# How These 3 Founders are building on Open Claw | E2248 Source: https://www.youtube.com/watch?v=vhFsdz8_L7g Recap page: https://rapidrecap.app/video/vhFsdz8_L7g Generated: 2026-02-12T02:04:30.744+00:00 --- ## Quick Overview The founders of OpenClaw demonstrated how their AI agent, operating via email and other channels like Slack and Telegram, streamlines complex tasks such as pulling PostHog data for analytics, automating research for blog posts, and handling customer support bug triaging, all while maintaining control through a dedicated agent instance and providing a superior alternative to generic LLMs by integrating directly with internal tools like GitHub and their codebase. **Key Points:** - OpenClaw, a company incubated at Launch, uses an agentic system powered by LLMs to automate internal workflows, enabling founders to skip tedious tasks like manual data extraction and initial bug analysis. - The system utilizes dedicated agent instances (like 'Ewa' and 'Claude') that can be given access to specific tools (PostHog, GitHub, etc.) and communicate via email, Slack, and Telegram. - A key use case demonstrated was generating a blog post draft from a longevity research study in under 5 minutes, highlighting the speed advantage over manual research (which takes 30+ minutes). - The company aims to make complex, multi-step tasks simple enough that even non-technical users or less-trained employees can execute them, effectively acting as a Chief of Staff. - The founders are self-funding, having raised $1.25M in seed funding and avoiding large seed rounds to maintain control and rapid iteration. - The core value proposition is integrating AI directly into existing workflows (Email, Slack) rather than relying on a separate dashboard, viewing AI as a team member, not just a chatbot. - The company has five products in market, including those focused on analytics pipelines, email as an action layer, content pipelines, and customer support/bug fixes. ![Screenshot at 00:50: The guest demonstrates the OpenClaw system by showing a Telegram conversation where the AI agent 'Ewa' successfully queries PostHog data and creates a Google Sheet summary of the 'Top 10 Active Tempo Users' based on a natural language prompt.](https://ss.rapidrecap.app/screens/vhFsdz8_L7g/00-00-50.jpg) **Context:** The interview features Preston, the host of 'This Week in Startups,' speaking with the co-founder and CEO of OpenClaw, an AI-powered automation company. OpenClaw was incubated at Launch and focuses on building agentic AI tools that integrate into existing workflows to automate complex, multi-step business tasks. The discussion centers on how their product, which uses LLMs like Gemini, streamlines operations for founders and teams, particularly in areas like analytics, content creation, and customer support. ## Detailed Analysis The discussion centers on OpenClaw's agentic approach to workflow automation, differentiating it from standard LLM prompting by integrating AI directly into existing tools like email, Slack, and Telegram. The CEO, Presh (presumably Presh Denesh Kumar), explains that their system allows users to delegate multi-step tasks to AI agents (like 'Ewa' or 'Claude' instances) that can access specific tools (like PostHog or GitHub) and report back results, often via email or Slack. This is contrasted with generic LLMs that are often sandboxed. Presh highlights a key use case: summarizing a research paper into a blog post draft in minutes, saving 30+ minutes of manual work. He emphasizes that this system allows for sophisticated, multi-step task execution, acting as a virtual Chief of Staff. The company is currently self-funded after a $1.25M seed round and is focusing on building out these agentic capabilities, including integrations that allow the agent to perform actions like adding items to an Amazon cart or interacting with their codebase for bug fixing. The core benefit is allowing founders to focus on high-leverage activities like achieving product-market fit and scaling the business, rather than routine operational tasks. ### OpenClaw Overview - Co-founder/CEO Presh Denesh Kumar runs OpenClaw, which was incubated at Launch; the company has five products in market covering analytics pipelines, email as an action layer, content pipelines, and customer support/bug fixes. ### Agentic Automation Demo - The system uses LLMs (like Gemini) to perform multi-step tasks; an example shows the agent querying PostHog data and creating a summarized Google Sheet based on a natural language prompt in Telegram. ### Email as an Action Layer - Agents are given dedicated inboxes (e.g., agentname@iru.com) to autonomously handle tasks like research, scheduling follow-ups, and analyzing documents by being CC'd on relevant emails. ### Customer Support Automation - A cron job searches the support inbox for bugs, the agent researches the bug in the codebase, and then posts a summary/solution to a dedicated Slack channel, all without manual intervention. ### Business Viability and Funding - OpenClaw is currently self-funded, having raised $1.25M in seed funding, aiming for a valuation where founders retain high equity percentages (not diluting more than 20-30% per round) to maintain control and rapid iteration. ![Screenshot at 00:00: The host, Presh, is introduced against a snowy window backdrop, wearing a New York Knicks hat.](https://ss.rapidrecap.app/screens/vhFsdz8_L7g/00-00-00.jpg) ![Screenshot at 00:50: A screen share displays a Telegram conversation demonstrating an agent named 'Ewa' fulfilling a request to compile a list of 'Top 10 Active Tempo Users' and generating a Google Sheet link.](https://ss.rapidrecap.app/screens/vhFsdz8_L7g/00-00-50.jpg) ![Screenshot at 01:08: The host asks the guest, Vishnu, when he first saw OpenClaw go viral on Twitter.](https://ss.rapidrecap.app/screens/vhFsdz8_L7g/00-01-08.jpg) ![Screenshot at 02:10: A screen share displays a Notion document outlining OpenClaw's founder use cases, specifically detailing the 'Analytics -\> Action Pipeline' and 'Email as an Action Layer' sections.](https://ss.rapidrecap.app/screens/vhFsdz8_L7g/00-02-10.jpg) ![Screenshot at 10:41: A promotional graphic appears detailing the 'ROSES Template for Business Applications,' a structured prompt engineering framework.](https://ss.rapidrecap.app/screens/vhFsdz8_L7g/00-10-41.jpg)