IBM revealed plans today to expand its AI and data platform watsonx with new capabilities designed to assist enterprises in deploying and governing AI solutions. The updates include new generative AI models, data tools, and model governance features that aim to help organizations overcome barriers to AI adoption.
IBM has clearly been hard at work since the launch of its watsonx platform, which serves as an umbrella for AI capabilities designed to help enterprises leverage artificial intelligence more effectively. While watsonx already offered a range of AI services, IBM now intends to include new generative AI foundation models and enhancements that go beyond rudimentary tasks. Specifically, the tech giant plans to introduce its own Granite series models, underpinned by a "Decoder" architecture, which will support natural language processing tasks, such as content generation and insight extraction.
The company is also expanding its catalog of third-party models. Meta's new Llama 2-chat 70 billion parameter model and the StarCoder language model for code generation are now available on the watsonx platform via IBM Cloud.
Beyond models, IBM will be adding new data capabilities in watsonx focused on data discovery, visualization, and preparation. This includes the addition of a tuning studio aimed at helping clients tailor foundation models to their specific needs using their own data.
Additionally, users will be able to generate synthetic data from proprietary data schemas or internal data sets, and use vector databases for data storage and retrieval. The company is also planning to infuse its generative AI capabilities into watsonx.data, to help users "discover, augment, visualize, and refine data for AI through a self-service experience powered by a conversational, natural language interface".
Finally, the company also announced a technical preview for watsonx.governance, an automated tool for tracking foundation model details and enabling human approvals at critical points in the workflow.
IBM's strategy is centered around providing a holistic AI ecosystem for enterprises. watsonx aims to be an end-to-end solution for clients, covering a range of use-cases from application modernization to customer service and HR. For example, the watsonx Code Assistants will offer tailored foundation models to generate code recommendations, watsonx Assistant is set to improve customer service with conversational AI, and watsonx Orchestrate will help HR professionals automate repetitive, high-friction tasks and back-office processes.
The company also disclosed plans to integrate watsonx capabilities across its hybrid cloud software and infrastructure offerings. For example, IBM’s IT Automation products like Instana and AIOps Insights are expected to include Intelligent Remediation features, embedding watsonx AI for more streamlined IT operations.
Developers will be able to get their hands on many of these new capabilities and models next week (September 11-14) at TechXchange, IBM's premier technical learning event in Las Vegas.