Why should you care about purchase invoice processing?

Processing purchase invoices is one of the accountants’ work tasks that take huge amounts of time, a whopping 20–50% of the accountants’ weekly working time! 

Processing invoices is repetitive, manual, time-consuming, and accountants spend a lot of time working on them, even as it doesn’t provide much value to the end customer or the accountant. In this part of the white paper, we’ll dive deeper into the issue with purchase invoices with the following themes:

  • What are purchase invoices
  • Why do they matter
  • The invoice handling process
  • How can automation help this process

Purchase invoices in general 

There are many kinds of purchase invoices but they all are a form for a seller to get money from the buyer. However, there are huge differences in the formats that invoices arrive in.  

You’d think technology has already made invoice sending and processing into a smooth and agile process, but unfortunately, there are still many old-school customs around. In many countries, PDF invoices sent by email are still the most prominent invoice format and even traditional paper invoices are still used in (thankfully) small amounts. PDF invoices might sound like an advanced option, but thanks to modern technology there are way easier and more effective ways to send and receive an invoice.

E-invoices are the most modern format meaning invoices that are in a structured format from start to beginning. This format allows AI capabilities to be maximally utilized. E-invoices are forecasted to massively increase their market share in the coming years (Schubak 2022, Lastiri 2022).  In some countries, like Finland, the diffusion has already advanced very far, with e-invoices covering the great majority of B2B trade. It is widely agreed that this same trend awaits other countries as well in the near future. Other Nordic countries such as Sweden, Norway, and Denmark are also at the forefront of this development. (Schubak 2022, Lastiri 2022)

The invoice formats play a significant role when moving towards more automated and efficient accounting.


COST.              “According to a 2021 Ardent Partners study, the average cost to process an invoice with little automation is $10.95.1 Yikes — that’s a big hit to the bottom line! Those that have developed a more mature level of automation are processing invoices at only $2.25 on average.” (Esker, p. 5)

“Take Valdese Weavers for example. By automating their manual AP operation, the company is now saving $40-50,000 annually thanks to drastic reductions in paper and postage costs, and another $18,000 by reducing time spent on manual tasks.” (Esker, p. 5)

TIME.              According to the previously mentioned 2021 Ardent Partners study, the average time to process an invoice with little automation is 11.9 days1, while those with mature levels of automation process invoices in 3.3 days on average — that means companies that leverage automation process invoices 73% faster

Purchase invoice processing

As noted, purchase invoice processing takes a lot of time but doesn’t really create additional value. What is specifically stealing all those precious hours within this process?

The most time-consuming phases of this process tend to be the first three phases – reading invoices, filling in the needed information, and posting the invoices. In addition, challenges are already forming at the starting point, depending on the arriving invoice material. 

E-invoices are the most automation-friendly format and present the highest quality of incoming invoice materials. Automation solutions allow also other kinds of formats to be automated as well, but the quality of scanned invoice material doesn’t usually match the quality of e-invoices, which creates challenges in later phases of the automation process.

It is important to recognize these separate steps in the process because they are in the key position in understanding purchase invoice automation and different solutions for it. But what they all share in nature is that they are all very manual tasks that need a lot of ‘clicks’ from the accountant rather than needing the accountant’s actual insights. 

The challenge with purchase invoice processing

One of the biggest challenges in purchase invoice processing is how many steps it includes. They all pose their own challenges that slow down the invoice-handling process and affect the other parts of the process as well.

Purchase invoice processing pain points

1. Old-school invoice formats

A large variety of different kinds of invoice formats pose challenges for processing because they are complicating the utilization of automation solutions later down the road. 

This first bottleneck can be approached in two ways. The first solution is to include scanning solutions such as OCR  that scan the information from PDF- and paper invoices to a format that can be processed with a computer or AI. However, this scanning is often far from perfect, resulting in poor quality invoice material that prevents them from becoming fully touchless since often all the relevant information is not gathered and the accountant has to check whether there are errors.

The second way to beat this bottleneck is to have it completely removed by the diffusion e-invoices. E-invoices are already in a structured format and therefore don’t need any scanning. The use of e-invoices is constantly increasing but is highly dependent on local laws and regulations of each country as well as the culture of buyers demanding e-invoices. This bottleneck has mostly been overcome in Finland and Norway for this exact reason; e-invoices represent a large majority of all invoices sent. Luckily, the research suggests that those hurdles will be solved quite rapidly in other Nordic countries such as Sweden and Denmark and are starting to become more common outside the Nordics as well.  (Schubak 2022, Lastiri 2022).

2. Invoice data-filling and posting 

Data-filling and invoice posting have great significance when it comes to the efficiency of invoice processing since they steal a huge amount of time from the accountant. This bottleneck is easy to overcome with the help of automation and AI.

It is possible to automate data-filling by predicting accounts, cost centers, VAT statuses, and other dimensions needed for the invoice. These types of automation solutions are already easy to find on the market but they only solve a part of the problem instead of offering a comprehensive solution. 

Having only automated data-filling means that the accountant stills needs to check every single invoice regardless of the automated data entry. This means that it still doesn’t reduce the number of invoices that the accountant has to process daily. Luckily, more comprehensive solutions are entering the market. 

When talking about purchase invoice automation the solutions differ significantly. 

In reality, some automation solutions may only automate one step of the entire invoice-handling process and other solutions are able to automate them all. Thus, it is important to understand the differences between available solutions.

Keeping the two bottlenecks in mind – purchase invoice automation can be divided into two categories 1) Automation focusing on data entry and 2) Automation focusing on processing invoices.  

These two can be further divided into four levels of automation.


The accountant creates and maintains company-specific posting rules manually. 

The purpose of Level 1 Automation is to reduce the manual work required in processing invoices, as the information is entered into the invoices according to posting rules made by the accountant. These rules however need to be maintained and revised when something changes.


AI creates and maintains company-specific posting rules automatically.

The purpose of Level 2 Automation is to reduce the manual work required in processing invoices, as the AI enters additional information for the invoices on behalf of the accountant. The accountant doesn’t have to keep track of updating the posting rules.


Reduces the number of invoices that the accountant has to process.

In Level 3 Automation, automation is taken a step further. The accountant can define the criteria that, if met, invoices can be processed completely automatically without any human intervention. Level 3 Automation requires that Level 1 or Level 2 Automation is already in use.


Progressive Automation represents the highest level of purchase invoice automation.

In Progressive Automation, the AI works autonomously and tracks the accountant’s work and based on this learns to identify invoices that can be automatically processed on behalf of the accountant. All this happens without any required action from the accountant.

The purpose of Level 4 Automation is to reduce the number of invoices that the accountant has to process without writing and maintaining rules.

As you can see, to automate these processes, you can use rules managed by either a human or AI. As one could guess, the AI does the job a lot faster in both cases. The key benefits of integrating AI into the financial services industry lies in obtaining supreme accuracy and speed for critical processes” (Frost & Sullivan 2017a). Artificial intelligence excels in repetitive back-office tasks such as checking purchase invoices (Forbes 2019).


AI Inside is the most widely used purchase invoice automation solution. With the help of AI Inside, more than one million purchase invoices from nearly 30,000 companies are processed every month.

“When contemplating the providers we learned that Finland has come further than Norway regarding technological advancements in the public sector and the accounting software market. It was a natural choice to pick a provider from the most advanced market.” JORID Trandem – Product Owner – Tripletex AS

AI Inside offers developers an easy way to update a single part of the purchase invoicing process or the entire purchase invoicing process. AI Inside is a combination of technology and knowhow that allows the software developer to update the purchase invoice process to meet the future challenges of customers.

Trusted by some of the biggest players