Cognitive Search: At the Core of Cognitive Computing

What exactly is cognitive computing? Well, if you ask 10 academics or scientists, you’re likely to get 10 different answers. Look it up on Wikipedia. You’ll see that in the academic and scientific community, there is no agreed upon definition. It’s just marketing jargon. Ouch.

On the commercial side though, we may not have a precise definition for cognitive computing, but we’re very aware of its potential to help our businesses.

Cognitive computing and AI have increased the tools in our toolbox. Machine learning, deep learning, neural nets, natural language processing. These technologies are the foundation on which companies are building new capabilities that affect top-line revenue and bottom-line profitability.

Cognitive Search: At the Core of Cognitive Computing

Cognitive Experiences Have Search at the Core

Many analysts put very good enterprise search — what we call cognitive search — at the core of cognitive computing. And no matter what you call it, search today is a very different animal than search even 10 years ago. Forrester Research defines cognitive search as

“The new generation of enterprise search solutions that employ AI technologies such as natural language processing and machine learning to ingest, understand, organize, and query digital content from multiple data sources.”

 In large organizations, information is spread across hundreds of business applications and very possibly thousands of databases. Plus, there are unstructured data sources from ordinary file shares to NoSQL, MPP, and columnar databases, as well as data lakes built on Hadoop. So, of course, finding the information is the first challenge but by no means the only challenge. Or even the most complex.

“Building cognitive applications that provide humans with a natural-feeling interface to software requires technologies fundamental to search, such as NLP and indexing. Enterprises can demystify cognitive interactions by taking advantage of these familiar technologies.”

—    Forrester Research

If you build a search tool that can connect to any data source, that’s step one. But steps two through whatever are building a tool that learns from the way individuals interact with information. Two people may ask the same question, but that doesn’t mean the same answer will satisfy both.

That’s where context comes in. What jobs do these people do? How do they use the information they find? What kinds of follow up questions do they ask after their first query? Do they also add information to a data store? Over time, a cognitive search platform like Attivio’s learns these things. It begins to understand who you are. 

The Human Condition: Solving Problems 

When we talk with customers, a lot of what we talk about is problem solving. And, when you're trying to solve a problem, first you must remember the facts, find the facts, or find someone who knows the facts. Once you’ve found the information, you need to understand it in context — how are the pieces of information connected? Then, you can develop an approach to solving the problem. If you’re trying to automate that process, it’s inherently a search problem. But not necessarily search in its traditional form. That’s not cognitive search.

“More than half (54%) of global information workers are interrupted from their work a few times or more per month to spend time looking for or trying to get access to information, insights, and answers.”

—    Forrester Research

The key to getting off the ground with cognitive search solutions is picking use cases where the problem is clearly defined, the process goals achievable, and the expected ROI understood.

Any cognitive computing initiative should start with search. Digital transformation starts when you can unify knowledge from disparate sources and understand the intent behind search queries at an individual level. And that’s what I’ll talk about in my next post.

 

Gartner Magic Quadrant for Insight Engines 2019
Attivio was recognized for our completeness in vision and ability to execute.