Slack and Claude Cowork: Revolutionizing Productivity with AI

Building a Custom Slack Inbox

Yash Tekriwal, who serves as the Head of Education at Clay, has developed an AI-powered Slack inbox. This ingenious system reduces the daily notification count from 150 to about 30, focusing only on those that are truly relevant. Instead of merely reacting to incoming messages, Yash has designed a mechanism that categorizes, prioritizes, and routes notifications to create a clear and actionable workflow.

Key takeaways from this approach include the fact that Yash receives between 100 and 150 Slack notifications each day, but only 30 to 40 of them require real action. By classifying messages into different categories such as direct messages (DMs), group mentions, discussions, and channel mentions, and then sub-categorizing them into action required, to read, and for information, he has transformed a notification overload into a manageable workflow.

Using AI to build deterministic tools, rather than just performing tasks, is crucial. There is an important distinction between asking AI to categorize items repeatedly and using it to create code that manages structured data via APIs. Yash utilized OpenClaw to create a Slack digest that extracts notifications through API endpoints.

The multi-model orchestration of Perplexity Computer frees you from the process. Unlike Claude Code or Codex, Perplexity automatically uses different models for different subtasks: Sonnet for data retrieval, Gemini for planning and coding, and Opus for complex constructions.

The cloud deployment and native connectors of Perplexity Computer are revolutionary. Unlike local coding agents, Computer operates in the cloud with pre-authenticated connectors to Gmail, Slack, Notion, Asana, etc. Applications are automatically deployed and shareable via URL.

The anti-task list framework suggests spending an hour a day automating what you never want to do again. Instead of a to-do list, maintain an anti-task list of tasks you hate.

SaaS is not dying; it is about to explode into micro-software. Yash asserts that the idea that AI will kill SaaS is misguided. Instead, we will witness an explosion of specialized software built on existing platforms.

The future of productivity is custom software. Every knowledge worker will eventually have custom applications optimized for their specific mental model and workflow.

Claude Cowork 101: Automating Your Workday Without Touching Code

JJ Englert, Head of Enablement and Community at Tenex, uses Claude Cowork to build a fully automated daily operating system. This system drafts emails, reviews his work, plans his day, and coordinates his tools.

Key points to remember from this method include the fact that Cowork is the first true bridge for non-technical knowledge workers to move beyond discussion and actually get things done with AI.

To start each project, it is advisable to create a "brain" file that explains to Claude who you are, how you work, and who you work with.

Always include good and bad examples when building skills. AI does not know what success looks like for you unless you show it.

Create an anti-task list of things you never want to do again and automate them with skills.

Cowork teaches you skills that transfer directly to Claude Code.

Projects do not need to be work-related; use them for home maintenance, wedding planning, recruiting, or anything else.

Detailed Workflows from This Episode:

• How I AI: Yash Tekriwal on Mastering Slack with a Custom AI-Powered Dashboard

• How to Turn a Text Digest into an Interactive Kanban Dashboard

• How to Prototype a Website Redesign Using an AI with Browser Access

• How to Build a Custom AI-Powered Text Digest for Slack Notifications

• How I AI: JJ Englert's Guide to a 'Daily Operating System' with Claude Cowork

• Build a Multi-Persona Sub-Advisory Board for Instant Feedback

• Train Claude Cowork to Write Emails in Your Personal Style

• How to Set Up a 'Daily Operating System' in Claude Cowork