Trusted by support teams at








Extendly: a 500% improvement, scaling across 1,000+ agencies and up to 400,000 end users.
Why enterprise teams choose Botpress
Every time a human support agent steps in — corrects a response, handles an escalation, resolves a ticket differently — that signal feeds back into how the AI agent performs next time. Resolution rates don't plateau after launch. They compound over time, shaped by the actual decisions your team makes.


Volume, auto-resolution rate, CSAT, escalation rate, and time to resolution in one view.
Surveyed and AI-inferred scores, broken down by product tag, topic, and time period.
See which topics the AI agent handles consistently — and where to invest next.
What changed, when, and by whom. Compare versions, audit, roll back.
Enterprise customers get a named customer success manager who knows the account before go-live and stays with it after. They track deployment health, flag issues before they become escalations, and bring product expertise into your specific setup.

Audited annually against security, availability, and confidentiality criteria. Compliant with EU data protection and US healthcare data privacy standards.
PII and sensitive data are stripped before anything reaches an LLM provider. Documents are never sent for ingestion or training — only anonymized chunks that can't be traced back to a client.
No retention on the provider side, no training on your data. You control retention periods for messages and media to meet policy or regulatory requirements.
A full trail of what the AI agent said, what it decided, and why. Data Subject Requests can be fulfilled without dependency on Botpress.
User inputs are sandboxed away from system prompts. Out-of-scope inputs never reach the model; topics and phrases can be blocked outright.
Helpdesk operations exposed as MCP tools — connect internal systems without custom integration work. Works with Claude Code, Claude Desktop, and Cursor.


Enterprise agreements are scoped to the deployment, not the headcount. Pricing is based on conversation volume — not on how many human agents are on your team.

We built it directly into the prompt: whenever someone mentions fraud, security, or cancellations, route to a real person immediately. Having that level of control over sensitive conversations was huge for us.
