Shiny Blog

Google-like Search in the Enterprise

[fa icon="calendar"] Nov 30, 2018 11:05:25 AM / by David Yoon

searchImage

Why settle for your enterprise search looking like this...
shinysearch01
when it could look like this?

updateSearchimage

Actually that's unfair. Most organizations would be ecstatic with version 1. A single, relevant result as opposed to a typical search that throws up a page of mostly misses while the desired document isn't even indexed, making it impossible to find via search.

Version 2 just seems like science fiction at this point. The holy grail of enterprise search. How is enterprise search supposed to navigate permissions, identify patterns, extract meaning, recognize intent, and provide structure around your search query?

Believe or not, we're there already. Let's break it down.

1. Connect your enterprise data sources. Unstructured information is essentially invisible to the organization. Workers can't find information that the enterprise doesn't know it has. Email, shared folders, legacy repositories, and user desktops hold just as much relevant data as your ECM and it should be included with every search.

2. Build document metadata. Let's face it, information governance in the enterprise is an afterthought at best. On the web, users spend countless hours optimizing their pages to increase searchability. Every page is filled with loads of metadata, relevant links, keywords and more. In the enterprise, users are far less thorough with their tagging. And even when they do, they've proven remarkably bad at it. IDC sees that less than 50% of enterprise content is correctly indexed, meta tagged or efficiently searchable.

Get machines to go the last mile. Artificial intelligence can process vast amounts of data quickly and accurately. Extract relevant entities like project names, product identifiers, and submission dates with the help of natural language processors. Categorize data by department, document type or customer deal. AI enriched documents are simply easier to search. 

3. Show your work. Machine learning is both incredibly smart and incredibly dumb. All your information is properly tagged and identified now what? Search can now pull similar documents concerning a particular off-shore rig, identify other PII markers associated with a searched for name, bring up support tickets, sales presentations, and documentation around product queries. AI makes it easier to cluster "like" documents illuminating connections. Recognizing intent and providing structure around content provides answers to the business to make better decisions.

Engage with all your information. Automation improves your employees ability to react to change, and speeds customer response times. Machines make it possible with a lot more science and far less fiction.

Topics: AI

David Yoon

Written by David Yoon

Director of Marketing at Shinydocs Corporation. Delivering information management that embraces your current work culture and optimizes your technology investments.

Subscribe to Email Updates

Recent Posts