Advanced Computing in the Age of AI | Monday, June 24, 2024

Anthropic Launches Tool Use, Making It Easier To Create Custom AI Assistants 

Anthropic has announced the general availability of the new Tool Use feature for its AI assistant Claude. The new feature enables users to build their own AI-driven solutions. The “AI agents” hook up to any external API allowing users to create an email assistant, a bot for online shopping, or other personalized solutions. 

The Tool Use feature is available across the entire Claude 3 mode family on the Anthropic Messages API, Amazon Bedrock, and Google Vertex AI. Pricing is based on the volume of text Claude processes, measured in “tokens.”

The new feature also has the capability to work with images to analyze visual data. For example, a virtual interior designer can use the tool to process images remotely and provide personalized solutions.  

The evolution of Claude from a basic chatbot to a sophisticated AI assistant capable of mimicking a human assistant is truly remarkable. Not only can the tool be used to automate tasks and provide personalized recommendations, but enterprises can also use it to improve efficiency by extracting structured data from unstructured text and answering customer questions by searching databases or using web APIs. 

Tool Use also features built-in functions to improve developer experience. This includes the ability to instruct Claude on tool selection and streaming capabilities for real-time responses in applications such as customer support chatbots. 

The introduction of the new feature comes at a time when there are a growing number of AI assistants on the market and new features are being rolled out regularly. OpenAI is transforming its chatbot into a voice assistant, while Google recently introduced various AI-driven features for searching and online shopping. 

Anthropic has been reportedly testing beta versions of the new feature for the last couple of months. The company shared success stories through a blog post of a few early adopters of Tool Use. 

StudyFetch, an AI-native learning platform, uses Claude's Tool Use feature to power its personalized AI tutor, Spark.E. According to Ryan Trattner, CTO, and Co-Founder at StudyFetch, Tool Use feature has resulted in a “42% increase in positive human feedback” by enabling the platform to act "agentically" to deliver better UI. 

Intuned, a browser automation platform, and Hebbia, an LLM platform for knowledge work, have also benefited from Tool Use with better developer experience and more streamlined customer workflows. 

The capabilities offered by Tool Use are impressive, but there could be a few key challenges in using the new feature with existing business systems. Companies will have to streamline their systems and CRMs for smooth integration of Tool Use into their ecosystem. 

Claude relies on accurate descriptions of tools to determine which tool to apply based on the user input. This means the users will have to clearly describe each tool in the AI model to enable Claude to select the appropriate tool for the task. 

While businesses will have to figure out how to leverage the power of Tool Use for their workflows, there is no doubt that the introduction of Tool Use marks a significant step forward in the evolution of human-AI interaction in a business environment. As more companies adopt this new technology, we can expect AI to take on a more autonomous role in the digital ecosystems. 

Organizations that can implement AI to augment human capabilities, not replace them, would be best placed to benefit from both, AI and humans. New AI capabilities, such as Tool Use, can handle the more complex, repetitive, and data-heavy tasks, while humans can focus on creative, strategic, and interpersonal tasks. 

Related Items 

Anthropic Breaks Open the Black Box 

Amazon Invests Another $2.75 Billion Into Anthropic 

Anthropic in Talks with Menlo Ventures to Raise $750M