@Ryan B it is one or the other. I use assistable for conversational AI.

@Ryan B Our main solution is a whatsapp chatbot that runs in ghl triggered by leads from Meta Ad campaigns. For chat-widgets, we use GHL widgets (sadly as they are really crappy looking)

Last time I asked, it was not working and I’m not sure it’s been resolved….

@Saied Balajio Here's what Claude says: XML prompts offer several key advantages over standard text prompts for AI agents: Structure and Clarity XML provides explicit hierarchical structure through nested tags, making complex instructions much clearer than paragraph-based text. Instead of relying on natural language markers like "first," "second," or "additionally," you can use tags like <step1>, <step2>, or <examples> to create unambiguous sections. Selective Processing AI agents can easily parse and extract specific sections of XML prompts. For instance, if you have <examples>, <constraints>, and <output_format> sections, the agent can process each independently and apply them at the appropriate stages of reasoning. Reduced Ambiguity XML eliminates many interpretation issues common in natural language. When you write <required>Generate exactly 3 options</required> versus "Please generate about 3 options," the XML version leaves no room for ambiguity about what constitutes compliance. Better Instruction Separation Complex prompts often mix different types of instructions - context, examples, constraints, and output requirements. XML allows clean separation: <context>, <examples>, <constraints>, <output_format>. This prevents the agent from confusing examples with actual instructions. Conditional Logic Support XML naturally supports conditional instructions like <if_user_asks_for_code> or <when_handling_sensitive_topics>, making it easier to create adaptive prompts that respond differently to various scenarios. Easier Prompt Engineering XML makes it simpler to modify, debug, and version control prompts. You can comment out sections, reorder components, or modify specific parts without affecting the overall structure. Consistency Across Interactions XML prompts tend to produce more consistent outputs because the structured format helps the AI maintain focus on each component of the instructions throughout its reasoning process.

@Ron G TBH it’s no extra work upfront… formulate a solid text prompt and run it through Claude to convert to xml…

following

I've only seen workflows in n8n via Telegram that require voice notes to be transcribed. **Following**