AI-Driven Customer Experiences are the Future of Knowledge Management
Knowledge management, at its essence, collects, catalogs, and curates data in order to make it usable to an organization.
Today’s knowledge workers expect more. They expect both raw data and analytics—scratch that—answers at their fingertips. They expect automation to eliminate repetitive tasks and streamline workflows. Some key technologies and techniques are required to implement these advanced capabilities including effective search, scalable storage, and powerful AI.
Search is Everything
The volume of data modern organizations have to access is just too large to curate every byte into a well-thought-out, hand-crafted ontology that a user can easily navigate. Instead, users expect the ability to simply ask questions and get answers. Some answers are a number or phrase (i.e., “What was last month’s sales forecast for QLS?”), where some answers are collections of documents or charts or analytics.
Context is King
Both users and data exist in a particular context. If I’m in marketing, a “qualified lead” may mean something different than if I’m on the sales side of things. Marketing has its “marketing qualified leads” and sales has its “sales qualified leads.” A well-designed system should understand this distinction and return the answer that is right for me.
Context is truly user experience (UX). It depends on the back end, including security (plus knowing authenticated users), organizational metadata, and the data management system. It also depends heavily on the interface itself.
Key to this is data unification. If a sales rep has to look up an account’s information in several places, that’s a lot of search and a lot of inefficiency. The rep should find everything needed in the most obvious place. The system should execute the search taking into account the context of the customer currently being served.
User Interfaces are Conversational
Modern user interfaces must provide a full-featured user experience beyond the search box and list of links.
Modern user interfaces are bi-directional: they must provide and collect behavioral data from the user. This data is used to personalize the experience. Answers shouldn’t just be provided based on my position in the company but based on the kinds of things that I tend to look for. If I tend to click through to charts and analytics, then the system should give me more charts and analytics. If I tend to just want an answer (“Just give me the number, I don’t need a story!”), then give me just an answer.
Collection and Curation Must Be Automated
The old tried-and-true KM tools and methods are not enough anymore. By the time you achieve the ideal organization of all your data, it will not only change, but you’ll have drowned in the data. Using AI techniques like machine learning and deep learning, systems can be trained to automatically categorize and curate data. Instead of humans manually creating and tagging information into ontologies, humans can teach a computer how to do it for them.
Machine learning (ML) and deep learning is not a luxury only for organizations with the resources to do more advanced capabilities. It is an absolute necessity. Modern computing has taken us from the 1990s dearth of data to a deluge of data pushing us to the point of drowning. The problem is almost never that we don’t have data, but that we cannot sort and curate the data into answers fast enough given the volume. Where modern computing created this problem, modern computing has the answer.