Everything an agent needs to run a conversation
Official channel APIs, a visual builder, connections to your data, governance, a shared inbox, analytics and a developer API — multi-tenant, with SSO, RBAC, audit logs, data residency and an uptime SLA.
One agent brain across every channel
Orchestrate agents visually
A visual builder for triggers, agent steps, data lookups, actions and handoff — with prompt-defined agents, templates, guardrails, approvals and testing against real conversations.
Prompt-defined agents
Describe the agent's job, tone, sources and limits in plain language; the builder handles orchestration around it.
Guardrails and approvals
Source restrictions, confidence thresholds, blocked topics and human-in-the-loop approvals for sensitive actions.
Test with real conversations
Replay past conversations against a draft agent before release, and keep change history for every version.
Trigger
Message · event · webhook
Agent step
Prompt-defined
Lookup
Connected source
Action
CRM · payment · calendar
Approval or handoff
Human in the loop
Trigger
Message · event · webhook
Agent step
Prompt-defined
Lookup
Connected source
Action
CRM · payment · calendar
Approval or handoff
Human in the loop
Connected to the systems that hold your data
HubSpot
CRM
Salesforce
CRM
Zoho CRM
CRM
Shopify
Commerce
WhatsApp Catalog
Native
WhatsApp Payments
Native
Razorpay
Payments
Google Sheets
Data
Zapier
Automation
Webhooks
Automation
REST API
Automation
Controls that hold at scale
Agents are useful only if they are safe to leave running. Governance is built into the platform rather than bolted onto the agent.
Source-bound answers
Agents answer from connected sources and cite them; unknown means handoff, not a guess.
Guardrails and approvals
Blocked topics, confidence thresholds and human-in-the-loop approvals for sensitive actions.
Change history and testing
Every agent version is tested against real conversations before release and kept in history.
Draft agent
New version
Replay real conversations
Test set
Approval
Reviewer sign-off
Release
Change history kept
Draft agent
New version
Replay real conversations
Test set
Approval
Reviewer sign-off
Release
Change history kept
Humans where they matter
Every conversation the agent does not finish lands in a shared team inbox with full context, so the human never starts cold.
One inbox, every channel
WhatsApp, Instagram, Telegram, LINE and web chat in one queue with assignment rules and SLAs.
Handoff with context
The transcript, customer record, agent notes and the reason for escalation arrive with the conversation.
Team roles
Seats, roles and workspaces map to how the team is organised — by brand, outlet, region or function.
Agent stops
Low confidence · policy
Inbox queue
Assignment rules
Team member
Transcript + record
Resolved
Outcome logged
Agent stops
Low confidence · policy
Inbox queue
Assignment rules
Team member
Transcript + record
Resolved
Outcome logged
Build on the platform
A REST API and webhooks for the whole platform: conversations, messages, agents, workspaces, outcomes and analytics — with API keys scoped per workspace.
Built for procurement, security and scale
Security & control
ISO 27001 processes, SSO, role-based access, PII redaction and audit logs across every workspace and channel.
Observability
Every conversation, agent decision and handoff is logged and queryable — with analytics on reply time, resolution and revenue by channel.
Deployment & rollout
Managed and Network tiers run pilots against defined metrics, then roll out in waves across outlets, brands and regions.
AI governance
Agents answer from approved sources, within guardrails, with human-in-the-loop approvals and tested against real conversations before release.