Ben  Franklin Ben Franklin | 15 Mar 2021

In Part 1 of this series, we talked through Azure Cognitive Search’s ability to save time and effort on content categorisation. In Part 2, we’ll be discussing how Cognitive Search can be implemented as an intrinsic component of a web platform’s navigation. And not only because it goes beyond the standard search bar. By illuminating relationships between content that may have previously been hidden, Cognitive Search creates a novel path on the user journey. 

Change website workflow

Microsoft’s mantra for Azure Cognitive Search is ‘Ingest, Enrich, Explore’. Ingest and Enrich take care of extracting data from content as it is uploaded by implementing a suite of cognitive skills and making that data accessible to searchers based on the parameters put in place by the business.

For an organisation processing a lot of data, such as a frequently changing product catalogue or a library of thousands of internal records that inform ongoing projects (for example, Microsoft case study Howden), this is a potentially huge resource saver. It saves time spent tidying and re-tidying data so that it can be retrieved, and it saves time retrieving it. This is true even when it is used at a fairly basic level. 

When we get into ‘Explore’, however, the power of Cognitive Search to provide actual insight into the data becomes apparent. It isn’t just providing a static list. It’s showing interactions and relationships. This affects the end user experience (UX) in a rather profound way.

Obviously the search function on any web platform affects UX. Cognitive Search, however, is truly integral to UX in a way that an ordinary search function often is not. Search ceases to be a simple magnifying glass icon or a bland blank bar in a page corner or down at the footer. It is not an afterthought to the navigation. Cognitive Search effectively is the navigation in some respects. 

Tailor search capabilities

While the cognitive skills do the bulk of the work trawling and enriching the data behind the scenes in order to generate search results, the end user has to do their bit via the interface in order to get what they want.

By configuring the index, you can determine the extent of the user’s interaction with the data. In particular, choosing whether content is filterable, sortable and/or facetable is going to directly impact that navigation. This configuration upends the conventional user journey, in a way. Instead of finding a path through the website to the content, the user inputs directions, and the content comes to them.

Creating faceted navigation category hierarchy, for example, provides end users with a menu of options that whittles down the results based on conditions provided by the user rather than by the cognitive skills. Think of searching an online clothing retailer with a sidebar full of product specifics.

You can get more precise results according to what you’re after by choosing which of these facets are important to you: perhaps colour, size range, and style. As you choose, the search becomes more refined, hopefully giving you something closer to exactly what you want. 

Reveal content connections

Any website search result can be narrowed in this way. But the application of cognitive skills and the machine learning ability means that for Azure Cognitive Search businesses, the ‘Explore’ moment can be truly tailored to the end user. Indeed, how the search results appear needn’t stay as the usual scrollable list, as a clever demonstration from Microsoft Developer illustrates.

This demonstration visualises the links between data points in the JFK assassination, compiled from thousands of digitised documents. Cognitive Search allows for a rapid realisation of the relationships among all types of content. How these relationships are displayed—whether that’s a simple list or an expanding spider diagram as in the JFK example or something else entirely—drives the user farther along their journey. 

Go back to the Eiffel Tower photograph from Part 1. Recall that it showed a smiling person holding a branded product in front of the Paris tourist destination. Originally you were searching for the Eiffel Tower, ‘Iffel Twoer’ or French cities or something similar, and your search results brought you that photograph. A dynamic search result might also have linked in results showing the product that was featured in the photograph, linking the relationship between the two. 

Perhaps this logo is an iconic Parisian brand, and the association of French urbanity and this particular item is important to its sales. Seeing where this logo shows up in social media in terms of mood, age and gender of the people in the photos, even what time of day when it appears might provide a lot of information about target demographics. Or perhaps like the JFK files, it is a piece of evidence in a crime, and this photo is linked to hundreds of other fragments. One thing keeps cropping up, though: that logo.

What your business does with Azure Cognitive Search is your business—literally. It is highly customisable. It is flexible and powerful. And it is changing how we understand, interpret and use data to solve business problems from the practical to the abstract.

Is Azure Cognitive Search right for your business? Get in touch with our Technical Team to talk more about it.