Picture the scene; invoices are being received into several different email inboxes across the accounts team. A steady stream of invoices are also still arriving in through the post. Suppliers are calling chasing payments. And nobody can exactly pin point where any specific invoice actually is! Sound familiar? Well you’re not alone.
In fact only 24% of Australian businesses have actually implemented an automated accounts payable system.
This is despite market intelligence by Aberdeen Group that shows that the cost of processing an invoice is $11 per invoice, and the same Aberdeen group report found that processing costs halved with automation.
Interesting, so just where do I start?
Well stripping everything right back to the basics. Every accounts payable automation solution has 3 key fundamental pillars.
- Capture – the process of capturing the key data from your invoices.
Every invoice that arrives into your business needs to be captured to have the key data extracted from the invoice. The technology that is universally used to complete this task is called optical character recognition software (OCR). But not all OCR software is made equal.
Some of the older technologies on the market use what is called ‘zonal OCR’. This is where all invoices that are received into the business are initially trained by showing the software the exact location on the invoice where it can find each key field of data. By key fields we’re referring to the important details that you are hoping to capture such as: client name, date, ABN, Invoice number, P.O Number, sub-total, GST & total.
The challenge with zonal OCR is twofold. Firstly, the initial set up and training of all of your supplier’s invoices to create an invoice template is a time consuming (and potentially costly) affair. Secondly, each of these invoice templates need to be managed by your internal team ongoing. If a supplier changes the format of their invoice (and you won’t be warned about this change) you’ll need to retrain your capture software to learn this new invoice format.
This isn’t a big challenge if you only receive a small number of invoices per month. But if you receive thousands, this will soon drive you crazy.
The better option is to choose an OCR capture software that reads and recognises the metadata it is looking to extract from the entire page regardless of where it is located. The more advanced versions of these OCR capture software solutions also have machine learning. So that if the software picks up some incorrect data, it can be trained and will learn for the next time it receives a similar invoice from the same supplier.
The cost to implement and maintain these more advanced OCR capture solutions are significantly reduced as invoice templates are not required to be created. And new suppliers can be onboarded to your business seamlessly without the need for training.
- Workflow – applying your own specific business rules
So now you have all the required information captured off the invoice, it’s time to validate this information to ensure it is correct. In an ideal world you would have a robust P.O. generation system implemented internally. And if so your solution would simply need to match the P.O. against the invoice received, and push it for payment. But life isn’t always that simple?
Here is where we take your everyday real world manually steps, and map them into an automated workflow for invoice approving. For example, you may have a simple internal rule that says all invoices under $500 + GST get approved by the finance manager, and all invoices above $500 + GST need to be routed to the CFO for their approval.
The choices and complexity of your internal rules based workflow approval (e.g. specific suppliers getting routed to specific department heads for approval or multiple approval levels etc) is pretty much limitless with a powerful workflow solution.
It’s important too that you ensure your solution has mobile functionality so that approvers can receive invoices to approve while working remotely, and so this doesn’t hold up your approval process. This could cause you to miss out on lucrative discounts for paying invoices on time.
Your workflow solution should also allow an approver to push invoices to another person to approve if they are unsure if the invoice is legitimate. Once approved by the other party the invoice will come back into the workflow and continue on its specified route.
Lastly ensure your workflow solution allows an approver to reject an invoice and give a reason why (i.e. no P.O. number provided on the invoice or incorrect invoice amount etc). This invoice can be pushed back to the supplier to amend and resubmit the correct invoice or add the additional relevant details.
- Integration – connecting with your finance package
Once you’ve captured all the data and routed it through your customised approval process, it’s time to push the information into your finance system or ERP for payment. This is where the integration between your AP automation solution and your finance system is important.
The best and most secure form of integration is a custom built open API integration. And this is what you should try to implement where possible.
This allows you to push and pull information such as P.O. details, supplier tables, GL codes etc for matching invoices against. And it will process invoices in real time. But this is not always possible. Most new cloud based finance and ERP systems will have open APIs which allow you to achieve this, but some of the older legacy on-premise accounting packages may not be so accommodating.
However you still have options. Flat file or CSV imports allow you to import the captured and approved data, and push it into your accounting system in the correct format at scheduled periods through the day.
While not being completely automated this is still a very useful workaround if your older system proves to be a little challenging.
To learn more about implementing a successful accounts payable automation solution or to talk through some of the challenges that your business is experiencing, feel free to reach out to Bryan Murphy (firstname.lastname@example.org).