It depends on the modules you choose: a basic customer-service agent (LLM connector, conversation memory, one Zalo or web channel) is far lighter than a full platform with Agent Studio, streaming avatars and conversation analytics. The tool converts each module into token AI usage and person-months so you can see a cost range before talking to TOT.

If you only need a basic chatbot with low conversation volume, renting Dify or n8n can make sense. But once subscription fees scale with conversations, you need deep workflow customization, or you want customer data inside your own infrastructure, building usually pays back quickly. The tool lets you put both numbers side by side.

A customer-service agent stores conversation history, contact details and sometimes transaction data, so the system must comply with Vietnam's Decree 13/2023 on personal data protection. When you build it, TOT designs access control, encryption and audit logs from the start and keeps data in infrastructure you control rather than sending it to a third party.

These are reference estimates based on an effort framework and current token AI pricing, not a committed quote. Real cost depends on integration complexity, number of channels, data volume and SLA level. TOT runs a detailed assessment before issuing a formal quote.

The tool gives a cost estimate with a reference price list for each AI plan. It is not an official quote — for a detailed quote on your AI software (by scope, features and integrations), contact TOT to scope it.