IBM Advances ‘Watson Anywhere’ with New Clients and Innovations Designed to Make It Even Easier to Scale AI Across Any Cloud
KPMG and Air France-Klm Highlight the Value of Being Able to Bring Watson to Their Data, Wherever It Resides
Recognizing that organizations are slow to adopt AI, due in part to rising data complexities, IBM announced new innovations that further advance its Watson Anywhere approach to scaling AI across any cloud, and a host of clients who are leveraging the strategy to bring AI to their data, wherever it resides.
“We collaborate with clients every day and around the world on their data and AI challenges, and this year we tackled one of the big drawbacks to scaling AI throughout the enterprise – vendor lock-in,” said Rob Thomas, General Manager, IBM Data and AI. “When we introduced the ability to run Watson on any cloud, we opened up AI for clients in ways never imagined. Today, we pushed that even further adding even more capabilities to our Watson products running on Cloud Pak for Data.”
Increasing data complexity, as well as data preparation, skills shortages, and a lack of data culture are combining to slow AI adoption at a time when interest in AI continues to climb. Between 2018 and 2019, organizations that have deployed artificial intelligence (AI) grew from 4% to 14%, according to Gartner’s 2019 CIO Agenda survey.1 Those figures contrast with the rising awareness of the value of AI. According to the 2019 MIT Sloan Management Review and Boston Consulting Group study, Winning with AI, 9 out of 10 respondents agree that AI represents a business opportunity for their company. Adding to growing enterprise complexities, a 2018 IBM Institute for Business Value study said that 76% of organizations surveyed reported that they are already using at least two to 15 hybrid clouds, and 98 percent forecast they will be using multiple hybrid clouds within three years.
The innovations announced today are designed to help organizations overcome the barriers to AI. From detecting ‘drift’ in AI models to recognizing nuances in the human voice, these new capabilities can be run across any cloud via IBM’s Cloud Pak for Data platform to begin easily connecting their vast data stores to AI.
As evidence of the growing appeal of this approach to enable AI to run on any cloud, IBM today announced a number of clients that are leveraging Watson across their enterprises to unearth hidden insights, automate mundane tasks and help improve overall business performance. Companies like multinational professional services firm KPMG, and Air France-KLM, are leveraging Watson apps, or building their own AI with Watson tools, to facilitate their AI journeys.
“IBM’s strategy for developing AI tools that enable clients to run AI wherever their data is, is exactly the reason we turned to OpenScale – we needed multicloud scalability in order to give clients the kind of transparency and explainability we were talking about,” said Steve Hill, Global and US Head of Innovation at KPMG. “Supporting the client’s environment, whatever it may be, reflects the understanding that IBM has not only about AI, but about the expanding enterprise.”
KPMG, the multinational professional services network and long-time IBM alliance partner, for years has integrated the latest cognitive and automation capabilities into services across its businesses from governance, risk, and compliance to taxes and accounting. Earlier this year KPMG turned to IBM to collaborate on a new service that would provide KPMG clients greater governance and explainability of their AI, no matter where that data resides, no matter what cloud and no matter what AI platform the company was using. KPMG developed the KPMG AI in Control solution leveraging the Watson OpenScale platform, to give clients the ability to continuously evaluate their machine learning and AI algorithms, models and data for greater confidence in the outcomes. Last month, KPMG teamed with IBM to release a joint offering of this solution to clients called KPMG AI in Control with Watson OpenScale.
And to help accelerate customer service and improve the passenger experience, Air France-KLM and its customer service team developed a voice assistant called MIA (My Interactive Assistant), which uses IBM Watson Assistant with Voice Interaction. Air France-KLM collaborated with IBM to develop the voice assistant to improve the customer experience by reducing file processing time. Since the beginning of the pilot in July in a single country, MIA has responded to 4,500 calls from people needing additional information about their flights or their travel plans.
