@tusharjain_25 At 25 people, I'd split it: chat on Slack/Teams, but daily ops (tasks, approvals, client status) in one simple internal dashboard. Chat tools are bad databases. What's the main thing breaking today?
@bahrulum69 Tip: spec the admin dashboard as carefully as the storefront: order statuses, stock, refunds, roles. That's where most ecom builds get underscoped and go over budget.
@CaptainShipSim With 160 replies, one filter helps: ask each person what the site should DO for the business (leads, bookings, sales) and how they'd measure it. The good ones will answer that before showing a portfolio.
@0xfrenchman Before choosing a dev, it helps to nail 3 things: the single core flow users must complete, iOS/Android/both, and what data needs a backend. Makes quotes way more comparable. What's the app about?
@FedericoS18132 For split tests via third-party checkout redirects, make sure the variant ID survives the redirect into Checkout Champ, or attribution breaks. That's where most offer tests silently fail.
@Shanawajweb For real estate voice agents, the make-or-break is the handoff: qualify (budget, area, timeline), book the viewing, write it to the CRM, then text the agent a summary. Happy to share a quick flow diagram if useful.
@doitwellnow Tip for picking from 50 replies: ask each dev for one live small-business site they built + what it does for leads (forms, booking, CRM). A site that captures and routes leads beats a pretty one.
@oladapizee Booking + follow-up automation tip: keep the source of truth in one place (calendar or CRM) and let n8n only react to webhooks. Saves you from double-bookings when the AI assistant and humans both edit.
@GardRogne Great list. One thing I'd add when vetting: ask candidates how they handle errors/retries in n8n and who gets alerted when a workflow fails. That separates demos from production systems.
@scallro For Gmail + support in n8n, the win is usually: classify incoming mail → draft reply with AI → human approves → log to your helpdesk. Start with your top 3 ticket types, not everything. Happy to share a flow outline.
We saw hype in chatbot's than we build a chatbot into our CRM for ticket system. With a huge rag system. With most powerful models.
But the problem is customers don't want to chat with a bot.
Well thats it. Chatbot is not the solutions you always need.
#chatbot