Nortinia AI Chat is an embedded chat AI that answers from your own knowledge base, in your own brand voice — in webchat, WhatsApp, Messenger, LinkedIn DM, email. With live human handoff. 14 channels, one conversation thread, one admin workspace. No voice calls — that's why it's faster and cheaper than the Nortinia AI Assistant.
A typical 2-person live-chat shift runs €4,500-7,500 / mo. Nortinia AI Chat Growth is €599 / mo. 90%+ savings, 24/7 coverage.
Channels in one thread — webchat, WhatsApp, Messenger, LinkedIn, email…
0%
Cheaper than a 2-person live-chat shift
0
Native languages — English, German, Hungarian
0/7
Coverage — replies on nights, weekends and holidays too
Live chat agent vs AI Chat
Replace the chat shift with an AI chat.
A typical B2C / B2B live-chat shift (2-3 agents, 8-16h coverage) costs €4,500-7,500 a month. Nortinia AI Chat Growth is €599. Same work — at machine speed, 24/7, never quits.
Traditional setup
2-3 person chat shift + tools
€4,500-7,500/ mo
Chat-window is dark at night and on weekends
Every new agent needs 2-3 weeks of brand-voice ramp-up
One agent handles 1-2 conversations at a time, max
Extra Intercom / Zendesk / Drift licence — $200-500 / seat / mo
Churn: every 6-12 months you train a new agent
Nobody re-opens the thread that went cold last Tuesday
Plus the CEO time to schedule shifts. Plus extra chat-platform licences. Plus the customer who never got a reply on Friday.
Nortinia AI Chat
One AI chat that works 24/7
€599/ mo
Replies at night, on weekends, on holidays — keeps context
Brand voice from minute one — learns from your existing content
Handles unlimited parallel conversations at once
14 channels in one admin workspace, one message thread
Doesn't forget: picks up the thread weeks later where you left off
One-click live human handoff — even books the calendar
Live in 1-2 weeks. Scales with traffic. Doesn't take time off, doesn't quit. Just chats.
Simple as a conversation
State the knowledge. Show the voice. Embed the widget.
Most chatbots ask you to be the UX designer, the prompt engineer and the knowledge-base curator. Nortinia AI Chat does all of that. You just supply the knowledge, show it your voice, and drop the widget — from there it talks for you.
1
Give it the knowledge base
PDFs, website, Notion, Google Docs, Zendesk export, GitHub README, e-commerce catalogue. Any source. RAG (Retrieval-Augmented Generation) indexes and structures it automatically — you just upload.
“No prompt engineering. No "intent training".”
2
It learns your brand voice
From your existing content (site, newsletters, past support chats) the AI learns your tone of voice. Casual where you're casual, formal where you're formal. You approve one sample conversation — from there it writes that way.
“Your customer won't notice it's AI.”
3
Drop the widget, you're live
One script tag on your site, or webhook to Messenger / WhatsApp / LinkedIn. Five minutes. From there it chats 24/7. Weekly report: who asked, where there was no answer, where you need more knowledge.
“By morning, 28 new leads in the CRM.”
What it actually does
Ten jobs it takes off your plate — text-only focus.
Not a feature list — an outcome list. Each item is concrete work you used to pay people for, or simply didn't have capacity for. No voice calls (go to AI Assistant for that) — everything else is here, cheaper.
01
Knowledge-grounded RAG
Vector-indexed up to 10k documents. Answers only what's in your source — no guesswork. Cites the source if you ask.
Source-accurate02
Brand-voice learning
Learns your tone of voice from your existing content. Formal or casual? Technical or playful? Funny or serious? You approve a sample, then it writes that way.
Tone-consistent03
14 channels, one admin
Webchat, WhatsApp, Messenger, LinkedIn DM, email-as-chat, Telegram, Viber, Instagram DM, X DM, Slack, Discord, SMS, in-app, Microsoft Teams — one thread, one workspace.
Omnichannel04
Live human handoff
Hands off to a human teammate whenever the AI is uncertain, or whenever the user asks. Books the calendar or opens a ticket in HelpDesk. Nobody feels "stuck in a bot".
Soft handoff05
Sentiment + intent routing
Detects when someone is frustrated, angry, or buying. Routes to sales, support, or retention queue. Silently tags the thread so you see it in reports.
Realtime06
Lead scoring during the chat
Every conversation ends with a lead score: cold, warm, hot. Auto-writes to your CRM (HubSpot, Pipedrive, Salesforce): what they came for, what they asked, what we offered.
CRM integration07
Multi-language natively
English, German, Hungarian at native level. Replies in the customer's language if they switch — automatic detection. The German connector knows Sie / du.
