In this 3-part guest blog, digital transformation expert John Stinchcombe considers how end customers can create a business justification for invoice process automation.
Part 1: Introduction, how automated invoice processing solutions work, key success factors for this software application
Part 2: Factors to consider when building a business justification, risks vs rewards.
Part 3: Implementation and ongoing costs, financial and non-financial outcomes, summary.
Having now discussed the background of automated invoice processing and what it takes to have an effective and efficient solution, Part 2 will talk about the elements that need to be considered when building a business justification.
Factors to Consider When Building a Business Justification
There are two important factors to face when considering the automation of Invoice Processing and building a justification.
Firstly, there will be numerous points in the process where exceptions or omissions are introduced by the supplier. These need to be handled properly and this may require further checking and liaison with the supplier and colleagues within your organisation. In the same vein are substitutions for ordered items. These are exceptions and will also require human intervention which naturally, in turn, impacts automation rates and thus potential productivity savings.
Secondly, the entire process is subject to an inherent time constraint or SLA (Service-Level Agreement) determined by your supplier’s payment terms or agreement. For example, if payment terms are 30 days, then the entire process needs to be completed within that time otherwise, the supplier will chase or worse, they may impose more onerous sanctions such as late payment charges or even credit stop. If your organisation suffers from high peak volumes relative to the median volume. This can cause significant work and is an area where automation can really make a positive impact.
As automated Invoice Processing systems replicate human tasks, there are several factors that affect the overall efficacy of the approach. It’s worth flagging a critical and fundamental difference between automated systems and human based processes. Automated systems rely on hard-data and immutable logic, whereas humans have access to more creative approaches to problems and can also use tacit knowledge and experience.
Tacit knowledge can’t easily be replicated by software and so we need to reduce our reliance on this for invoice processing.
The following table summarises the 8 factors that will affect overall levels of automation and thus the effectiveness of the solution. It’s an essential step in trying to establish a business case; the degree to which each one impacts the overall automation rates needs to be assessed for individual customer cases.
|Percentage of invoices using purchase order references||For every invoice that doesn’t reference a Purchase Order, an AP user will need to locate and retrieve the relevant PO and enter it via the validation screen manually. Provided this PO details are available to the solution, this process will initiate verification against PO line-item details. Analytics will highlight the ‘offenders’ and pressure can be applied to ensure greater compliance.|
|The quality of paper-based invoices and the resultant scanned image.||For paper-based invoices, this is a critical factor and can be best summed up in the old IT adage ‘garbage in, garbage out’. If an OCR process is required to process poor quality invoices the output will be poor quality data. Although printing techniques and scanning technology has improved this area, it remains an important factor. This is especially the case when a significant proportion of invoices are submitted as paper – poor quality OCR will significantly hamper attempts to automate the process.|
|Percentage of invoices sent in via email as ‘born digital’ electronic formats||Stating the obvious here but if invoices are submitted via email rather than physically as paper documents, then the task of scanning and managing paper invoices is eliminated. The task of filing or destroying the originals goes away too. Automated systems handle the task of monitoring email mailboxes 24/7 very well without operator intervention.So-called ‘born digital’ versions of invoices (such as PDF or HTML format that are typically submitted via email attachments) don’t require an OCR process to convert the scanned image to text. As a result, the absence of any OCR errors will mean that users will not have to correct poorly extracted data (as may be the case if paper invoices or images are of substandard quality). This also allows data extraction algorithms to work to at optimum levels. However, submission of electronic born digital invoices does not mitigate two issues: (a) suppliers’ errors and omissions and (b) due to an unorthodox layout or format, the system’s data extraction logic may still fail to locate some required data fields; in both cases, the verification processes will highlight occurrences and users will be required to intervene. Machine learning may adapt to these but the initial encounters of these will require human intervention.|
|Percentage of invoices submitted as compatible e-Invoices.||Like ’born digital’ format invoices above, e-Invoices don’t require an OCR process. However, the biggest benefit is that e-Invoices don’t require data fields to be identified and extracted – the structured data-set (aka individual invoice fields) are already distinct and simply need to be mapped. However, like all invoices, e-Invoices may also contain supplier errors and omissions; these will be trapped via verification routines for AP team members to manage.|
|The quality and accuracy of your supplier Master File data.||The value of an accurate supplier master file data cannot be understated in an automated solution. Matching supplier details on submitted invoices against a master supplier data is a key element of an automated approach. It’s an area where manual system’s reliance on AP user’s tacit knowledge is a benefit and a liability as humans can and do make mistakes. In contrast, an automated system requires a very close match and is rigorous in its approach. Poor quality data will make this automated process less effective and flag too many low-confidence scenarios and as a result, require increased AP team intervention.|
|The quality of your purchase order and data.||Per the Supplier Master data issues above, poor quality purchase order information may make the matching of key identifier invoice line items (such as SKU / Article nos) with their corresponding purchase order more difficult. The wider the gap between the two, the more AP team intervention will be needed – overall automation rates will reduce the efficacy of the system will fall short of it’s potential.|
|Mismatches between supplier SKU or Article references and your corresponding Purchase Order.||While this is unusual, this is a potentially difficult area if occurs and will require supplier liaison or changes to your purchase order process; automated systems are reliant on consistent data matches that may have been handled previously by tacit knowledge of AP team members. Built in analytics for the AP Invoice Processing system will flag these specific scenarios quickly so that they can be addressed.|
|The quality / timeliness of your goods receipting process and data.||As the default scenario for no-receipt flag is to hold an invoice in the system, delays in updating this crucial flag will ensure that invoices are passed for further up-stream processing such as GL Coding, Approvals and extraction for payment in the AP system.|
Risks vs Rewards
As with any IT solution, there are risks and rewards that must be considered and assessed. Clearly, the risks need to be manageable and minor and rewards need to exceed a threshold of ensuring the implementation beneficial for the business.
