How Talla's Smart Knowledge Base Uses AI

April, 17 2018

Subscribe and stay up to date

No spam, we promise! You will only 
receive essential emails.



One of the most common reactions when we demo Talla's solutions to prospects and customers is, "this is great, but where is the artificial intelligence?" The short answer is "everywhere," but we understand the confusion, so let's examine how Talla's knowledge base employs A.I. software.

First, in the minds of many consumers, artificial intelligence is any technology that seems "magical," which is to say that it performs functions that seem beyond the scope of traditional software. Everyday software, however, uses A.I. all the time. Magic has become commonplace. Thus, our customers can be staring A.I. in the face and not realize, and so can you. 

For example, digital smartphone and smart-speaker assistants like Apple's Siri and Amazon's Alexa use A.I. to understand spoken commands and turn them into search queries. Google uses A.I. to "correct" search queries with typos and to suggest better results. Nest thermostats use A.I. to guess when and how you'd like your home's climate control settings adjusted.

The common thread in A.I. software is that it learns and adapts over time to improve its performance, and it does that without being reprogrammed by a human. Moreover, A.I. systems can infer what you mean by looking at data that isn't included in your request, because they understand context. Talla's knowledge base software does this today -- even if it isn't obvious where the magic lies.

Talla's internal search engine is "aware" of the structure of your knowledge base content, so it can distinguish between titles, section headers, and body content. If you have a master policy document (as denoted by a title), with a section on parking policies (as denoted by a section header), and a paragraph that addresses the Boston office's parking policy (as denoted by body content), Talla's search engine would jump to the correct paragraph of the document when you search a phrase like "Boston parking policy".

Moreover, Talla's search engine would "learn" over time if the overall document, and the specific paragraph it suggested, were the best result for that query, and adjust its results accordingly.

In a more advanced scenario, you could simply search "parking policy" and, given your user profile in Talla, the search engine would already know you work in the Boston office and give you the Boston policy excerpt without you even including the word "Boston" in your search query. The Talla A.I. would learn whether this technique works for you, and works in general, and over time would adapt its search behavior to be more useful.

Within Talla, all this behavior is tied to a Siri-style chatbot that can understand chat-based questions and requests and convert them to more typical search queries -- all which will improve over time.

These features deal with the fundamental problem of information discovery, which is to say making it easier for knowledge base users to find the right information inside your knowledge base. That's how the Talla knowledge base uses A.I. today.

 In the future, Talla's knowledge base will use A.I. to tackle problems of information creation and information curation, determining what is missing or inaccurate in your knowledge base, and helping you develop and organize your knowledge base content in a more effective fashion. (But that's a blog post for another day.)

If you'd like to see Talla's A.I. solutions in action, or discuss what the future of A.I. can be for your company, contact Talla today.

New call-to-action 


View all posts

Subscribe and stay up to date

No spam, we promise! You will only 
receive essential emails.

Subscribe and stay up to date