Today we are releasing the new Claude App for Slack, now in beta. Built on top of Slack’s platform, the app can summarize threads, answer questions, and more. Now any company has the chance to have a “virtual teammate” who can help make work more fun and productive.
Claude was created using Anthropic research focused on helpful, honest, and harmless AI systems, using techniques like Constitutional AI. Claude is capable of a wide variety of tasks while maintaining a high degree of reliability and predictability. You can start working with Claude in Slack today for free, during our beta experience.
Claude can take on many different roles, just like a friendly and hardworking colleague. Claude can remember and summarize entire Slack threads, answer questions about website content, iterate on tasks, and more! For functions across your organization, from marketing to product management, Claude can help you be more efficient. Here are some concrete examples of how to use it:
- Get a summary, with bullet points and prioritized action items, of lengthy Slack threads or long websites.
- Turn conversations into structured data inputs for CRM entries, engineering tickets, tables, and more.
- Share a website with Claude and ask questions about the content.
- Brainstorm ideas with a group, with each participant able to mention Claude and further refine your output.
We believe in Claude’s ability to be a force multiplier for human productivity and communication. By embedding Claude as an assistant in Slack, people can have instant access to AI-generated insights, right from where they’re already working.
Steve Wood, SVP of Product, Platform at Slack, says, “We’re excited that Claude will enable organizations to become more productive in Slack. Claude’s natural conversational skills, ability to be instructed, and capacity to retain context over large amounts of data is powerful. Anthropic’s emphasis on AI safety research has created a truly helpful assistant that will empower every team working in Slack to focus on higher level work.”
You can interact with Claude in the following ways:
- Channels and group DMs, by mentioning @Claude in your message. Claude will respond in a thread that anyone in the channel or group DM can see. Anyone can make additional requests by mentioning @Claude again.
- 1-to-1 direct messages with Claude, by simply messaging it. Claude will respond right in the DM, just like messaging a coworker!
At Anthropic, we’ve been relying on Claude internally for more than a year. Claude helps us pull critical information from websites, edit or generate code (we love using it for help with pandas and Kubernetes), critique or improve memos, draft customer communications, help us summarize or expand internal communications, and generally fuel our creativity. We are excited to make a difference for Slack customers around the world, and we welcome your feedback. Add Claude to your Slack workspace today!
In Slack, Claude only sees messages in threads where it has been explicitly @-mentioned, or when you send it a DM. Anthropic does not use this data to train models.
Claude has some limitations: It may incorrectly assess its own ability or memory, and it may hallucinate, or make up information. Claude will often make mistakes with complicated arithmetic and reasoning, and sometimes with more basic tasks. Lastly, it does not have general internet access, although it can access specific links that you share with it. We want to be open about the technology’s limitations, and are working hard to improve these.
As more and more people work with Claude, we are committed to deploying our technology to the world in a responsible, thoughtful way. If you’re interested in API access to Claude for commercial purposes, please submit an inquiry to our team.
Originally published at Anthropic
Our humans need coffee too! Your support is highly appreciated, thank you!