By Agam Shah
Amazon.com Inc. on Tuesday introduced two artificial-intelligence services designed to help companies gain better insight from business data spread across multiple sources, from documents to conversations with customers.
Both services, Contact Lens for Amazon Connect and Amazon Kendra, can be previewed immediately on the company's cloud service, Amazon Web Services.
The tools, announced at Amazon's annual cloud event in Las Vegas, aim to help the company's cloud customers incorporate functionality such as natural language processing, but without long waits often associated with AI-related projects. Amazon says it is making this possible by integrating machine learning into the two new stand-alone services.
"There's no machine-learning expertise required for either of these services. They're just plug and play. You don't have to get into all the weeds and get the training data and label the data and all those sorts of things," said Matt Wood, vice president for artificial intelligence services at Amazon Web Services.
The new AI tools process a client company's data and attach meaning to that information through methods that identify patterns, Amazon said.
Implementing AI isn't easy for companies. International Data Corp. in a July study said that most of the 2,473 global organizations surveyed had some failures in AI projects, and a quarter had up to a 50% failure rate.
Amazon Kendra, an AI-enabled enterprise search tool, answers queries by combing through a variety of data sources within an organization. The search tool can be implemented on websites and interfaces such as chatbots.
Kendra uses deep learning models to understand text from multiple sources and across a number of domains including life sciences and legal and financial services. Mr. Wood said.
"As people use this service more, the machine-learning models under the hood get better at ranking and providing relevant answers to the questions," Mr. Wood said.
For example, an employee at a client company might ask, "When will an IT help desk open?" and Kendra will respond with an opening time and a map to the help desk.
Workgrid Software LLC, a unit of Liberty Mutual Insurance Co. that specializes in workplace chatbots, has been testing Kendra to see how technology could help its virtual assistants give better answers to questions posed by employees of its client companies. Kendra does the heavy lifting of providing the needed AI tools and connecting data sources, said Gillian McCann, head of cloud engineering and artificial intelligence at Workgrid Software.
"That's the piece I'm really trying to explore with Kendra...to give me this further, more in-depth search. Rather than us building it, we could use an AWS service," Ms. McCann said.
Contact Lens was built for Amazon's cloud-based call-center service, Amazon Connect, which lets clients build a virtual contact center. The new tool uses machine learning to assess customer conversations. Amazon Connect customers can sign up for preview access of the new tool.
Contact Lens uses machine learning to assess the sentiment of a call, which could be positive or negative.
"We also detect long periods of silence and also clashes -- when the agent and customer are talking over each other," Mr. Wood said.
Using Contact Lens, companies can search transcripts of customers' conversations with agents by keywords, sentiment and other topics. Through dashboards and reports, they will be able to measure trends such as call quality over a certain period. Based on that, companies can implement changes, such as revising scripts or training employees.
"You're getting better visibility; it's bringing into focus the interactions that your customers are having with your organization, day by day, minute by minute, and allowing you to inspect that on a level that just wasn't possible before," Mr. Wood said.
Research and advisory company Gartner Inc. projects that 40% of the contact center interactions will be fully automated by 2023, up from around 25% in 2018.
Write to Agam Shah at email@example.com