The co-creation initiative with Regione del Veneto’s ICT Directorate evolved a passive document archive into an intelligent, strategic asset. This systemic shift redefined invoice control, payment processing, and project monitoring. The result is clarity within a complex operational flow, providing tools that ensure payment accountability within defined timelines.

A complex public mandate - scenario

The ICT & Agenda Digitale Directorate for Regione del Veneto drives an extensive digital transformation program. Its responsibility is dual: managing the information systems for the administration's 3,000 employees while simultaneously guiding the digital transition for a highly complex territory. This territory includes nine healthcare authorities, some with up to 9,000 employees, and 560 municipalities, ranging from metropolitan centers like Venice to small towns.

This forward-thinking vision faced an operational reality common in large organizations: a growing volume of documents, contracts, and reports, distributed across fragmented silos and difficult to retrieve. An extensive, yet inactive, knowledge base.

A race against time - challenge

The challenge extended beyond mere archiving. Failure to process an invoice within 30 days directly impacted directors' key objectives, creating a high-pressure environment.

The existing process contained specific systemic vulnerabilities:

• Handling variable input and data extraction: Incoming documentation arrived in numerous formats, both physical and digital. Reading, classifying, and precisely extracting structured data - such as an IBAN - was a complex, manual, and time-intensive task.
• Locating supporting documentation for compliance: Upon invoice arrival, personnel had to manually locate all associated documentation and verify a long list of administrative and technical requirements, sometimes involving up to 300 checks per single invoice.
• Reducing approval uncertainty: Validating the information and assessing payability required managing an extensive control checklist. Each check corresponded to data found across various documents, requiring coherence, correctness, and validity assessment.
• Eliminating process fragmentation and dead time: Document retrieval demanded time and constant contact with colleagues, who were often remote. This resulted in significant 'dead time', where the 30-day clock advanced without real progress.
• Achieving model scalability: This fragmented model was replicated across 30–40 regional directorates, creating systemic inefficiencies that were difficult to monitor. Furthermore, changing existing processes or supporting IT systems was resource-intensive and complex, as was changing user work habits.

The objective - expected outcome

The goal was to transition from this laborious, high-risk process by introducing a tool that could:

• Reduce the team effort required to validate and approve an invoice via a simplified UI.
• Accelerate payment times by simplifying the operator’s work and supporting control activities.
• Facilitate granular control involving dozens of documents and checks, making every piece of information simple to manage and verify.
• Avoid changes to existing information systems.
• Avoid changes to existing processes, ensuring immediate personnel adoption.
• Provide full monitoring, offering visibility into the progress and timing of every phase.

KLens: knowledge activation through intelligence - the solution

The project was executed through co-creation. Working directly with the Region Veneto team, KLens was developed via continuous iteration, translating technical expertise into an intelligent user experience that adhered to existing processes.

KLens is an intelligent system that transforms static documents into a dynamic asset. It is structured around four core elements:

• Automated classification: AI, contextualized to the Public Administration domain, analyzes file content for correct classification, even with inconsistent naming or multiple versions.
• Document summarization: The system generates concise, clear summaries, allowing immediate comprehension of even the most complex documents.
• Natural language query: This moves beyond simple “search.” Users can pose complex business questions (“Which PNRR projects exceeded the Q3 budget?”) and receive precise, evidence-based answers.
• Complete integration with existing systems: KLens connects to the administration’s current systems, enriching them with structured data and analyses.

Efficiency, control, and strategic transformation - the outcome

The project has resulted in strategic outcomes with significant organizational impact: From passive archive to active system The project and invoice document archive became a proactive source of knowledge that supports daily decision-making. Users can now query documents in real-time. The system responds by locating specific files, extracting relevant data, and indicating the necessary references to the operator, which saves time and optimizes performance. Previously hidden information and answers are now readily accessible. Administrative simplification, efficiency, and process transparency Automatic classification and analysis drastically reduce time and errors, ensuring full information traceability. The system automatically checks over 300 conditions per invoice and presents the control results to the user via a traffic-light system: green signifies all controls passed and the invoice is payable; red indicates a need for verification. In the latter case, the system highlights the failed controls and inconsistencies, including the exact sources, supporting the user without requiring changes to their work method.

Enhanced governance: clarity on process bottlenecks For the first time, executives have a clear view of the payment process and its blockages. Knowledge of 'error' or 'non-compliance' cases was difficult to track without laborious manual classification. Now, every control at every stage is tracked by design, providing clear data on where the process can be refined. “We often asked, ‘Where did the process stop?’ No one tracked that before. Today, we have a tool to see where and for how long. This is a crucial element for truly controlling what needs improvement. A previously latent documentary asset has been reactivated, enabling more informed governance and freeing up time for higher-value activities.

Lessons learned: a model for pragmatic AI

The project established an operational model for introducing AI in complex environments.

Co-designing the system to drive innovation The customer holds the knowledge of needs, processes, and necessary actions. The design and IT partner contributes experience from other projects, novel ideas, strong competencies, and the capacity to envision the future. The final result emerged from this joint working model.

AI as digitalization without process change Concrete value emerges when technology integrates with existing work and processes, avoiding structural change. AI’s ability to simplify the customer interface and speak a business language, rather than a screen-based one, facilitates a change in workflow without imposing unnecessary changes to core processes and systems. This significantly simplifies the typical IT change management process, making adoption virtually immediate. No one was asked to change their work method or learn new tools.

Adapting to legacy systems Integration with existing infrastructure ensures rapid adoption and protects prior investments. Existing data and systems are a hidden value that can now be seamlessly integrated into AI-driven workflows. Furthermore, unstructured data can be extracted and structured for decisions requiring certainty and verified facts, such as invoice payment.

Engaging users early Active participation from the team was a decisive factor, thanks to early involvement and expectation management. Showing users how their work could change, and involving them to understand business needs, helped manage the common 'fear' of new technologies, test system limits, and identify optimal application areas.

Key insights for your organization

The experience with the Region Veneto provides a replicable roadmap:

• Secure innovative partners: Durable solutions arise from shared work between domain and business expertise, and technical competencies spanning Design, Customer Experience, and AI.
• Activate value from existing data: Archives contain hidden value. When extracted and made available through business-focused queries, this knowledge permeates the organization, empowering teams and dismantling silos.
• Digital transformation is simplified: Many projects failed due to change management complexity. The IT interface is now far more business- and people-oriented.
• Process bottleneck resolution is accessible: Digitalizing with AI enables the measurement of process effectiveness, clarifies bottlenecks, and then allows for resolution. Starting with clear, measurable projects, such as invoice payments, builds momentum for broader transformation.
• Involve and value users’ business knowledge: Experimentation is key to understanding technology's value and successful application. The people who understand the business and its operational necessities are the core of a successful project. Both roles must be present and engaged during key phases.