Activity
Mon
Wed
Fri
Sun
Dec
Jan
Feb
Mar
Apr
May
Jun
Jul
Aug
Sep
Oct
Nov
What is this?
Less
More
16 contributions to AI Automation Society
Help me out: How Are You Structuring Your AI Workflow, Handoffs, and Project Tracking?
Hi everyone, I’m currently rebuilding the foundational work processes for my projects. While everyone’s individual tech stack is different (Claude Code, Codex, Cursor, VS Code, various chat interfaces, etc.), I believe the underlying processes our agents and tools follow should be relatively structured, repeatable, and clearly defined. Right now, I'm trying to rebuild my own process to solve the classic challenges around clear work responsibilities, clean agent handoffs, and knowing where things stand. I’d like to know how you’ve set this up in your own operations. Specifically, how do you handle: 1. Scope & Assignment: How do you define what work needs to be done for a given project, and how do you decide who (or which specific tool/agent) does it? 2. Platform & Agent Responsibilities: How have you cleanly separated what each AI platform or agent does versus what the human-in-the-loop is responsible for? 3. State & Progress Tracking: How do you track where a project is at any given point in time? (I’m currently re-evaluating whether project management tools like Linear are the right fit, or if there's a leaner way). 4. Remaining Work & Milestones: How do you dynamically evaluate what remaining work is needed to hit an objective? 5. Completion Criteria: How do you definitively know when a project or slice is actually finished? 6. Stack Organization: How do you orchestrate and separate your tools to stay aligned and drive projects across the finish line without constant friction? If you’ve built out a structured process that works—whether it's for solo development, client builds, or managing multiple systems—please walk me through how you designed it, what your handoffs look like, and how you keep everything aligned. Would love to hear what’s working for your stack so I can apply those principles and best practices to mine. Appreciate any insights you can share! @Nate Herk @Nate Herk
Mastering Agent Architecture: How Do You Design Your Projects in Claude Code and N8N?
Hey everyone, As many of us continue to build applications and automate workflows, a key challenge is figuring out how to structure our projects effectively. I'm curious about how you all approach this in Claude Code and/or N8N, since both platforms involve similar concepts of organizing tasks. I'd love to hear about your strategies and best practices for creating and structuring workflows, agents, and sub-agents. Could you share your insights on topics like: - When do you decide to create a new workflow, agent, or sub-agent? What's your process for determining what needs its own specialized function versus what can be handled within an existing one? - How do you differentiate between global and project-specific assets? Are there specific use cases where you find one more effective than the other? - What are some examples of workflows or agents you've built and the problems they solve? Sharing a few concrete examples would be incredibly helpful in understanding your approach. Your insights would be a huge help to me and the entire community as we work to build more efficient, scalable, and well-organized automations.
0 likes • Sep '25
@Nate Herk
0 likes • Sep '25
@Karan Paliwal What are some examples of workflows or agents you've built and the problems they solve? Sharing a few concrete examples would be incredibly helpful in understanding your approach
Idea for a Dynamic AI Knowledge System for any CRM (HubSpot & GHL)
@Nate Herk @Nate Herk And all community members, I've got an interesting challenge that might benefit the whole community. The Problem: I'm using Claude to help build out my SaaS business with HubSpot as my CRM. The issue is that Claude's knowledge of HubSpot is outdated. HubSpot releases new features frequently, but Claude doesn't have access to their latest documentation or knowledge base updates. This creates a gap between what HubSpot can actually do today versus what Claude thinks it can do. The Opportunity: I'm looking to build a system that creates dynamic understanding between: - My AI assistant (Claude or any other tool) - HubSpot's current capabilities (from their live knowledge base) - My specific business needs and goals Essentially, I want the AI to have real-time awareness of HubSpot's latest features and understand how to apply them to my use cases. My Question: How would you architect something like this? I'm thinking it could involve: - Scraping/ingesting HubSpot's knowledge base - Creating a dynamic context layer for the AI - Building a bridge between the AI and the CRM's current capabilities This would be valuable for anyone using AI assistants with frequently updated CRM platforms, not just HubSpot. What are your thoughts on the technical approach? Have you seen anything similar built? Thanks!
Is the SaaS Business Model Dying? The Rise of Fully Agentic Tools
I've been thinking a lot lately about the changing landscape of software — especially with so many new tools becoming fully agentic, taking on tasks previously handled manually or by traditional SaaS products. Is anyone else wondering: Are SaaS businesses and the SaaS business model on the decline? Could they even be “dead” in the next 12 months? - With agentic platforms automating not just features but entire workflows, what does this mean for SaaS founders and users? - Is the classic subscription-based SaaS model becoming obsolete, or will it adapt and survive? - Are we heading towards a future where “pay-per-agent/action” replaces monthly recurring revenue? I’d love to hear what this community thinks. - If you run a SaaS, are you worried, adapting, or excited? - As users are agentic tools “good enough” to replace SaaS, or will there always be space for both? Let’s start a real discussion: Do you think SaaS as we know it will be dead in a year’s time? Or do you see a path forward? Looking forward to your thoughts!
Poll
4 members have voted
0 likes • Jul '25
@Nate Herk
Where Do YOU Draw the Line? Seeking Your AI Automation Rules
Hey everyone, With AI tools like browser agents, voice agents, and platforms like Genspark and Manus AI, we can now automate nearly all of our non-physical work—from making phone calls and building presentations to deploying apps. This brings up a critical question for all of us: Where do you draw the line between full automation and keeping a human in the loop? I'm trying to create a framework for myself and would love to hear your thoughts on: - Trust & Security: Do you let AI run code locally, access private API keys, and build/deploy applications for you? What are your non-negotiable security boundaries? - Client-Facing Tasks: Would you trust an AI to call customers or important leads on your behalf? If so, under what conditions? - Your Personal "Automation Ceiling": What tasks do you intentionally keep for yourself, even if AI could do them? What's your reasoning (quality control, personal enjoyment, strategy, etc.)? As we get closer to AGI, defining our roles alongside AI is more important than ever. I'm looking forward to hearing how you're all navigating this!
0 likes • Jun '25
@Nate Herk , please promote this post so we get maximum discussion on this very important topic.
1-10 of 16
Jonathon Tamm
4
70 points to level up
@jonathon-tamm-2465
Hi

Active 3d ago
Joined Sep 21, 2026
Powered by