Sorting each ticket from scratch
Your team has to check the request, urgency and ownership again for each recurring enquiry.
The goalPre-sorted tickets that your team can review and take on.
AI agents for IT service providers and MSPs
I build AI agents for IT service providers and MSPs in Germany, Austria, and Switzerland. They triage tickets, prepare server diagnostics, and write support reports. A technician approves actions on customer systems. Every action is logged.
Does this sound familiar? These recurring tasks may be worth examining before introducing an agent.
Your team has to check the request, urgency and ownership again for each recurring enquiry.
The goalPre-sorted tickets that your team can review and take on.
Before working on the fix, technicians need to collect information and server logs.
The goalPrepared diagnostic information to support the technician’s decision.
The issue is resolved, but the summary for the ticket and customer still needs writing.
The goalDraft support reports that your team reviews rather than writes from scratch.
When a ticket changes hands, the next technician has to piece together information and previous actions.
The goalClear summaries for the next technician to work from.
Potential time savings mean little if nobody knows who approves actions on customer systems.
The goalLimited permissions, technician approvals and a log of every action.
Without baseline measurements, it is hard to tell whether an agent actually reduces work.
The goalCompare minutes per ticket, first-contact resolution and escalations before and after introduction.
These are possible starting points, not promised savings. The check helps us assess whether an agent fits your tasks.
I started programming at 12 and have been doing it for 14 years. I am an IHK-certified Fachinformatiker, a German vocational qualification in IT.
I held the CTO role at two startups, both spin-offs. My project portfolio documents the software I have built.
At START Hack Vienna 2026, I won a challenge with an agent that logs into servers and writes support reports.
The START Hack project also won the ElevenLabs special award for best use of voice.
The awards recognise a hackathon project, not production deployment at MSP customers. They are not evidence of customer savings. View the Techbold project and source code
In a one-day workshop, we choose ticket types, define permissions, and record what the agent must never do. The workshop is paid. I credit the fee toward implementation if you commission me to build it.
See details→I implement the agent at a fixed price. For two weeks in shadow operation, it only makes suggestions. It then goes into operation with the agreed approval stages. Actions on customer systems need technician approval; every action is logged.
Send an enquiry→After launch, I can operate and expand the agents I built. We measure minutes per ticket, first-contact resolution rate, and escalations to decide what to change next.
Discuss operation→Start hereFour service-desk inputs, 30 minutes, free: Ticket Time Check
ProcessWe agree on tasks, approvals, and measurements before implementation. How introduction works
In the free 30-minute Ticket Time Check, we estimate the technician hours spent on routine tickets. We use four inputs: monthly ticket volume, technician count, minutes per routine ticket, and the share of fixed-price contracts. You can request the call before gathering these figures.
Talk to my AI assistant
My AI assistant asks about your service desk and prepares an enquiry below. You review it before sending. This is not a call with me.
Ready when you are
Please do not share passwords, customer tickets or other confidential information.
Your enquiry
Write a few lines, or let the voice assistant prepare a draft. Nothing is sent automatically.