Initial Situation
A public administration organization in Austria operates a portal through which companies, ranging from sole proprietorships to large corporations, submit their filings digitally. Other departments work with the same data. Externally, the portal appears as a single point of access. Behind the scenes, however, there are separate procedures with their own rules and legal foundations. The knowledge required for this is contained in manuals and FAQs, and to a large extent, resides with individual long-term employees. Those seeking answers rarely find them immediately and turn to the service desk via phone, email, or contact form, often providing only a brief note regarding their actual request. The small first-level team handles the majority of these inquiries themselves. The workload is high.
The strategic question was how to gradually relieve the service desk using generative AI, while maintaining high standards for sovereign data storage, GDPR compliance, and government-grade access control.
Our Solution
AMAI guided the client through a structured consulting process to evaluate the feasibility, operating model, and cost-effectiveness of AI-powered service desk automation, and presented the results as business case ready for decision-making.
Requirements Workshop and Current State Analysis:In a joint workshop with the service desk team, processes, typical requests, and pain points were systematically recorded. This resulted in a structured requirements catalog covering pain points and a target vision, ranging from structured request intake and usable knowledge access for users to transparent feedback of second-level results to the specialized departments.
Technical Target Vision with Phased Sovereignty:AMAI designed a multi-stage expansion model for automation: a source-based knowledge bot for public and internal information as a self-service offering, a support bot for structured request intake and qualification, and, looking ahead, a learning service desk agent based on anonymized historical cases for internal staff. The architecture relies on strictly source-based answers (grounding) and deterministic access control: user identity and permissions are verified before document retrieval, ensuring access protection is not left to model behavior. The operation is modular and can be implemented with European model and hosting alternatives, from managed cloud services based in the EU to self-operated data centers.
Business Case Development:Based on the target vision, AMAI assessed the costs and potential benefits for each expansion stage and developed a transparent cost-benefit and ROI perspective. Conservative benchmark corridors from comparable systems compensate for the metrics (such as ticket volume and processing times) that the client still needs to collect. The result is a roadmap with capacity planning and a phased investment outlook as a sound basis for decision-making.
Results & Business Impact
The business case provides the client with a foundation for their further AI strategy in the service desk. The entry point can remain small: first, relief through self-service, then more complete requests, and only later an internal agent that learns from past cases. The automation can be integrated with the ongoing development of the portal without the need to commission a large-scale project upfront.
The conceptual work makes the requirements for data protection, access control, and operations tangible. The project can subsequently be connected to the platform in stages.
