Financial AI · Omnichannel chatbot platform

ChatBot

One conversation engine.
Every channel your customers reply on.

A configurable, reusable and operable chatbot platform for overseas collections and marketing. It picks up every inbound SMS, WhatsApp and Viber message, replies with AI and keeps a human-like, multi-turn conversation going — always within the rules of each channel and market.

  • SMS
  • WhatsApp
  • Viber
  • Other IM
  • Voice
Book a ChatBot demo Watch a conversation →
Robot types on one engine
0
Text and voice channels
0+
Typical reply time
<0s
Conversations tagged and logged
0%

Who does what

The hub owns the channels. The AI owns the conversation.

A clean split keeps each side simple: the call hub connects channel resources and executes delivery; the AI side decides what to say and which robot to use — joined by standard text and voice interfaces.

CALL HUB

Channels & execution

  • Channel resources — SMS gateways, WhatsApp accounts, voice lines
  • Reach execution and delivery status
  • Packages each Robot instance with its channel resources
  • Unified data tables and callbacks
STANDARD INTERFACES
Text APIMessage in → reply, tags and next action out
Voice APIAudio stream in and out · ASR and TTS run inside the AI side
AI SIDE

Conversation & strategy

  • Playbook Chat Engine — scripts, utterances, dialogue control
  • Intent tagging and knowledge-base management
  • AI reach strategies on the decision-engine flow canvas
  • Visual strategy configuration, list intake and task output
BUSINESS TEAMS Configure scene robots Configure reach strategies Use overseas robot capabilities

How it works

Two modules. One engine underneath.

The Chatbot module decides how to talk; the reach-strategy module decides when to use which robot. Both run on the same large-model capabilities and AI infrastructure.

AI product framework

The AI side focuses on conversation and strategy decisions; the call hub focuses on channel resources and execution packaging.

AI PRODUCT SYSTEM

01Chatbot robots

Owns how to talk

Text robotSMS · WhatsApp · IM — the text conversation base
Voice robotOutbound voice — the voice interaction base
Playbook Chat EngineOne multi-turn engine: node transitions, context, variables, FAQ, knowledge base and reply modes — driving both text and voice robots

02Reach strategy orchestration

Owns when to use which robot

High-code strategy flowsRules, decision flows and robot-invocation timing
Visual orchestrationLow-code configuration so business teams build strategies fast
LLM decision-makingComplex conditions, dynamic robot choice, strategies from natural language and smart recommendations — from rule orchestration to intelligent decisions

AI large-model capabilities

LLM dialogue — in-house multilingual and third-party modelsIn-house dialogue tagging modelKnowledge baseFAQPrompt managementVariables & contextState managementContent generationDecision-flow engine

AI infrastructure

Model gatewayInferenceVector searchLog monitoringConversation recordsStrategy execution logsConfig centreVersion management

Call hub

  • Channel resources
  • SMS · WhatsApp · lines
  • Scene robot packaging
  • Delivery chain management

Business teams

  • Configure scene robots
  • Configure reach strategies
  • Use overseas robot capabilities

See a conversation

Same engine. Different channels, different jobs.

Pick a scenario and watch the playbook node, intent tags, variables and next action update with every message.

chat://whatsapp WhatsApp

What the AI side delivers

Five commitments, one platform.

  1. 01

    One Playbook Chat Engine

    A single engine that offers two base capabilities to the business: text robots and voice robots.

  2. 02

    Text and voice robots

    Text robots carry SMS, WhatsApp and other IM conversations; voice robots carry outbound calling scenarios.

  3. 03

    Shared AI foundation

    Both share one dialogue engine on top of LLMs, the knowledge base, FAQ and the variable system.

  4. 04

    Standard interfaces

    Text and voice APIs for the call hub, with ASR and TTS built into the Chatbot itself.

  5. 05

    AI reach strategies

    Visual strategy configuration on the decision engine's flow canvas — list intake in, tasks out.

Channel-compliant by default

Approved WhatsApp and Viber templates, the 24-hour service window, sender IDs and opt-out keywords like STOP handled automatically.

Quiet hours & frequency caps

Per-market contact windows, daily caps and cool-downs apply across every robot and channel.

Human-like, multi-turn

Handles typos, mixed languages and changes of mind, and keeps the thread across days.

Multilingual overseas

English, Bahasa, Thai, Vietnamese, Tagalog and more — including local slang and code-switching.

Tags that feed operations

Intent and outcome tags flow back to the hub and into Lingque strategies and XinBI dashboards.

Versioned and auditable

Every playbook, prompt and strategy is versioned; every conversation and decision is logged for review.

Frequently asked questions

QHow is ChatBot different from the LLM Voice Agent?

They share the same Playbook Chat Engine. ChatBot focuses on text channels — SMS, WhatsApp, Viber and IM — while the voice robot powers outbound calls. One playbook can run on both.

QCan the bot say something it shouldn't?

Playbooks define what can be said at each node. Regulated messages use approved templates; LLM-generated replies stay inside guardrails, and disputes or complaints go to a human.

QHow does it connect to our channels?

Through the call hub, which holds your SMS gateways, WhatsApp Business accounts, Viber senders and voice lines. ChatBot only needs the standard text and voice interfaces.

QWho decides which robot and channel to use?

The AI reach strategy. Operators draw it on the decision engine's flow canvas today, with LLM-assisted decisions coming next.

QWhere do conversations and data live?

In our Singapore cloud, a dedicated cloud, or fully on-premise with private models — with conversation records and strategy logs kept for audit.

Text it. See how it replies.

Share one collections or marketing scenario and the channel you use. We'll build a playbook and let you chat with it on WhatsApp or SMS.

What a pilot looks like

  1. Week 1 · playbook, FAQ and templates
  2. Week 2 · channel approval and simulation
  3. Weeks 3–4 · live conversations on a pilot segment
  4. Then · A/B against your current channel
Book a ChatBot demo Talk to an engineer