The cloud provider unveiled new features for Contact Center AI, new versions of Document AI, and improvements to the AI platform for machine learning operations (MLOps) practitioners.
Google on Tuesday rolled out several new products and capabilities within its Cloud AI portfolio, including new products and features in Contact Center AI and new versions of Document AI. It also announced improvements to the AI Platform for machine learning operations (MLOps) practitioners.
Google considers its AI expertise as a key selling point for Google Cloud. “We are steadily transferring advancements from Google AI research into cloud solutions that help you create better experiences for your customers,” Andrew Moore, head of Google Cloud AI & Industry Solutions, wrote in a blog post Tuesday.
Google’s Contact Center AI (CCAI) software, which became generally available last November, enables businesses to deploy virtual agents for basic customer interactions. The service promises more intuitive customer support through natural-language recognition.
The new features introduced Tuesday include Dialogflow CX, the latest version of Dialogflow, available in beta. Dialogflow is the development suite for building conversational interfaces such as chat bots and interactive voice responses (IVR). Dialogflow CX is optimized for large contact centers that deal with complex (multi-turn) conversations. It makes it easy to deploy virtual agents in contact centers and digital channels, and it offers a new visual builder for creating and managing virtual agents. It’s available now, in beta.
Google has also updated the “agent assist” feature in CCAI, which transcribes calls, recommends workflows and provides other kinds of AI-driven assistance to human call center agents. Now, a new Agent Assist for Chat module provides agents with support over chat in addition to voice calls, identifying caller intent and providing real-time, step-by-step assistance.
Lastly, CCAI customers can now create a unique voice for their virtual agents with Custom Voice, available in beta. With Custom Voice, customers can make changes to their scripts and add new phrases without scheduling studio time with voice actors. Customers have to go through a review process to ensure their Custom Voice use cases aligns with Google’s AI principles.
While CCAI spans industry use cases, Google on Tuesday also announced new industry-specific tools — starting with Lending Document AI, a new version of Document AI tailored for the mortgage industry. Document AI extracts structured data from unstructured documents. Lending Document AI, now in alpha, specifically processes borrowers’ income and asset documents. This can speed up the loan application process.
Additionally, Google announced Procure-to-Pay Document AI, now in beta. This helps companies automate the procurement cycle, typically one of the highest volume, highest value business processes. This tool, now in beta, provides a group of AI-powered parsers that extract data from specific documents like invoices and receipts.
Lastly, Google on Tuesday unveiled new features in the AI Platform designed for machine learning operations (MLOps) practitioners.
“Even for the ML experts, the long-term success of ML projects hinges on making the jump from science project and analysis to repeatable, scalable operations,” Moore wrote in his blog post. “Often, analyst teams will hack together an activation process that can be extremely manual and error-prone with too many parameters, decoupled workflow dependencies, and security vulnerabilities. In fact, an entire discipline called MLOps has emerged to solve this issue by operationalizing machine learning workflows.”
To improve MLOps, Google is introducing AI Platform Pipelines, a fully-managed service for ML pipelines that will be available in preview by October this year. With the new service, customers can build ML pipelines using TensorFlow Extended (TFX’s) pre-built components and Templates, making it easier to deploy models.
There’s also a new Continuous Monitoring service to monitor model performance in production, which is expected to be available by the end of 2020.
To help AI teams track artifacts and experiments, the new ML Metadata Management service in AI Platform provides a curated ledger of actions and detailed model lineage. It’s expected to be available in preview by the end of September. Additionally, Google will be introducing a Feature Store in the AI Platform to provide a centralized, organization-wide repository of historical and latest feature values. It’s expected to be available by the end of this year.