Description
IBM Watson is IBM's official AI brand page, tracing Watson's history and its role today as the foundation for IBM's watsonx AI portfolio.
Read More
IBM Research began work on Watson in 2007, as a challenge to build a system able to compete with Jeopardy! champions. By 2011, Watson beat the two highest-ranked human players, Ken Jennings and Brad Rutter, on the show, ranking its confidence in candidate answers to respond in under three seconds. In 2013, Watson technology became available as a cloud development platform, spurring a wave of AI application startups. In 2014, IBM Watson Discovery Advisor was designed to surface unexpected connections in raw data. In 2017, IBM unified its natural language processing work into IBM Watson NLP, so IBM products and AI applications could draw on the same underlying stack. In 2020, Watson Assistant released a beta intent-detection model combining traditional machine learning, transfer learning and deep learning, aimed at faster and more accurate intent recognition with less training data. In 2023, IBM announced watsonx, developed over three years, to manage the lifecycle of the foundation models behind generative AI and to build and tune machine learning models.
IBM now presents watsonx as the current AI portfolio built on Watson's foundations. It includes watsonx.ai, for training, validating, tuning and deploying foundation and machine learning models; watsonx.data, for scaling AI workloads across enterprise data; watsonx.governance, for responsible, transparent and explainable generative AI workflows, including for models built on third-party platforms; watsonx Assistant, for building and deploying AI-powered virtual agents without writing code; watsonx Orchestrate, for offloading time-consuming employee work; and watsonx Code Assistant, which gives developers AI-generated code recommendations.
This page is aimed at enterprises evaluating IBM's AI portfolio and at anyone researching Watson's history in AI. IBM also offers IBM Garage consulting to help shape an AI deployment strategy.
The Watson brand itself is presented here mainly as history: current product development and capabilities sit under the watsonx name rather than under a standalone Watson product.
- CategoryNatural Language Processing Tools
- Pricing ModelContact sales
- DeploymentCloud
- PlatformsWeb
IBM Watson's Core Features
Documents Watson's path from the 2011 Jeopardy! win to today's watsonx portfolio
watsonx.ai: train, validate, tune and deploy foundation and machine learning models
watsonx.data: scale AI workloads across enterprise data sources
watsonx.governance: transparent, explainable AI workflow management, including for third-party models
watsonx Assistant: no-code building of AI-powered virtual agents
watsonx Orchestrate: automates time-consuming employee workflows
watsonx Code Assistant: AI-generated code recommendations for developers
How to use IBM Watson?
Review Watson's history: Read the page's timeline from the 2011 Jeopardy! win to the 2023 watsonx launch.
Pick a watsonx product: Identify which portfolio product fits your need, such as watsonx.ai for models or watsonx Assistant for virtual agents.
Explore watsonx.ai: Follow the link to train, tune or deploy a foundation model.
Set up governance: Use watsonx.governance to track and explain AI workflows, including third-party models.
Talk to IBM: Use the IBM Garage links to get consulting help planning an AI deployment.
IBM Watson's Use Cases
- Enterprise AI history reference
- Foundation model management
- AI governance
- No-code virtual agents
