Automated financial controls

Important results from the project

The Health and Care Administration in Lund Municipality receives, like many other public authorities, lots of invoices, time reports and related decisions about different types of support and needs that the municipality´s citizens need. An increase needs from residents in the form of support, increases the pressure on the administration to handle the influx of documents, control and later pay to external assistance companies. Use AI and good access to data and information as the basis and possibilities for good control

Expected long term effects

LUND Vård- och omsorgsförvaltningen introduced AI components, specifically Machine Learning that can identify errors in data, time reports both faster and with greater precision than is allowed today. Results are a reduced administration per user and better and deeper analysis of invoices, time reports and decisions, therefore find errors that are not found today due to lack of time or absence of analytical instrument.

Approach and implementation

1.Create data pipeline to get new insights of data from Procapita, Raindance data sources 2.Speedup invoice processing 3. Analytics solution for joined economic data (Raindance) and Health care data (Procapita) to get KPI´s and an NLP algorithm to extract timesheet data to automatically compare approved hours for care service and invoices