A diagram/model of what AI is being used for
PHOTO: Shutterstock

Search has always been a key enterprise technology going back to the days of the first enterprise content management systems. This is hardly surprising given how important finding the right data is for any of the applications used by enterprises in their business processes.  Since the rise of big data and the use of big data sets, search has become even more important. If enterprise data is the real wealth of a business, then search is the tool that uncovers that wealth. But what do you do with the increasingly large amounts of data that enterprises now have access to? Like many other areas of business technology search vendors are turning to artificial intelligence to enhance the power of their search offerings.

AI-Driven Search At Work

There are many kinds of AI-driven search much of which is being adopted by enterprises to improve the performance of their websites. According to Daryl Plummer, vice president and Gartner Fellow  spoke at the Gartner Symposium/ITxpo 2017 in Orlando last October. He said by 2021, early adopter brands that redesign their websites to support visual and voice search will increase digital commerce revenue by 30 percent.  With visual and voice search rapidly increasing in popularity and on the way to being dominant mobile search modes, enterprises need to experiment to identify the best ways to capitalize on this consumer shift.

Many enterprises have taken heed. Los Angeles-based Forever 21 — a clothing outlet — is moving beyond text-based online search to offer shoppers AI-powered visual search and navigation. Their customers will no longer be limited to searching for a new looks online by typing fashion terms into a search bar. They will now be able to search items by simply clicking icons that represent the features they want in an outfit.

Dubai, UAE-based Bayt.com, the Middle East’s top job site has added new AI powered features that are designed to match talent with opportunity in a much more powerful and efficient manner enabling the 33,300,000 job seekers (Bayt’s figures) and the hundreds of thousands of users who visit the site every day to browse over 15,000 jobs, working to match job seekers with the right career opportunities.

Google has also recently launched a new job search tool as part of its Google for Jobs suite to capitalize on the $200 billion U.S. recruitment market. Google differentiates itself from LinkedIn, CareerBuilder, Monster and other online job boards by using machine learning and AI to help job seekers find jobs that are specifically tailored to their needs. 

Related Article:  8 Examples of Artificial Intelligence (AI) in the Workplace

The Elements of a Solid AI Search

So, what does AI-driven search offer Kavita Ganesan founder of Salt Lake City-based Opinosis Analytics pointed to three basic AI elements that search engines should have now. They include:

Query completion - Why not guide users to the right query instead of letting them blindly type out keywords that may yield poor results? Query completion also allows users to be productive as they don't need to type out every single word.

Related searches - Sometimes users may not really know what exactly they are looking for or what keywords to use to get the best search results. It would be great if we can suggest what to look for based on their initial keywords. This would help users narrow down and nail their search very quickly. For example, if you search for Activate 2018 on Google, you are looking for a very specific conference in Montreal. However, the first few items in the search results list brings up a different conference with the same name. Google helps me disambiguate by suggesting other queries that you should use. The problem is solved with one of the related searches.

Related articles/products - When a searcher is interested in a search item for example, an article about AI in the healthcare domain, you may want to explore other related articles to increase our knowledge on that topic. Suggestions within the search results (e.g. right under the description of the search result), can help users explore and find more of what they really want. This not just helps the users, it also helps the enterprise with user engagement. Keeping users longer on the website means users are getting more value from what is on offer and leads to increased conversions.

Identify Your Search Problem

Patrick Reinhart, senior director of digital strategies at San Francisco-based Conductor pointed out that one of the biggest problems search engines are struggling with these days is figuring out what problem they are trying to solve for their users. Most companies look to Google for that answer, which makes sense. The problem, however, is that these companies also lose their identity because they are just trying to clone what Google is doing, which is impossible. “Google knows what it wants to accomplish — it always has. It wants to answer your questions along with indexing and organizing the internet,” he said. 

“Now, you may have noticed that I said it and not they, and that's because Google's Rankbrain has allowed the company to step out of the world of being told what to do and is now figuring out most of what it's doing on its own. Google utilized this machine learning AI to make its algorithm work better and faster to drive towards its mission.” That's what makes a good search engine, a clear mission that a collection of people can work towards and avoid getting distracted with details that take them away from the main goal. “This is where a lot of other search engines are missing the mark. They don't know what the problem is that they want to solve, they have no identity,” Reinhart said.

Once other search engines figure this out, AI can assist in driving towards that goal and make the experience better by taking day to day tasks out of the engineers’ hands so that they can focus on the bigger picture and overall mission of their search engine.

Related Article: Searching for Brand Success With Voice Assistants

Content Classification And Group Results

There are two other main features a modern search engine must have, AI driven content classification and clustered group results, according to Dr. Manjeet Rege, of the School of Engineering at St Thomas University in Minnesota. Both features provide users with custom and accurate information.

AI content classification - Previously, search engine robots were programmed to only look for keywords in a static catalog of data. Now, our search engines can take search terms and customize them based on user behavior. AI now allows search engines to notice if one website meets the searcher’s needs more than another and will bump up that URL to the top search result. With the help of AI, content classification can now help find not only content that matches your search term, but multimedia as well, including images and videos.

Twenty years ago, search engines were mainly text based. A website with an image would be retrieved only if there was matching text around it. Now, with the help of AI — contents within images and videos could be discovered. So even if a web resource without any text or captions exist, that will still be retrieved because the content inside the images and videos. 

AI is now learning based on how users are reacting. If the top search result doesn’t meet your needs and you quickly leave the page, AI is tracking that movement. If the fourth search result meets your needs and you stay on that page for a while, AI tracks that data. Eventually, AI’s dynamic ranking will recognize the fourth result as the more useful search result and will slowly make it the first one that pops up

Clustered group results - Clustered group results refer to the presentation of information the search engine finds for you. If we were to search the word "leaf" 10 years ago the search engine robot would have provided URLs and pictures of leaves on trees or bushes. Now, when we search the word, ‘leaf,’ we are given a choice between either leaves on a tree or the compact car known as the Chevy Leaf.

AI has helped the search engines recognize words can have more than one meaning, and to best serve users, you provide them with all the options.“As AI keeps moving us forward we can expect to see an increase in voice-based search engines. We can already see this being explored with Alexa and Googl," said Rege.

According to a study conducted by Dimensional Research in tandem with M-Files, nearly 50 percent of workers said they have struggled with documents and content scattered in disparate locations across their organization. In addition, the Dimensional Research survey found that 40 percent of workers had to search three or more locations to find a file or document and that nearly half of workers were unsure if they had the most recent version. 

The copious amount of data generated daily is a nightmare for employees, and that’s where AI comes in. New developments in AI set the stage for an intelligent digital assistant that is designed to help employees do data-oriented tasks, such as prioritizing emails based on urgency, balancing workflows and proactively attaching relevant content to emails and ultimately make their lives easier in this age of information overload.