act, but a demand.

The challenge of understanding language like this is what makes features so important. Next-generation AI not only understands the individual features within unstructured data, it understands the relationship between them.

This lets sophisticated AI understand complex concepts and identify patterns of behavior rather than simply identifying problematic words and phrases. After all, one high-pressure conversation might not mean anything. But a sudden spike in high-pressure conversations may signal a declining relationship with a client or something else that puts the organization at risk. 

Moving Forward with Smarter AI

Technology that only looks at the words being said misses essential context, and this can easily backfire for organizations looking to mitigate risk. For example, organizations involved in legal disputes have to be careful of how their technology defines privileged information. An overly broad definition causes organizations to withhold critical information and opens them to lawsuits. On the other hand, too narrow a definition would result in giving away proprietary or sensitive information.

Moving from context-less technology to context-aware technology will give businesses far greater ability to understand their unstructured data - for legal disputes and beyond. Context-aware technology is all about Features. The more Features AI understands in a data set, the more accurately and quickly it can extract truly relevant information to help users understand their data. With context-aware technology, users get:

  • A more nuanced understanding of a data universe
  • The ability to act more quickly from unstructured data
  • More time focusing on truly relevant information

The nuanced understanding of unstructured data is especially key for high-risk incidents, because what you’re looking for isn’t always going to be said explicitly. Someone attempting to collude with the competition probably isn't going to come out and ask your competitor to collude with them. However, they might talk about sharing proprietary information. They might exhibit abnormal behavior (e.g. emailing outside of their usual business hours).

Next-generation AI can identify these concepts and changes in behavior because it understands the Features of a broader communications universe. extending across emails, chat sessions, legal and business documents and plenty of other forms of communication that take place across ant given organization.

If you're thinking about investing in artificial intelligence software, ask potential vendors how many Features their solution understands. This will give you insight into how sophisticated their AI really is.

Interested in learning more about AI and machine learning? Subscribe for exclusive content from the team at NexLP.

Jay Leib

Jay Leib is the CEO of NexLP, Inc. Jay has successfully founded and led software start-ups throughout his career. Most recently, Mr. Leib was at Chicago-based Relativity (formerly kCura) as the Chief Strategy Officer. Relatiivity is a Chicago based company focused on the eDiscovery vertical.