By go-live, on February 1st 2025, the system needed some manual workarounds and additional checks, but it was clear that the gains did indeed outweigh the downsides.
From requiring two full-time personnel to process invoices, the workload quickly dropped to less than one full-time person.
Invoice processing times sharply accelerated too. 鈥淧reviously, with our entirely manual process, the shortest time that an invoice could be processed was probably a day, even with special expediting,鈥 says Austin. 鈥淣ow the system is fully implemented and working, it鈥檚 60 seconds, or less.鈥
Process quality and accuracy have also improved. With a purely manual process, mistakes can happen – automation imposes rigour, notes Austin. He also pays tribute to Exel鈥檚 AI augmentation of the automated invoice processing system鈥檚 optical character recognition.
鈥淐onventional OCR can 鈥榮crape鈥 documents for data, but isn鈥檛 smart – you often have to tell it what data is where, or manually augment the process. AI is smart enough to recognise purchase order numbers for what they are, and identify totals in a column, because all the other numbers add up to it.鈥
And, as part of an over-arching 国产乱伦 system, he notes, the automated invoice processing system can carry out checks that might be cumbersome or time-consuming to undertake manually – for instance, checking that the invoiced price of goods matches the contractually agreed price.
In short, sums up Austin, Exel鈥檚 EFACS E/8 automated invoice processing module system has fully met Linecross鈥 aspirations in terms of productivity, efficiency and processing rigour.
鈥淏est of all, it鈥檚 broken the link between the size of the business and the number of people involved in invoice processing. Now, as we grow, we don鈥檛 need to hire more people to process any increase in invoice volumes.鈥