This community is about Claude Code and n8n, so I figured it’d be worth sharing exactly how I’m using both in a production automation system. Setup: KnockKnock (AI SMS) + GHL (CRM/calendar) + n8n (orchestration) + Claude (intent classification) Here’s the full flow: Trigger (GHL): Missed call or form submit → GHL fires a webhook to n8n. First n8n check: query GHL to see if an active conversation already exists on this contact’s number. If yes → stop. Prevents double-threading when a lead calls twice. First message (KnockKnock): KK sends an SMS within 60 seconds. “Hey [Name] — saw you reached out about [service type]. Is that still something you’re looking into?” Short. Human. No pitch. Intent classification (Claude via n8n): When the lead replies, n8n passes the text to Claude with a simple system prompt: “Classify this SMS reply as HOT, WARM, or COLD. HOT = clear interest or booking signal. WARM = soft interest. COLD = not interested or wrong person.” Claude returns a single word. n8n routes via switch node. HOT branch: Send GHL booking link → appointment auto-creates in GHL calendar → pipeline stage updates → owner notification fires. WARM branch: Contact enters GHL 48-72hr nurture sequence. If they re-engage → back into HOT routing. COLD branch: Tag in GHL. Remove from active flow. Long-term drip. Write-back to GHL (every branch): n8n updates the GHL contact record: • Last intent: hot/warm/cold • Service mentioned: extracted from Claude’s output • Execution ID: [YYYY-MM]-[workflow-name]-[version] • Last message timestamp The Execution ID matters: query it at the start of every n8n run to detect duplicates. When debugging a specific contact, it gives you the exact execution to pull up. Stack summary: GHL → trigger, CRM, calendar, pipeline, notifications n8n → all logic: routing, Claude calls, duplicate detection, write-back KnockKnock → AI SMS layer Claude → intent classification (one clean word output = easy switch node routing)

New to n8n? Here's what a node actually is. A node is a single building block in an n8n workflow. It can start the workflow, fetch, send, or process data, control the flow logic, or connect to an external service. You link multiple nodes together to create a complete automation.

I'm Aatar Atta, and I help businesses automate operations, reduce costs, and scale with AI. I've joined this community to learn from Brendan and its members, to share my knowledge and contribute to making this community more valuable, and to network with like-minded people.

Same model. Same task. Same success rate. Half the bill. A new study from UC Berkeley and Arena put a name to something many of us building with AI agents have felt but rarely measured: the harness tax. The harness is the agent wrapper around the model: Claude Code, Codex CLI, Pi, and others. It turns out the harness can change your inference costs as much as the model you pick. A few findings stood out: → Claude Fable 5 solved ~97% of tasks in Claude Code, Codex CLI, and Pi. But Claude Code cost $1.33 per rollout, while Pi cost $0.67. → On SWE-bench Lite, Claude Code cost about 2x more than Pi and 1.6x more than Codex. Success rates were within 2 percentage points of each other. → Much of the gap appears before the agent does anything. Claude Code starts with 27,000+ tokens of context, compared with ~2,000 for Pi. → Models don't always perform best inside their own vendor's harness. Simple setups are often surprisingly competitive. The most interesting part for me: even when two setups post identical benchmark numbers, they often fail on completely different tasks. A leaderboard score won't tell you which one fits your codebase. The researchers' advice is practical: 1- Test a few model and harness combinations on your actual engineering workload 2- Measure cost per solved task, not raw cost per run 3- Pick the cheapest option that clears your reliability bar 4- Retest whenever the model or harness changes If you're running agents at scale, whether for your own product or for clients, this is the difference between a margin and a leak. The best agent setup isn't the most elaborate one. It's the one that solves your problems at a cost you can sustain.

Hi everyone, I’d like some advice on the project I’m describing below. For the past weeks, I’ve been building a WhatsApp chatbot demo in n8n to handle appointments and answer FAQs for a medical clinic. I’m using Google Sheets and Google Calendar. I don’t have much experience building this kind of thing in n8n, and I’ve noticed that the AI ​​model (Gemini) makes several errors in its responses. It mixes up dates and times, and sometimes claims there are no free slots when there actually are. I know this is part of the troubleshooting and fine-tuning process, but I already have a fairly extensive System Message with many rules designed to prevent these errors, and it keeps making them. So, my question is: for a WhatsApp chatbot, is it better to use fixed options (like a menu) to avoid these errors, rather than having the AI ​​respond with free-form text? My goal is really to give customers the feeling that they’re talking to a person, not a robot. However, if the AI ​​is prone to so many errors, I’m wondering if it would be better to use buttons, fixed menus, or a hybrid approach?m What do you recommend? Thanks in advance.