Flicking the switch
So how is a ten-fold boost in productivity possible? It’s simply a question of automating as many aspects of the AP function as possible. The real costs involved in AP are the man-hours involved in manual tasks, including: finding purchase orders and good received notes; checking and matching these; manual data entry into core business systems such as ERP; manual validation; processing complex invoices which may involve checking against service level agreements, and more.
Even more costs are incurred in tasks such as long-term filing and storage of documents, staff turnover and teaching systems to new staff. If these tasks can be automated, then accounting staff can be redeployed in more strategic roles such as data analysis, and not be involved in time-consuming paper trails.
The ultimate aim is sometimes referred to as “lights-out accounting [http://www.dvtl.co.uk/index.pl?art=14]” – in other words, switching off the lights while the work proceeds automatically. And although it’s not yet possible to flick off the light switch all the time, the latest-generation invoice data capture and processing solutions really do deliver gains in productivity of up to 1000%, helping to turn AP into a profit centre, not a cost centre.
Let’s take a close look at exactly how this can be achieved. We will use a real-life example of the AP department of a large business in the licensed trade, which undertook a phased migration from paper-based manual processing to fully automated invoice processing.
Raising a glass to automation
DrinkCo is a UK-based business that owns several well-known pub chains. As one of the largest pub operators in the UK, the company processes 600,000 invoices per year. As a Peoplesoft (now Oracle) Accounts user, DrinkCo wanted to deploy an invoice capture [http://www.dvtl.co.uk/index.pl?art=31] solution to bring key financial data directly into the Peoplesoft solution and accelerate processing.
With ROI being a key factor in the decision to deploy, DrinkCo undertook a proof-of-concept exercise with Digital Vision [http://www.dvtl.co.uk] that demonstrated that ROI was possible within months, rather than years.
The proof-of-concept was based around an invoice capture solution from Paradatec, integrated with data extracts from the Peoplesoft system so that invoices can be matched with purchase orders and goods received notes. This demonstrated that high levels of accuracy were possible with data capture on the majority of invoices received.
Before optimisation of the capture process, data capture on general invoices was 73%. By focusing on invoices from large-volume suppliers, this was raised close to 100%. Furthermore, the process proved that the artificial intelligence system performed regardless of invoice layout, and that that the invoice capture engine’s AI rules could be adapted to handle different invoice layouts, or those from new suppliers.
On completing the proof of concept, Digital Vision gave a detailed specification to DrinkCo to integrate the Paradatec capture solution, SpeedKey validation and Contempus invoice processing with Peoplesoft Accounts, with invoice images stored in an IBM Content Manager repository.
Going live
When the green light was given to proceed with the automation project, additional pressures reared their head. One of DrinkCo’s offices was to be closed and the work transferred to its Midlands HQ. This also meant redundancies amongst AP staff – which applied additional pressure to get the automated invoice processing system live as quickly as possible. As staff was leaving, a backlog of invoices was building up.
The solution was deployed in just 8 weeks. The backlog of 4500 invoices built up while staff were leaving was dealt with by just four staff, who validated and corrected the backlog in less than three days. Under its previous system, clearing a similar-sized backlog would have taken 12 staff over 10 days – clearly demonstrating the savings in time and manpower possible with the automated solution, and immediately delivering ROI. The AP team continues with four validation staff, and the project achieved full return on the original investment in just 8 months.