Amazon today launched General Availability Kendra, an AI and machine learning service for enterprise search. It was previewed last December during Amazon Web Services (AWS) on: Invent 2019 in Las Vegas and is now available to all AWS customers.
Companies generally struggle with countless data areas. Over 93% say they store data in more than one place. Some of these buckets are inevitably no longer used or forgotten. A Forrester survey found that between 60% and 73% of all data within companies is never analyzed for insights or major trends. This is where services like Kendra come in – they use AI to return results that are more relevant to users or embedded in apps that return search queries.
After configuration via the AWS console, Kendra uses connectors to standardize and index different sources of information beforehand (from file systems, websites, SharePoint, OneDrive, Salesforce, ServiceNow, Amazon Simple Storage Service, relational databases, etc.). Customers answer some questions about their data (and optionally ask frequently asked questions (think knowledge bases and support documentation) and let Kendra use natural language processing to create an index to identify concepts and their relationships.
According to Amazon, Kendra̵
Kendra ensures that the search results comply with existing access policies by scanning permissions for documents so that the results only contain documents that the user has access to, and data is encrypted during transmission and at rest. Here are some of the other questions it can understand:
- “How do I set up my VPN?”
- “How long does it take for policy changes to take effect?”
- “When does the IT help desk open?”
- “What is the genetic marker for ALS?”
- “What are some of the suggested treatments for COVID-19?”
Queries in Kendra can be tested and refined before being deployed. They improve themselves over time as the underlying AI algorithms pick up new data. Companies can manually adjust relevance and improve certain fields in an index, e.g. B. the topicality of documents, the number of views or certain data sources. The ready-made web app for end users can be integrated into existing internal apps. It has signal tracking mechanisms that monitor which links users click and what searches they perform to improve the underlying models.
Last year, the company showed renewed interest in AI-based software-as-a-service (SaaS) products that capture, understand, organize, and query digital content from various sources. Beyond Kendra, Microsoft has ramped up the segment by launching Project Cortex, a service that taps AI to automatically classify and analyze documents, conversations, meetings, and videos from a company. In a way, this was a direct response to Google Cloud Search, which retrieves data from a range of third-party products and services that run both locally and in the cloud. Machine learning was used to provide query suggestions and determine the most relevant results.
In any case, the cognitive search market is skyrocketing – it’s expected to be worth $ 15.28 billion by 2023, after $ 2.59 billion in 2018, according to Markets and Markets.