AWS machine learning improves search
May 13, 2020
Amazon Web Services (AWS) has announced the general availability of Kendra, an enterprise search service powered by machine learning.
Kendra uses machine learning so organisations can index all their internal data sources, make those data searchable, and allow users to get precise answers to natural language queries. When users ask a question, Kendra uses finely tuned machine-learning algorithms to understand the context and return the most relevant results, whether that be a precise answer or an entire document.
For example, businesses can use Kendra to search internal documents spread across portals and wikis, research organisations can create a searchable archive of experiments and notes, and contact centres can find the right answer to customer questions across the complete library of support documentation.
Kendra requires no machine-learning expertise and can be set up completely within the AWS management console.
Despite many attempts over many years, searching for information within an organisation remains a vexing problem for enterprises. Many businesses and organisations struggle implementing internal search across their siloed troves of data, requiring their users to use keywords to find information.
Organisations have vast amounts of unstructured text data, much of them incredibly useful if they can be discovered, stored in many formats, and spread across different data sources such as SharePoint, Intranet, Amazon Simple Storage Service and on-premises file storage systems.
Even with common web-based search tools, organisations still find internal search difficult because none of the available tools do a good job indexing across existing data silos, don’t provide natural language queries, and can’t deliver accurate results. When users have questions, they are required to use keywords that may appear in multiple documents in different contexts, and these searches typically generate long lists of random links that users have to sift through to find the information they seek, if they find it at all.
Kendra lets users search across multiple silos of data using real questions not just keywords and leverages machine-learning models to understand the content of documents and the relationships between them to deliver the precise answers they seek instead of a random list of links.
Because natural language understanding is at the core of Kendra’s search engine, employees can run their searches using natural language; keywords still work, but most users prefer natural language searches.
As an example, an employee can ask a specific question such as “When does the IT help desk open?” and Kendra will give them a specific answer such as “9:30am”, and highlight the passage in the source document where it found the answer, along with links back to the IT ticketing portal and other relevant sites.
It supports industry-specific language from IT, healthcare and insurance, plus energy, industrial, financial services, legal, media and entertainment, travel and hospitality, human resources, news, telecommunications, mining, food and beverage, and automotive, with additional industry support coming in the second half of this year.
“Our customers often tell us that search in their organisations is difficult to implement, slows down productivity, and frequently doesn’t work because their data are scattered across many silos in many formats,” said Swami Sivasubramanian, AWS vice president. “Using keywords is also counterintuitive, and the results returned often require scanning through many irrelevant links and documents to find useful information. Today, we’re excited to make Amazon Kendra available to our customers and enable them to empower their employees with highly accurate, machine-learning-powered enterprise search, which makes it easier for them to find the answers they seek across the full wealth of an organisation’s data.”
Kendra encrypts data in transit and at rest and integrates with commonly used data repository types such as file systems, applications, intranet and relational databases, so developers can index their company’s content with a few clicks, and provide users with accurate search without writing a single line of code. It provides native cloud and on-premises connectors to popular data sources such as SharePoint, OneDrive, Salesforce, ServiceNow, Amazon Simple Storage Service and relational databases, with more being added throughout this year.
Developers can add data sources by selecting the connector type, and those connectors will maintain document access rights. Data connectors can be scheduled to sync automatically between the index and data sources to ensure users are always securely searching the most up to date content.
Kendra also helps ensure search results adhere to existing document access policies by scanning permissions on documents, so search results only contain documents for which the user has permission to access. Developers simply log into the Kendra console, point the service at their unstructured and semi-structured documents, and Kendra then creates an index across silos of data. Users can then deploy Kendra across their applications from the console by copying short code samples provided in the documentation.
Onix, a cloud consulting company with nearly 20 years of deep enterprise search experience, has helped hundreds of customers adapt to the ever-changing search landscape.
“Search capabilities have evolved over the years,” said Tim Needles, CEO at Onix. “Users now expect the same experience they get from the semantic and natural language search engines and conversational interfaces they use in their personal lives. Powered by machine learning and natural language understanding, Kendra improves employee productivity by up to 25%. With more accurate enterprise search, Kendra opens new opportunities for keyword-based on-premises and SaaS search users to migrate to the cloud and avoid contract lock-ins.”