To get the most out of Amazon Kendra, Ben Freiberg is linking AWS’ new search service with Alexa for Business – and optimizes the possibilities of corporate search engines through a language-based interface. In our three-part article he explains both components and the result of their combination.
Amazon Web Services (AWS) recently unveiled Amazon Kendra at its in-house exposition re:invent. Kendra is a machine learning search service that allows companies to develop their own powerful search engines for distributed knowledge and internal documents. The search is based on natural language instead of just using keywords.
What a catch phrase and impulse! If you think about the possibilities of Amazon’s language assistant Alexa, it seems obvious to try to improve the voice control of the search service a little more in order to make it more comfortable for the user – and to integrate Alexa for Business as an interface for a google-like company-wide search in Amazon Kendra.
This three-part article first shows the advantages and possibilities of each of the two systems and then explains what comes out when they are properly linked to one another.
Part 1: „Alexa, let’s talk business”
The everyday working routine is characterized by a great deal of standard questions. For example: “Where is the helpdesk?”, “How many holidays do I get when I move?”, “Who is team leader of Team XYZ?”, “When is the Christmas party?” and so on.
The answers to these questions are usually hidden in various documents in Sharepoints, Wikis, DropBox, Salesforce and similar sources and tools. To make this information as accessible as possible, AWS has developed the machine learning based search service Amazon Kendra. But before it can find and give the right answers, the right questions must be asked – and these questions must be asked in the right way.
These and similar questions are well suited for a language-based interface. And with Alexa for Business, AWS also provides the most suitable one for companies.
May I introduce? Alexa…
Next to Google Home (Assistant) and the Apple Homepod (Siri), Alexa is the best known and most popular representative of the Smart Speakers. In Germany alone, 32 percent of people now use digital speech assistants for private comfort and convenience, among the age group under 40 as many as 48 percent.
Alexa for Business transfers these into the office and is intended to enable organizations and employees to work more effectively with Alexa. To this end, the service offers numerous possibilities, including:
- source of information for employees and/or visitors concerning rules or processes e.g. Onboarding
- setup of conference calls
- sending of voice memos
- creation and management of appointments
- booking of conference rooms and setting up of meetings
- control of room temperature and lighting in offices and conference rooms
The decisive factor for this is that Alexa provides interfaces to many other relevant office applications such as Microsoft Office 365, Google G Suite or Exchange. Of course, this is only possible after the necessary access authorizations have been granted. In addition, there is a management interface which, alongside the administration of devices, users and skills, enables the analysis of various key figures, such as the utilization of conference rooms.
However, the use of Alexa for Business became truly meaningful only after it was adapted to individual business requirements. First and foremost, the ability to create so-called private skills for shared devices and make them available to registered users. In other words, skills for voice-controlled functions and access to company applications that can only be used by selected groups and can in turn be assigned to specific shared devices.
This is where Amazon’s platform thinking comes into play. For example, a specially set up marketplace should make it possible to publish skills that can also be used by other companies. Examples include skills to support travel planning (SAP Concur) or skills for device management, such as notification or even automatic reordering when printers run low on toner.
To be continued in part 2: “Kendra, fill me in!”
Ben Freiberg is a Senior Data Engineer and proven AWS expert at the *um location and AWS Competence Center in Frankfurt.
This post is also available in: German