Fewer repeat tickets
The chatbot answers from your Znuny FAQs in the customer portal and on your website. If a person is needed, it opens the ticket with the details.
A chatbot answers from your Znuny FAQs in the customer portal and on your website. New tickets land in the right queue. Agents get a summary and a reply draft. Your ticket data stays on your servers.

Your host
Constanze Krishnaratne
CEO
Your agents keep working in Znuny. OTAI takes over the repetitive part.
The chatbot answers from your Znuny FAQs in the customer portal and on your website. If a person is needed, it opens the ticket with the details.
New tickets land in the right queue, with priority and your dynamic fields already set.
Agents see a short summary at the top of the ticket and start from a reply draft. A person still checks and sends every reply.
OTAI runs on your own servers. Ticket text does not leave your network.
Example monthly savings for Znuny teams, net of the OTAI subscription. Open the full calculation to see how the numbers add up.
Small team · 50 tickets/day · 3 Queues
≈ €2,700–€3,000
estimated monthly savings
Mid-size · 200 tickets/day · 10 Queues
≈ €12,600–€13,300
estimated monthly savings
High volume · 750 tickets/day · 40 Queues
≈ €57,000–€58,500
estimated monthly savings
Four hidden costs — easy to miss, expensive when they stack up.
Agents read and sort before specialist work starts.
Wrong Queue → tickets ping-pong between teams.
Hours before the right team sees the ticket.
Reading threads, copying fields, re-prioritizing.
Concrete situations — not just spreadsheet math.
Lisa
Customer
Hours waiting for a reply
A password reset lands in the general inbox — Lisa waits most of the day.
4+ hoursAnswer in minutes via chatbot
The public chatbot resolves common requests instantly — no ticket needed.
~2 minMarco
L1 agent
Sorts every ticket by hand
Sorts 80+ tickets/day by hand — specialist teams still wait hours.
45 min avg. to right teamRouting on arrival
Classification sets Queue and priority before Marco even opens the ticket.
InstantSandra
Specialist
Reads long threads herself
Scrolls through 15 emails just to understand the current status.
8–12 min/ticketSummary on the ticket
Problem, status, and next steps are written back on the ticket.
~30 secMany routing targets, minimal tickets — Full On-Prem is not yet worth it.
5 tickets/day · 30 Queues
Estimated monthly savings
≈ −€250
per month (net of subscription range)
Where the savings come from
Queue classification
Direct routing to the right team without L1 triage hops. Main lever for fast handling.
Faster first handling / shorter cycle
Priority & other attributes
Correct priority cuts sorting effort.
Faster first handling / shorter cycle
Ticket summaries
Saves agent reading time, with minor impact on overall cycle time.
Agent time (not cycle time)
Answer drafts
SoonMinimal effect at low ticket volumes.
Not shipped yet — illustrative only
Customer chatbot
Premium · SoonPremium pack — stops ticket creation. Needs higher volume.
Not shipped yet — illustrative only
Subscription costs outweigh savings at 5 tickets/day. The math flips at higher volume.
A lean support team handling a steady daily ticket stream.
50 tickets/day · 3 Queues
Estimated monthly savings
≈ €2,700–€3,000
per month (net of subscription range)
Where the savings come from
Queue classification
No L1→L2 hopping: The right team starts on intake. Primary savings driver.
Faster first handling / shorter cycle
Priority & other attributes
Proper fields reduce mis-prioritization and re-sorting.
Faster first handling / shorter cycle
Ticket summaries
Cuts reading time per ticket for agents.
Agent time (not cycle time)
Answer drafts
SoonReview a draft (~30–90s) instead of typing — useful on ~25% of tickets.
Not shipped yet — illustrative only
Customer chatbot
Premium · SoonPremium pack only: Prevents simple requests from becoming tickets.
Not shipped yet — illustrative only
Tickets land with the right team hours earlier — SLAs are easily met.
Growing helpdesk routing across multiple specialized teams.
200 tickets/day · 10 Queues
Estimated monthly savings
≈ €12,600–€13,300
per month (net of subscription range)
Where the savings come from
Queue classification
Straight to the right team without manual triage — dominant savings lever.
Faster first handling / shorter cycle
Priority & other attributes
Proper priority from intake shortens time to meaningful work.
Faster first handling / shorter cycle
Ticket summaries
Saves agent reading time across all tickets.
Agent time (not cycle time)
Answer drafts
SoonLess typing on routine queries (coming soon).
Not shipped yet — illustrative only
Customer chatbot
Premium · SoonPremium: Deflects ticket creation. Not included in Standard.
Not shipped yet — illustrative only
Resolved up to a day faster — Queue classification drives most savings.
Heavy ticket load across many targets with complex initial routing.
750 tickets/day · 40 Queues
Estimated monthly savings
≈ €57,000–€58,500
per month (net of subscription range)
Where the savings come from
Queue classification
Eliminates L1 bottlenecks with immediate team assignment.
