AI that runs inside processes and cuts costs by 110x

The challenge: AI that's governed, auditable, and delivers measurable results

GTX Service embedded its core service process in WEBCON in 2019. Since then, the system has automatically collected all data. Every ticket, decision, service photo, and customer communication is stored in a structured way.

When the company decided to adopt AI, it already had years of clean process data from WEBCON. What typically slows AI down wasn't an issue here.

• Can we use our existing data to train AI models safely? WEBCON had been collecting data in a structured, controlled way since 2019, so there’s no separate AI infrastructure needed.
• Can we audit every decision AI makes? Every process decision is part of the case history. AI embedded in WEBCON operates within the same audit trail as any other user action.
• Can we measure the real impact of AI on time and cost? The data from before and after deployment lives in the same system. WEBCON gives you a ready baseline, so measurability comes from data continuity, not separate analytics tools.
• Can we keep AI inside our systems and out of shadow tools? AI didn't arrive on a separate project. It got embedded directly into existing WEBCON processes.

„A customer used to wait 4 hours for a response. Today, the average response time is 5 minutes – at a cost of just 10 cents. And 68% of all inquiries are already handled by AI.”

Dariusz Robak, COO, GTX Service

AI embedded directly in the processes

The core principle was clear: AI cannot be a standalone tool. Today, a service technician doesn't switch between apps. The model responds inside the same form where the employee runs the diagnosis. A customer service agent doesn't copy emails into an external tool. AI is built into the request-handling process.

WEBCON platform keeps everything under control. Every decision made by AI is recorded in the process, embedded in the case context, and available for analysis in reports. The organization decides which LLM models run, which processes they power, and what data they can access.

How AI works in daily operations

110x lower cost per service query

More queries coming through multiple channels, with the same number of consultants, created a classic choice: slower response times or more headcount. Neither was acceptable.

An AI agent reads incoming emails, classifies them, and generates responses based on similar historical cases. Today, 68% of service requests are handled automatically. Consultants only review exceptions. Response time dropped from 4 hours to 5 minutes, and the cost per request fell to less than 3 cents. AI agent, works 100 times faster.

Instant verification of service photos

GTX Service trained a neural network using a database of photos from previous service jobs and descriptions that employees had prepared earlier. A technician photographs the fault, and AI checks the photo quality in real time.

In the past, when a human did the review, it took around 15 minutes. If the photo was wrong, the device was often already disassembled and couldn't be photographed again. Now the technician can start the repair immediately.

GTX Service didn't start with AI. They started with process digitization in WEBCON

Digital workspaces and structured data came first. Language models and neural networks came later. AI became the tool that let the business scale without proportional cost growth.

"We didn't build this in a day. And we didn't start with AI. It First came operational excellence, a digitized, repeatable service process. Then the data that process generated. Then the question: can we get more out of this? That's when AI came in. And after that, a whole new business built on the same foundation. Every step came from one before it."

Krystian Dobrzyński
Process Management & Business Intelligence Manager, GTX Service

GTX Service proves it: AI works when the digitized process comes first. 

How WEBCON builds the foundation AI needs

Most AI projects never make it past proof of concept. GTX Service did, because the process foundation was already there.

The platform makes AI auditable and controllable from the very beginning. Every model’s decision leaves a trace in the process. Data gets collected directly in the workflow, which removes silos and manual data handling. And because AI runs in a familiar environment, user adoption goes up.

Start with the foundation. Contact us and see how WEBCON can help you deploy AI safely, measurably, and at scale.