‎I like the Reddit example because community-driven growth can be surprisingly powerful when the product actually solves a problem people are already talking about. You don’t necessarily need to “market” the product aggressively. Sometimes being genuinely useful in the right conversation is enough to get those first few users.

I’d probably start with an AI-native platform if I were building a support stack today. Not because Zendesk or Intercom are bad, but because I’d rather have the AI built into the workflow from the beginning than bolt it onto an older process and hope everything fits together.

‎I think the biggest benefit is having one place to manage everything instead of constantly fixing five different bots. But I’d still be a little careful with the idea of one agent for everything. Chat, email, and voice all have different expectations, so consistency is great, but the responses probably still need some channel-specific tuning.

I’ve run into this too. Customization is great when you’re trying to solve a specific problem, but it can quietly turn into a maintenance headache later. I think the sweet spot is customizing the parts that actually make a difference for users and keeping everything else as simple as possible.