Faster first handling / shorter cycle
Priority & other attributes
Proper priority shortens wait time before work begins.
Faster first handling / shorter cycle
Ticket summaries
Saves significant agent reading hours.
Agent time (not cycle time)
Answer drafts
SoonReview drafts instead of typing — applies to ~25% of tickets (coming soon).
Not shipped yet — illustrative only
Customer chatbot
Premium · SoonPremium: Deflects common requests so tickets are never created.
Fewer tickets created
Direct team routing on intake plus Premium chatbot ticket deflection.
Queue classification drives cycle-time savings. Other features save agent time or deflect tickets.
Summarizes long email threads instantly. Reduces agent reading time.
From Entry · available now
Review and send drafts instead of typing. Speeds up responses.
Depends on pack
Answers common questions directly to deflect ticket creation.
Premium pack only
Illustrative examples for European support teams. Subscription figures are ranges (Full On-Prem from ~€500/month depending on agents + pack) — not a fixed price list. Manual costs cover triage, wait time, and misrouting. Chatbot and answer drafts are coming soon.
Every feature, what it gives you, and how it looks.
Signed-in customers ask in the Znuny customer portal. The chatbot answers from your Znuny FAQs. If a person is needed, it creates the ticket.
Your benefit: Fewer repeat tickets, and the tickets that still come in already have the details.
Visitors who are not logged in ask on your website. The chatbot answers from your public Znuny FAQs.
Your benefit: Simple questions get answered before they become tickets.
Agents open AI Chat in Znuny and ask about internal FAQs and procedures.
Your benefit: Agents find the answer without searching the FAQ module.
Install two Znuny packages and start OTAI with Docker Compose. The model is already included. OTAI Studio opens from the Znuny menu.
Your benefit: You go live in days, and your admins never leave Znuny to sign in.
OTAI reads each new ticket and moves it to the right queue. It uses your queue names and descriptions. No ticket history is needed.
Your benefit: Tickets reach the right team without anyone sorting them by hand.
OTAI sets the priority from what the customer writes.
Your benefit: Urgent tickets stand out at once.
OTAI fills Znuny dynamic fields, such as department or product, with values from your list.
Your benefit: Reports and filters work with clean values instead of empty fields.
A short summary appears at the top of the ticket in Znuny.
Your benefit: Agents know what a long ticket is about in seconds.
OTAI prepares a reply to the last customer message, using your FAQs. The agent reads it, edits it, and sends it.
Your benefit: Faster answers, and an agent still checks every reply. OTAI sends nothing on its own.
Znuny offers the free Znuny-LLM add-on for agents. Here is where OTAI goes further.
Znuny-LLM: Suggests queue, type, service, and SLA in a sidebar for the agent to check.
Open Ticket AI: Moves the ticket to the right queue and sets priority and dynamic fields when it arrives.
Znuny-LLM: Built for agents and admins. There is no chatbot for customers.
Open Ticket AI: A chatbot in the customer portal and on your website answers from your Znuny FAQs.
Znuny-LLM: Rewrites text the agent selects and suggests matching templates and FAQs.
Open Ticket AI: Writes a full reply draft to the last customer message, using your FAQs.
Znuny-LLM: You install Ollama or connect another model endpoint yourself.
Open Ticket AI: The model comes with OTAI and is already running.
Znuny-LLM: Znuny 7.3 or newer.
Open Ticket AI: Znuny 6.5, 7.2, and 7.3.
Sources: Znuny-LLM announcement · Znuny-LLM setup
What a typical rollout looks like. We map your Queues in Studio; Runtime classifies new tickets and writes queue, priority, and dynamic fields back to Znuny.
Connect the helpdesk, describe your setup in Studio, classify tickets zero-shot, run Full On-Prem.
01
Enable Generic Interface and connect OTAI with a web-service token.
02
Use queue names and descriptions from your helpdesk setup — no ticket history.
03
Zero-shot classification for queues, priority, and dynamic fields.
04
Runtime classifies new tickets locally and writes results back to Znuny.
Full On-Prem AI runs inside your network to set queues, groups, priorities, and custom fields securely.
On-prem configuration and control in your environment
New tickets arrive for AI processing
Local classification on your GPU — air-gap ready
Routing, priority, and custom fields updated automatically
No ticket data leaves your network. Everything runs on your infrastructure.
Free Cloud Trial to evaluate. Full On-Prem quotes based on agents and tier.
One month · full features
On request · agent-based software
Questions about using OTAI with Znuny.
No. Full On-Prem classification is driven by your helpdesk setup — not by uploading historical ticket archives to an external service.
What OTAI uses instead:
That architecture fits teams migrating from OTRS-era setups who already treat queue definitions as the authoritative routing catalogue. Ticket content stays inside your network throughout inference.
Read the Setup-based classification guide and the OTOBO/Znuny plugin setup for connector details.
Deploy, configure, and run OTAI with Znuny.
Automate ticket classification on-premise using your existing Queue setup.

Your host
Constanze Krishnaratne
CEO