Anthropic has introduced a new feature to its AI chatbot, Claude, enabling users to create personalized automated assistants. Known as “Tool Use” or “Function Calling,” this feature integrates with external APIs to perform various AI tasks, such as purchasing items and analyzing data.
The new tool allows for the creation of email assistants, shopping bots, and other customized solutions. For instance, it can generate product recommendations based on a user's purchase history or provide quick responses to customer inquiries, including order tracking and technical support. Users need access to an API and basic coding skills to set up an assistant.
Additionally, the tool can process images, making it useful for applications like virtual interior design consultants that analyze room images to offer decor suggestions. This feature is accessible through Anthropic's Messages API, Amazon Bedrock, and Google Vertex AI.
Cost and Availability
The pricing for the AI assistant is based on the volume of text processed, measured in “tokens.” Typically, 1,000 tokens are equivalent to about 750 words. During the beta phase, most users preferred Anthropic's Haiku option, which costs approximately $0.25 per million input tokens and $1.25 per million output tokens.
Dianne Penn, a project lead at Anthropic, mentioned that the Claude tool has been in beta testing since April with several thousand customers. One notable startup, Study Fetch, used the tool to develop a personalized AI tutor named Spark.E.
How this Sits in the Industry
The development of AI assistants or agents is a growing trend in the technology sector. Google, for instance, has introduced various AI-driven shopping and search functionalities, while OpenAI is working on a voice assistant capable of real-time interaction and environmental observation. Anthropic's new feature empowers users to create tailored assistants, offering an alternative to similar solutions from Google and OpenAI.
Several clients have already been utilizing Tool Use. StudyFetch, an online learning assistant, uses it to help students navigate course materials, while Hebbia, a startup, leverages the technology to extract metadata from documents and automate workflows for financial services clients.