Summary of risks is as follows:
- Failure to Meet Expected Levels of Automation
This is a balancing act as over-estimation of the productivity gains will result in a rapid face-saving exercise on the part of the sponsor, or worse. Whereas, under-estimating the gains may result in failure to obtain the go-ahead. To use an Americanism; ‘mileage rates may vary’ and the level of automation will be influenced or rather, determined by some of the factors outlined above. There is a degree of getting one’s house in order prior to adoptions and suppliers such as Mysoft can provide expert guidance in this area.
One of the largest areas of consternation for those considering these solutions is the subject of ‘straight through processing’ rates or proportion of invoices that can be processed at ‘near 100% unattended processing’.
Many end customers conflate Invoice Processing with routine accounting updates or housekeeping tasks that can, and do, run without operator intervention. Unless close attention is paid to the quality of data and only invoices that conform to very prescribed standards, even modest levels of straight through processing are unlikely to be possible. The reality is that invoices arrive from disparate suppliers using a variety of different packages and mechanisms to submit invoices. Similarly, they consider (subconsciously) that their invoices are going to be processed manually and that any irregularities will get picked up and handled. The bottom line is that new users should expect to review most invoices post-recognition and a significant volume of these will require some work on them to handle errors and omissions. However, averaged out, this amount of work will typically only represent 20-30% of the time and effort needed to manually process all invoices; put another way, 70% to 80% of the work is automated. If your organisation currently has six FTEs dedicated to invoice processing, you should be able to easily manage the same quantity with just two FTEs and have some additional ‘bandwidth’ for peaks of volume.
- Integration with Finance and/or ERP Systems
As this is a crucial factor, it’s right to flag this as a potential concern. It is fundamental to the viability of the system that information can flow readily, automatically and with integrity between the AP Invoice Processing system and the Finance system. Not least of these is an audit trail and recording of key process metrics. Most contemporary Finance and ERP systems use APIs and Open Systems methods to achieve reliable and resilient exchange of files and information; some integrations are simplistic requiring only that files are deposited into a specific location while others are more sophisticated and use integrations such as web services as a more robust data interchange mechanism. Invoice Processing systems require that extracts of supplier master file and purchase order/receipting information are uploaded periodically. Once the invoice dataset has been extracted and verified, this information along with meta-data as well as the invoice documents (or links to them) are uploaded to the ‘target’ Finance or ERP system.
Evaluating these mechanisms is a key planning tasks to address customer’s concerns regarding automation and integration capabilities.
- Data Privacy and Protection
Although rare for invoices to contain personal information, they do contain potentially sensitive commercial data. This is a matter for the DPA function in the supplier to provide a statement as to who, where and how long your information is processed and held. As always, common sense prevails and using reputable providers will mitigate concerns around this risk.
- Stake-Holder Acceptance
There are two sets of stakeholders. Firstly, suppliers. Suppliers are external and those submitting paper invoices will ideally need to switch to emailing invoices as attachments to a nominated email address. While not entirely necessary, those that can switch from paper invoices to ‘born digital’ formats will help make the adoption and success of invoice process automation projects a lot easier for reasons outlined above.
The second set of stakeholders are internal users such as the AP team as well as colleagues that may be required to approve invoices or apply GL codes to invoiced items. These users are generally pleased to welcome a new solution that unburdens them of the onerous and thankless task of processing invoices manually. Once these users understand the changes to the process and the need for data accuracy, they typically enjoy using a consolidated Invoice Validation screen rather than having to flip between email, PDF viewers and the Finance System.
Remote users are particularly pleased as browser access means that they can process invoices without the potential hassles of accessing thick client applications across the internet. The removal of the need to key information twice is a big win for these users. From an AP and Finance Management perspective, they benefit from a significantly accelerated process that enables them to see the bigger picture sooner – invoices are no longer backed-up in email mailboxes and through the user of dashboards and analytics, they’re able to focus on essential AP best practice functions such as cash-flow, supplier management including the identification of miscreant partners whose adherence to data-quality may be lacking.
Part 3 of this 3-part blog series will look at implementation and ongoing costs, financial and non-financial outcomes, and a summary about the real value.
Find Out More:
- Why your employees are your biggest asset in the age of automation
- See X3CloudDocs for yourself in this short explainer video
- Take a closer look at X3CloudDocs and download the data sheet