Ugrás a tartalomhoz
← Back to Nortinia AI Chat
AI
Nortinia AI Chat · AI chatbot

AI chatbot for websites and webshops with an embeddable widget

Nortinia AI Chat is an embeddable AI chatbot widget that learns from your own knowledge base (PDF, DOCX, HTML, webshop data) and answers via a RAG (Retrieval-Augmented Generation) engine with source citations. It is multilingual, remembers conversation context, and automatically hands off to a human operator when a case exceeds its scope — every step logged in a GDPR-compliant audit trail.

View details

What it does

Own knowledge base indexing (PDF, DOCX, HTML, webshop)
RAG engine with source citations and live retrieval
Multilingual: Hungarian, English, German and 20+ more
Embeddable widget — a single <script> tag, any page
Automatic human handoff (escalation) based on sentiment
GDPR-compliant conversation log, EU-region data storage

How people use it

SaaS product onboarding

New SaaS users ask about features, integrations and billing during onboarding. Nortinia AI Chat indexes the full documentation, help centre and changelog, so users get precise answers with source citations — and if it is a technical case, handoff to the right support team is a single click away.

Webshop customer service

Webshop visitors ask about delivery times, sizing and return policy. The AI chatbot indexes the product catalog and customer information pages, acts as a cart assistant recommending products, and only escalates the more sensitive complaints to a human — handling 80% of questions on its own.

Financial services

Financial service customers ask about account opening, pricing and KYC. Nortinia AI Chat answers only from public information, immediately routes KYC cases to a human advisor — every conversation logged for the compliance audit.

Travel agency

Travel portal visitors ask about destinations, visa requirements and cancellation policies. The AI chatbot is multilingual (the guest picks their own language) and shows live availability via the booking system — final booking is closed by a human sales rep.

Frequently asked

Can it be embedded into any website?

Yes. A single <script> tag in the head works on WordPress, Shopify, plain HTML, Webflow or a custom React/Next.js site. The widget size, colour, position and opening message are configured from an admin panel.

How much does the RAG engine actually use from our documents?

The RAG engine runs hybrid retrieval (vector + keyword) on every answer and cites the top-K document chunks. Each answer carries a source link so customers can trace back to the original text. The model only answers within the source frame — if no relevant document exists, it says so and routes the case to a human.

How does it handle conversations under GDPR?

Every conversation is stored on EU-region servers in an encrypted database. Retention is configurable (30/90/365 days), and customers can request deletion at any time — one click in the admin panel. The Data Processing Agreement (DPA) is provided by default.

Will it process our own organisational documents?

Yes. PDF, DOCX, HTML, CSV and any structured document export are supported. Documents are uploaded encrypted, indexed only in your own tenant, and never mixed with other customers’ knowledge bases. Refresh can be cron-driven or triggered via webhook.

Interested? Let us chat for 15 minutes.

Send a quote request and hear back within 24 hours. No commitment, no pressure.

View details