EN / DE / HU08
GDPR + audit log
Every conversation logged server-side, optional PII masking. User DSAR + erasure API. Every answer's source is traceable — fully auditable.
EU compliance09
Our own Nortinia AI Engine
Not OpenAI-dependent. The Nortinia AI Engine picks the best model per conversation — GPT, Claude, or a custom fine-tune. Cost- and quality-optimised.
Built in-house10
On-premise option
On the Scale tier, the whole Nortinia AI Chat stack runs on your own server / VPC. Data never leaves your company. SLA + dedicated customer success.
Enterprise
Not a chatbot — a chat AI
Forget the threatening, "sorry I don't understand" rule-based bots.
People learned to hate the old chatbots for a reason — they were static rulebooks driven by decision-tree logic. Nortinia AI Chat uses real language models, on YOUR knowledge base, in YOUR brand voice. It doesn't trick, it doesn't hallucinate — if it doesn't know, it says so and hands off to a human.
No "if-then-else" tree — every sentence interpreted in context
Recognises returning visitors — picks up last week's conversation
RAG: answers only from what's in your knowledge base — hallucination lock
"I don't know" is a valid answer — one-click human handoff
Brand voice consistent in every reply — not AI-smell, but you
Multi-turn memory: message 47 remembers message 1
Typical daily volume for one customer
Conversations answered412
Leads generated (hot)28
Human handoffs11
Average response time0.8s
Channels active8 / 14
Customer CSAT4.7 / 5
What you replace
How much do you save vs a chat shift?
Here are typical EU market salaries and chat-software costs. All of them are absorbed by Nortinia AI Chat Growth. Run the math.
WhatsApp Business APISeparate provider, 360dialog / Twilio per-msg
€150-450 / mo
≈ €300
Out-of-hours outsourcingWeekend + holiday + night coverage from an agency
€550-1,250 / mo
≈ €900
Lost customers (no reply)Estimated revenue loss from 6+ hour response times
€850-1,700 / mo
≈ €1,250
Typical monthly cost of a 2-person chat shift€4,500-7,500
Nortinia AI Chat Growth€59990%+ saved · 24/7 coverage
An AI chat doesn't take weekends off. Your customers do — and they ask questions on Friday night.
Token cost, transparently
Included. Not a surprise.
Every Nortinia AI Chat tier includes an LLM token budget. The AI uses these tokens to process question / answer pairs. The cards below show what you get, what it means in conversations, and what it costs if you go over. A new standard for every Nortinia product.
Starter
5M tokens
≈ 5,000 conversations / mo
Over the limit€0.02 / 1k tokens≈ €0.02 / extra conversation
Typical conversation length~ 1,000 tokens
Growth
50M tokens
≈ 50,000 conversations / mo
Over the limit€0.015 / 1k tokensVolume discount (-25%)
10,000 extra conversations cost€150
Scale
200M tokens
≈ 200,000 conversations / mo
Over the limit€0.012 / 1k tokensVolume discount (-40%)
50,000 extra conversations cost€600
What's a "token"? A token is roughly 4 characters or ¾ of a word. A typical 5-message exchange burns 800-1,200 tokens total. Your budget is optimised by the Nortinia AI Engine: short questions on cheap models, long context on stronger ones. The monthly report itemises exactly what you used.
How it starts
One real knowledge upload. Then it chats for you.
Honest answer: deploying AI Chat isn't a 5-minute script copy-paste. The real value lands when we load your full knowledge base, train it on your brand voice and wire it into your CRM. That's a one-time real investment. From there it chats for you, 24/7.
1
Knowledge base ingestion
Website, PDFs, Zendesk export, Notion, past support chats. Any source. We index it, vectorise it and build the RAG layer — that's what guarantees it won't hallucinate.
3-7 days
2
Brand-voice training + sample approval
We generate 50-100 sample answers from your existing content. You approve, tweak, refine. The voice model learns from them. From there it writes that way — consistently across every channel.
1-2 weeks
3
Channel integration + go-live
Widget on your website, webhook to WhatsApp / Messenger / LinkedIn, mail-forward for email-as-chat. Two-way CRM sync (HubSpot / Pipedrive). Weekly report wired up.
1 week
And you don't do it alone. The Nortinia consulting team walks you through end-to-end — from organising your knowledge base to your first live channel. Your one job: get in touch.
If your competitor switches to chat AI and you don't — your customers chat in their window, not yours.
The last generational shift in B2C customer service was live chat — around 2014, with Intercom. Whoever fell behind then is still replying by email six hours later. The next shift is AI chat. It's happening right now. Whoever starts 12 months from now will be behind — and their customers will have already gotten used to the competitor's chat window.
— Norbert Kubinyi, founder
Your customer already expects an instant reply.
