Integration

AI-Powered Routing for Government Services

Automate citizen request routing on-prem. GDPR-compliant, works with your existing ticketing system.

  • Automatic classification of citizen requests by department and service tone
  • On-premise deployment — no citizen data leaves your infrastructure
  • Works with Zammad, OTOBO, and other systems
  • GDPR-compliant, audit-ready logging
Integration Flow

How It Works

Step 1

Define Your Departments

Provide your department names, service categories, and routing rules as a QueueSpec — no real citizen data needed.

Step 2

We Train Your Model

Our EU-hosted training service generates synthetic data from your QueueSpec and trains a custom adapter. Typically same-day.

Step 3

Deploy On-Prem

Install OTAI Runtime Core in your government data center via Docker Compose. All inference runs locally.

Step 4

Go Live & Monitor

Incoming citizen requests are automatically classified and routed to the correct department. Monitor accuracy and retrain as regulations change.

Custom Model Training

Custom Models for Queue Classification

OTAI Studio configures your automation on-prem. Optional cloud training generates synthetic data and fine-tunes your model — then the signed artifact is downloaded to run inference locally. Fully on-prem training is available with a GPU. Other features like ticket summarization run on-prem without cloud training.

Cloud-Assisted Training Path

Cloud (EU) — Optional Training Services
Synthetic Data Generator

Creates training examples from your configuration metadata

Training Service

Supervised fine-tuning on cloud GPU hardware

Signed Model Artifact

Ready for one-time download to your infrastructure

Configuration Metadata

Queue names and descriptions only — no ticket data

Model Download
Your Infrastructure (On-Prem)
OTAI Studio

Configuration and control plane — always on your infrastructure

Your Ticket System

New tickets arrive for AI processing

OTAI Runtime + Connector

On-prem inference — CPU for routing, GPU optional for on-prem training

Tickets Enriched by AI

Classification, priority, and fields updated automatically

Privacy by design: No ticket data ever leaves your network. Only configuration metadata is sent for optional cloud training. Your custom model is downloaded once and inference runs entirely within your own secure infrastructure.

AI Capabilities

Fields AI Can Improve

OTAI doesn't just route to queues — it can classify and populate multiple system fields automatically.

Department

Department Routing

Route citizen requests to the correct government department automatically

Service Category

Service Classification

Classify requests by service tone (Permits, Registrations, Complaints, etc.)

Priority

Priority Classification

Set request priority based on urgency and regulatory deadlines

Sub-Department

Sub-Routing

Route to specialized sub-departments within main departments

Requirements

Any supported ticketing system (Zammad, OTOBO, or Znuny)
Docker / Docker Compose for OTAI Runtime
2 CPU cores, 4 GB RAM minimum

Government Service Templates

Pre-built QueueSpec templates to get started quickly with your system.

1

Citizen Services

Permits, Registrations, Complaints, Information Requests

2

Infrastructure

Road Maintenance, Public Transport, Utilities, Construction Permits

3

Internal Administration

IT Support, Facility Management, HR Requests, Procurement

ROI Estimate

See Your Potential Savings

Example monthly savings from replacing manual ticket routing with OTAI. See the full page for the assumptions behind these figures.

Quick ROI Estimate

See how much OTAI can save you

Small team (50 tickets/day)€3,500
Mid-size (200 tickets/day)€13,600
High volume (750 tickets/day)€54,500

estimated monthly savings vs. manual routing

See full ROI examples
Pricing

Editions

Pro: support & updates included. Enterprise: cloud-trained, no local GPU. Sovereign: fully air-gapped on your hardware. All tiers also available as a one-time perpetual license (no subscription required) — on request for the public sector.

Community

Evaluate & test

Free
  • Queue Prediction
  • Up to 5 queues / values per attribute
  • Open source (LGPL), self-hosted
  • Community support
Available Now
MOST POPULAR

Pro

Everything to run in production

500 € /month
  • Queue & Priority Prediction
  • Up to 50 queues / values per attribute
  • High training limit
  • High-quality training, modern fine-tuned models
  • Support & updates included
View Product

Sovereign

Fully air-gapped, on-premise training

On Request
  • Everything in Enterprise
  • Fully air-gapped — nothing leaves, not even metadata
  • On-premise training on your own GPU
  • Requires a GPU with at least 24 GB VRAM
  • Maximum data sovereignty & compliance
View Product

Enterprise

CONTACT SALES

Scale, SLA & largest models

  • Everything in Pro
  • Unlimited queues / values per attribute
  • Largest / highest-quality models
  • Dedicated SLA support
  • Cloud-trained in the EU — no GPU required
  • On Request
    View Product

    Frequently Asked Questions

    Common questions about using OTAI with your ticket system .

    Ready to automate
    your routing?

    Get a custom queue model trained on your queue definitions. Same-day turnaround, on-prem deployment.