Improving Efficiency in Accounts Payable with Automation
Intelligent operations are at the heart of the enterprise. Today, leading companies are looking to their leaders to be strategic enablers. Digital transformation is less about implementing upgrades to existing platform. It’s about finding new ways to operate the business and using technology in new ways. Ways that improve efficiency, remove barriers to performance, improve operations, and create agility.
Process automation (Robotic Process Automation or RPA) is one tool in the digital toolbox. It promises the benefits of improving quality, creating efficiency, and scaling operations at a cost below that of a human resource. In other words, leveraging technology to create operational efficiency.
How does process automation achieve these benefits? RPA technology is a mature technology that allows a software module (a robot or “bot”) to perform the same tasks as a person at a computer. The bot can mimic the keystrokes, mouse events, recognize images, and even extract content from documents in order to automate a task. By accessing data in a business application or a file, it automates the work by processing/ moving the data into another system. It can perform operations on the data such as copy/paste, reformat, validate against other sources, and input into another system. In addition, the process is performed with a higher level of quality over a human by avoiding keying errors. For many tasks, process automation can execute tasks 2 to 7 times faster than a person.
Example Business Case for Accounts Payable Automation
For the uninitiated, this can be challenging to visualize, so let’s use an example of a common automation ready task: accounts payable (AP). Why is this a good automation candidate?
Manual AP processes can cost businesses from $2.00 to $10.00 per invoice with a median just under $6 (it can vary by industry).ii It can also reduce employee productivity by at least 50%. According to APQC benchmarks, top performing organizations clear invoice errors in 3 days compared to the median of 5 days. The lowest performers clear errors in 7 days. [i] High performing businesses strive to lower the cost of processing and reduce the amount of errors/rework while processing more volume (transactions).
As for improving quality, we can model this using the 1-10-100 rule. This is a rule of thumb that estimates the cost of poor quality. The model generally states verifying data at the point of entry costs 1x, correcting the data costs 10x, and if nothing is done it costs 100x. It always costs more to correct and error than to prevent it.
Common challenges with the AP process include:
- Dealing with large volumes of invoices coming from different suppliers or cost centers, in many different layouts/designs and various delivery formats – post, fax, email, EDI.
- Errors associated with manual data entry – often the same transaction needs to be entered more than once, increasing the transaction cost.
- Labor-intensive processes required to input data can be complicated further if linked to different costs centers and /or involve currency conversion.
- Longer lead times due to manual routing, processing and sign-off procedures involving individuals in different departments and possibly different countries – early payment discounts are often lost, and fees can be incurred for late payments.
- Inefficiencies due to time spent gathering physical documents in order to match invoices with purchase orders, delivery notes, contracts and accounting system records.
- Risk of document loss or damage, as documents are received at multiple input points and may be routed or classified incorrectly. In some cases, there is a risk of attaching the wrong document.
- Inability to easily audit invoice processing and payments to match up document flow with the accounting system.
A simple AP flow could look like this:
- Receive an invoice in digital form (fax, email/PDF, XML/EDI).
- Open the invoice document.
- Open the accounts payable application and input the invoice data.
- Review the AP item. This may include checking for duplicates, validating GL coding, and PO matching.
- Send for approval (process control).
- AP posting for payment.
Example metrics for measuring efficiency or performance in AP can include:
- Total cost to perform the AP process, per invoice processed
- Cycle time, in hours, to enter invoice data into the system
- Cycle time, in days, to resolve an invoice error
- Number of FTEs that perform the AP process, per $1 billion in revenue
- Percentage of invoices that are manually keyed into the financial system
- Number of invoice line items processed per AP process FTE
- Personnel cost to perform the AP process, as a percentage of total process cost
“Regardless of industry, the lesser the human touch, the faster the process moves, and the fewer errors are likely to work their way in and require fixes later. Our research supports the logical assumption that manually entering invoices into the general ledger has a direct negative influence on the total cost of processing accounts payable per invoice”.[ii]
Automating the Process
To automate this process, an automation script is created to perform these steps. One or more bots run on a Windows machine to execute the script. This can be scheduled to run on a regular basis or run manually. At an abstract level, the automation scripts could appear as follows:
- Open the invoice from the digital source (image,
If the source is email, the bot would open a specific email account, open the email and access the AP document.
- Extract and read the AP document to fetch the data (e.g. payee name, PO, invoice no., amount, description, etc.).
- Open the accounts payable application and navigate to create a new AP item.
- Enter the data extracted from step 2 above into the AP form.
- Validate data against other sources, if necessary. (e.g. validate a PO number, or payee). If the item fails validation, save the item. Notify the appropriate parties (e.g. email) of the issue and action is required.
- Submit the AP item after successful validation. Optionally, the automation could send notification to involved parties.
- If manual approval is required, the script would wait for an approval (or rejection) and then post to payment.
- If rejected, the item could go back to AP team for review and action. The automation could send notifications as appropriate.
Each step is performing as though a human was present and interacting with the application or operating system. Given this is automation, we could schedule this process to run 24 hours a day checking for new items to process. Would you rather have an AP process cost of $2 or $10? Would you rather have fewer errors or more (1-10-100 rule)? It seems obvious and yet many companies haven’t taken the next step in continuous improvement. These efficiencies improve when you find opportunities to incorporate AI as part of the process to drive decisions.
This is a simple scenario and yet can be applied to almost any rule-based business process. According to respondents, organizations with successful automation efforts are more likely than others to designate automation as a strategic priority. Thirty-seven percent (37%) of organizations with successful automation efforts designate it as a strategic priority. Sixteen percent (16%) indicated their automation efforts were initiated to keep pace with competitors pursuing automation.[iii] What is your Intelligent Automation strategy?
Contact Votum to understand how we can help in your Intelligent Automation Journey.
[i] “AP Process Issues: Symptoms, Root Causes, And Risks”, Mary Driscoll, Digitalistmag.com, 31 August 2016