The product's engine. Trained on your business, not on open knowledge. They talk, decide and act on your systems, and escalate to human when needed.
Set each agent’s tone, goal and rules — no code required.
KEY CAPABILITIES
Each agent with its tone, language, scope and prohibitions. Configured by your team, not by us.
Reads what the user wants and how they feel in every message, and reacts accordingly.
Applies your sales team's criteria and returns qualified leads, with score recomputed on every interaction.
When needed, hands the case to the right person with full conversational history.
EVERYTHING INCLUDED
TYPICAL USE CASES BY INDUSTRY
Qualifies a master's candidate via WhatsApp on Sunday at 22:14, validates requirements against the SIS, uploads documents and books the interview with the director.
Confirms an appointment at 19:30, pulls the patient history from the EHR, handles the reschedule and notifies the specialist.
Receives a call from a B2B customer asking about a spare part, checks stock in the ERP and opens the sales order.
Qualifies a new tax-advisory lead, collects documentation with automatic verification and books with the tax partner.
FREQUENTLY ASKED
No. Models are not trained on customer conversations, cases or documents. Memory stays inside the customer's tenant and is not used to improve generic models.
It automatically escalates to the right human (rep, department or support) with the full conversation history and reason for escalation.
Yes. Voice, WhatsApp, email, web and SMS share the same agent, context and memory. A conversation can start by voice and continue on WhatsApp.
A pilot with a defined use case is in production in 14 days. The agent can start from just a URL of your business and be refined later.
In 30 minutes we'll show you how Twinny connects to your operation and starts delivering results from week 1.