MIA asks the customer for his reservation reference number (PNR) and extracts from this PNR all information relating to the passenger, including name, flight number and telephone number. If needed, the voice assistant can quickly pass the call to a human agent who can take over. The agent will already have all the necessary information on the screen in the background and will therefore be positioned to resolve the request. By design, the higher the number of requests handled by MIA, the more intelligent it will be over time. Other use cases are also under study.
New Capabilities Come to Watson Apps and Tools
IBM rolled out an array of new features and functionality to several of its key products today, including the following:
Watson OpenScale – With rising data privacy regulation and growing concern for how AI algorithms are reaching their results, bias detections and explainability are becoming critical. Last year, IBM launched OpenScale, the first AI platform of its kind to do just that – provide organizations the ability to look for bias and govern their AI and query it to understand how it arrived at its results. With such insights, clients can gain greater confidence in their AI and in their results. Today we announced a new capability called Drift Detection which detects when and how far a model “drifts” from its original parameters. It does this by comparing production and training data and the resulting predictions it creates. When a user-defined drift threshold is exceeded, an alert is generated. Drift Detection not only provides greater information about the accuracy of models, but it simplifies, and hence, speeds model re-training.
Watson Assistant – Building on IBM’s leading position in enterprise AI assistants, IBM today announced several new key features to the conversational AI product that allow users to deploy, train and continuously improve their virtual assistants quickly on the cloud of their choice. For example, the new Watson Assistant for Voice Interaction is designed to help clients easily integrate an AI-powered assistant into their IVR systems. With this capability customers are able to ask questions in natural language. Watson Assistant now can recognize the nuances of the way people speak and will fast-track the caller to the most appropriate answer. Clients can also blend texting and voice at the same time, allowing instantaneous information exchange. IBM also announced that Watson Assistant is now integrated with IBM Cloud Pak for Data which enables companies to run an assistant in any environment – on premises, or on any private, public, hybrid, or multi-cloud.
Watson Discovery – IBM announced several key updates to Watson Discovery, the company’s premier AI search product that leverages machine learning and natural language processing to help clients find data from across their enterprises. New to the platform is Content Miner, which allows for the searching of vast datasets for specific content types, such as text and images. A new simplified setup format helps non-technical users to can get up and running quickly, while a new “guided experience” dynamically recommends next steps in configuring projects. All of which results in a more agile data discovery process.
Cloud Pak for Data – IBM advances its first-of-a-kind, integrated data analytics platform, with key new features and support. The platform, which has supported Red Hat OpenShift, one of the leading Kubernetes-based container orchestration platforms, since its launch 18 months ago, is now certified on OpenShift. With full certification brings added confidence to clients knowing that all the components came from a supported source, container images contain no known vulnerabilities and most importantly that the containers running throughout are compatible across Red Hat Enterprise Linux environments, regardless of the cloud, and whether private, public or hybrid.
In addition to certification, this latest version now comes with a host of capabilities standard as part of the base platform. Among the new capabilities is Db2 Event Store, for storing and analyzing more than 250 billion events per day, in real-time, and Watson Machine Learning, equipped with AutoAI. AuoAI is IBM’s innovative automated model building program that enables data scientists and non-data scientists alike to build machine learning (ML) models with ease. As its name suggests, AutoAI automates such tedious and complicated tasks of ML, including data prep, model selection, feature engineering and hyperparameter optimization, to truly speed clients’ adoption of AI. Now these tools come standard with Cloud Pak for Data to be used and scaled across any hybrid multicloud environment.
In addition, Cloud Pak for Data now features open source governance in the base platform, enabling users for the first time to set policy for, and govern the use of open source tools and programs within the enterprise to enable more efficient model building, testing and deployment.
To empower developers to take advantage of the IBM Cloud Pak for Data platform, IBM also announced the Cloud Pak for Data Developer Hub. Here, developers have step-by-step tutorials, code patterns, ongoing support and information on in person workshops taking place in their area for hands-on labs.
Read More: AI And The Future Of Content Management