This article outlines four distinct workflows that can be implemented using a Hermes agent system to automate business processes and save time. These workflows are presented as actionable prompts that users can adopt immediately. The accompanying visual prop, detailing these prompts, is available for free access at clearmud.ai under the Resources section, specifically within the Hermes category.

1. Daily Business Watchtower Brief

This workflow focuses on generating a daily business brief, serving as a "daily business watchtower." The objective is to automate research and provide a concise summary of critical business information each morning.

Prompt Example: "Hermes, create a daily business brief for my business. Check the following sources, and give me what changed, top priorities, risks or blockers, opportunities, decisions I need to make, recommended next steps. Separate for your information from needs action."

This prompt instructs the Hermes agent to monitor specified sources, identify changes, and report on key business aspects. The output is categorized into information for general awareness and items requiring action. The system aims to reduce time spent on manual research and prevent distractions. An example of this workflow in practice is the "Oracle" agent, which provides a daily "morning operator brief." This brief is customizable through real-time adjustments to the agent's configuration. The system leverages GPT-4.5's reasoning capabilities to generate comprehensive reports, which are then refined by the user. This process has been applied to overnight builds for a micro SaaS idea, an open-source idea, and an existing product upgrade. The research output includes pinpointed information with source URLs for verification. The focus of this multi-agent system is on front-facing projects, products, and revenue-generating ideas, distinct from internal day-to-day operations. Overarching AI industry news and ClearMud-related YouTube updates are managed separately. The Hermes agent setup is designed to function as a "product OS" for all front-facing projects.

2. Content to Campaign Pipeline

This workflow transforms a single idea into a comprehensive content campaign. It demonstrates how an agent can automate the process of developing marketing strategies from initial concepts.

Prompt Example: "Hermes, turn this idea into a content campaign."

This prompt is exemplified by the "Prompt Browser" tool, which extracts prompts from YouTube videos. The tool identifies prompts mentioned verbally or displayed on screen, providing timestamps for easy access. The campaign development process involves identifying a core frustration (e.g., users spending excessive time on long videos to find specific prompts) and positioning the product as a solution. The output includes campaign concepts, core positioning, content pillars (e.g., tutorial fatigue, prompt extraction, faster building), and desired user behavior (e.g., build a prompt library). The Prompt Browser tool is free to use and allows users to extract prompts from any YouTube video without requiring sign-in, with an optional sign-in for saving prompts and future features.

3. Overnight Business Improvement Build

This workflow focuses on automating local business improvement builds overnight. It emphasizes the importance of human review and approval before deploying any AI-generated changes to production to avoid contributing to low-quality content online.

Prompt Example: "Hermes, run an overnight local business improvement build. Goal: I want to improve X, Y, and Z. For example, create one new custom workflow for us based off of all of our agent profiles."

A practical application involves instructing an agent to create a marketing kit nightly. This kit would promote a product or YouTube video, providing ready-to-preview and edit copy for text-based social media platforms. The prompt can be extended to include the creation of accompanying creative assets via an image generation model. The process involves scheduling this as an overnight improvement, with the output reviewed and refined by a human the following morning. This workflow also incorporates delegation rules, ensuring that tasks are routed to the appropriate department heads, such as a marketing department head (e.g., "Mouse"), and that relevant tasks are logged in their runtime inbox. The system relies on pre-configured rules of delegation and delegation maps, with a designated "single source of truth" and a Kanban board for task management. All builds are intended to remain local and for preview only, with no direct deployment to production without explicit human approval. This approach is part of a "building in public" philosophy, where extensive testing and agent profile creation (e.g., around 30 profiles with specific gateways and delegation responsibilities) precede the implementation of automated workflows. The Hermes agent system is designated for front-facing, revenue-generating activities, while internal operations are managed by a separate system (e.g., "Open Claw").

4. Delegation Map Builder

This workflow focuses on creating a delegation map for a business, ensuring that each agent role has a defined scope, routing rules, handoffs, and approval escalation paths. The goal is to avoid creating "dirty chatbots" and instead build a structured delegation system.

Prompt Example: "Hermes, design a delegation map for my business. This is where you fill out your personal details, your personal goals, all of the different lanes you want to focus on."

This prompt aims to generate command center, department lead, and specialist profiles with defined routing rules and approval gates. It is recommended to sketch out a delegation map manually before feeding it to the agent. The output can include starter and expanded organizational maps. The system is designed to be modular and adaptable to various Hermes agent dashboards. The process involves defining personal details, goals, and desired focus areas. The agent then creates the necessary profiles and routing rules. The example provided illustrates a traditional organizational chart with department heads and sub-agents, clearly showing dotted lines for communication and delegation. Specific examples include engineering roles (front-end, back-end, security, release) with an "architect" as the routing owner, and marketing roles (content, video directory, script writer, designer). The Hermes system is positioned as the front-facing AI agent system for revenue generation, while internal operations are handled by a separate, trusted system. The prompt emphasizes starting simply, perhaps with a single agent (e.g., an administrator managing API-related work and system connections), and then expanding. The example shows the successful creation of a local marketing kit for "Prompt Browser" by the "Mouse" agent, including a recurring cron job for overnight marketing kit generation and a note dropped into Mouse's runtime inbox. The review of the generated kit includes identifying blockers and recommending next steps, such as reviewing the kit's markdown file and selecting the strongest hook.