A Gen-Z buyer won't wait for an email. If they don't get a reply in 90 seconds, they move to the next search result. AI chat answers in 0.8 seconds.
"Office hours" is dead.
38% of prospects try to buy after 7pm, on weekends, on holidays. You're asleep. Your competitor's AI chat is selling them right now.
"Reply five messages behind" is becoming standard.
Whoever still runs a dedicated chat shift will be 5x more expensive than the market in two years. And slower — because customers are multi-channel, AI Chat is multi-channel, humans aren't.
Separate webchat shift + WhatsApp agent + email support = €4,500-7,500 / mo. Skipping it → your competitor sits in front of your customer 24/7.
Simple pricing — text-only
One chat's price. One shift's output. Tokens included.
In EUR, monthly. 20% off with annual commitment. The monthly fee covers the platform AND the included LLM token budget. Onboarding (knowledge ingestion, brand-voice training, CRM integration) is a one-time consulting project with the Nortinia team. Voice calls are NOT in this product — for that, go to AI Assistant.
Both products run on the same Nortinia AI Engine. The difference is one capability: voice. The AI Assistant handles voice calls (Telnyx / VoIP). AI Chat works in text only — that's exactly why its deploy is faster and its monthly price is lower. Read across and choose the one that matches your traffic.
AI voice calls (outbound + inbound)No — text-only focusYes, Telnyx / VoIP
Voice calendar bookingNo — chat onlyYes
Monthly price (Growth tier)€599€899
Frequently asked
What every CEO asks first.
No marketing fluff in the answers. Concrete numbers, concrete timelines, concrete answers.
What does it cost to start?
Two line items: (1) the monthly platform fee (Growth is €599, which already includes 50M tokens / mo); (2) a one-time onboarding consulting engagement with the Nortinia team — organising your knowledge base, training the brand voice, integrating channels. We quote the onboarding project specifically based on the complexity of your sources. You pay it once. After that it's just the monthly fee.
How long until it's live?
It depends on the size of your knowledge base. A simple B2B SaaS-style FAQ / 100-200 page site is typically 2-3 weeks to live. A complex e-commerce catalogue (10,000+ products) or a regulated financial service is 4-8 weeks. The "go live in 5 minutes" promise would be a lie — if you don't train on your own sources, all you get is a generic ChatGPT that will happily answer about your competitor too.
Why is it cheaper than the AI Assistant?
One reason: no voice module. The Telnyx / VoIP / sub-second voice latency infrastructure is the most expensive layer of the stack. If you only need text, you don't pay for it. Everything else (RAG, brand voice, multi-language, live handoff, CRM, GDPR) is exactly the same.
Does it really not hallucinate?
It only answers from what's in your knowledge base — that's RAG (Retrieval-Augmented Generation). If it can't find the answer, it says: "I don't have an accurate answer for that — please hold, I'm passing you to a human teammate", and hands off. It never invents product data or prices. It can cite the source for every answer if you ask.
What if I go over the token budget?
It doesn't stop. It automatically switches to overage billing: Starter €0.02 / 1k tokens, Growth €0.015 / 1k tokens, Scale €0.012 / 1k tokens. A typical 5-message conversation ≈ 1,000 tokens. The monthly report shows exactly what you spent, and we warn you when you hit 80% of the budget.
Is it GDPR-compliant?
Yes. Every conversation logged server-side, optional PII masking. Data stays in the EU (Hungarian Kubernetes cluster). DSAR and erasure APIs built in. Users can request or delete their conversations any time. We're cookie-free.
Can it really write in my brand voice?
Yes. During onboarding we ingest your existing content (website, newsletters, past support chats). The AI learns from it and writes specifically in your tone — technical, casual, high-cadence, whatever your style is. You approve one sample conversation before it goes live.
Which channels does it work on?
Webchat (script tag anywhere), WhatsApp Business (360dialog / Twilio), Messenger, LinkedIn DM, email-as-chat (mailbox → conversation thread), Telegram, Viber, Instagram DM, X DM, Slack, Discord, SMS (Twilio), in-app SDK, Microsoft Teams. One admin workspace, one conversation thread.
What if I want to stop later?
All your data (conversation history, lead data, knowledge base export) is exportable with one click in JSON or CSV. We disconnect channels immediately. No penalty, no termination fee. You own the data, always.
Can I upgrade to AI Assistant later if I need voice?
Yes, one click. The knowledge base, brand voice, CRM integration are all portable — we just bolt on the voice module. AI Assistant includes every chat feature plus voice calls and voice calendar booking.
Let's talk
Tell us the knowledge. Live in a week.
We reply within one business day. Demo, quote, integration scoping — we'll handle all of it. English or Hungarian.