Botpress

Operator's take

Most businesses that want an AI support agent face the same fork: pay steep per-seat fees to Zendesk or Intercom, or hand the project to a developer and wait months. Botpress is the middle path — drop it on top of your existing helpdesk (Zendesk and Intercom connect via OAuth, nothing migrates), configure your knowledge base and escalation logic yourself, and go live in days without engineering resources.

The core bet Botpress is making is that AI spend shouldn't be a separate line item you optimize against. Pricing is conversation-based — LLM inference, embeddings, and web search are bundled into each conversation you buy, with no separate token math. Free tier includes 100 conversations/month; Plus starts at 250 conversations included (scaling in $65 packs of 100), Team is $750/mo for 1,500 conversations, and Enterprise is custom. There's no per-seat cost at any tier — they make this a centerpiece of the pitch, framing it as $0 per seat against Zendesk's $115 and Intercom's $85. The result is a support operation that can resolve routine volume autonomously while escalating complex tickets with full context — no cold handoffs.

Where Botpress runs out of road is with genuinely non-technical users who just want something simple and quick. The platform is flexible enough that its flexibility becomes friction if you don't have at least a light technical background. It's also not the right tool if you're mostly building internal workflows rather than customer-facing conversations — tools like Make or n8n own that territory.

What it's good at

• Visual flow builder — drag-and-drop conversation design means you can map out branching logic without writing code; complex paths that used to take months get built in days.
• LLM-powered understanding — plugs into modern language models so the bot handles natural phrasing and context across a conversation rather than matching exact keywords.
• Custom knowledge base — train your bot on your own documentation, website copy, or support data so it gives answers that are actually accurate for your business.
• Multi-channel deployment — one build reaches web chat and WhatsApp out of the box, with voice available on the Enterprise tier; a broader Hub covers integrations to other tools.
• Built-in analytics — see where conversations drop off, which flows get escalated, and what questions the bot fails on; the feedback loop for improving the bot is built in.
• Conversation-based pricing, no per-seat cost — Free tier gives 100 conversations/month; paid plans (Plus, Team, Enterprise) scale conversation volume with AI spend bundled in — no separate token math, no per-seat fee at any tier.

What it's not

• Not plug-and-play for non-technical users — advanced customizations, API integrations, and anything beyond the template flows require comfort with logic-building; simpler tools like ChatBot by LiveChat have lower friction if your use case is basic.
• Not an internal workflow tool — Botpress is built for customer-facing conversation, not for connecting your internal apps and automating background processes; that's Make or n8n territory.
• Not deep on voice if you're below Enterprise — voice channel support exists but is gated to the Enterprise plan; if voice is your primary channel and you're not at that scale, purpose-built telephony platforms run deeper.
• Not developer-optional at scale — customizing integrations with complex backend systems or building production-grade error handling will pull a developer in regardless.