AI assistant for enterprise data in natural language
Nortinia AI Assistant brings enterprise ERP, finance, sales pipeline and warehouse inventory into a single chat surface — in plain Hungarian (or English). A multi-agent engine decides which modules to query per question, runs under role-based access with a full audit trail, and logs every conversation in a GDPR-compliant way. Spin it up in SaaS in days, or self-host on-premise inside your own environment.
What it does
How people use it
Small and mid-sized business
A small-business owner wants to see in one question which product sold best, how many new customers arrived yesterday and how many invoices are open. Nortinia AI Assistant answers across all three modules in a single chat, and if the owner wants to drill down, the breakdown comes in the same conversation — as an Excel export or an embedded dashboard.
Multi-entity group
Group management of three companies wants consolidated KPIs in a single question while lower-level managers only see their own entity. Nortinia AI Assistant segments data per entity, filters by role and aggregates at group level only where the asker is authorised to see it.
Sustainable ERP compliance
The compliance team wants an on-demand snapshot of NAV sync, ESEF reporting and internal control rules. The AI assistant walks across the connected ERP modules, flags missing items and routes control discrepancies directly to the responsible colleague.
Quick finance lookup
The finance lead wants a 30-second answer to what came in today, how much receivable is >60 days overdue, and whether anything is due today. The AI assistant returns this in one chat with links to the source invoices — bypassing multi-level menus and manual Excel filtering.
Frequently asked
Which LLM powers Nortinia AI Assistant?
Configurable. By default the SaaS version uses an OpenAI GPT-4 class model, but Claude (Anthropic), Mistral and Llama-family models are also supported. For on-premise deployments you can pick an open-source model (Llama 3, Mistral) running on your own GPUs so no prompt ever leaves your environment.
Does it learn from our own enterprise data?
The assistant accesses enterprise data live at query time (live lookup), not as base-model training — so changes are reflected instantly in answers and never end up in model weights. Company documents and knowledge bases are reachable via a vector index (RAG), but raw data never leaves the environment, not even in SaaS.
Which languages does it support?
Primarily optimised for Hungarian and English, but 20+ more languages (German, Romanian, Slovak, French, Spanish etc.) work via the supported LLMs. The role-based access log and audit trail are always written in the assistant’s primary language, while answers are returned in the language of the original question.
What permission levels can be configured?
Role-based access (RBAC) can be set per module, per field and per row (record): e.g. a sales rep sees only their pipeline, finance only its division, the CEO everything — with every decision logged. Permissions can integrate with your existing SSO (Microsoft Entra, Google Workspace, Okta).
More Nortinia AI Asszisztens solutions
Interested? Let us chat for 15 minutes.
Send a quote request and hear back within 24 hours. No commitment, no pressure.