Build or Buy? That’s the question

It isn’t self-evident for everyone that purchase invoice automation should be a built-in feature in the accounting software. Many software providers choose to offer their API to different kinds of automation developers and let their customers use these integrations if they want.

Well, why should you have it built in your software? We think there are two reasons: First of all, as discussed in this paper, purchase invoice processing is the bottleneck of accounting now and in the future, so automating it seems to increasingly become a part of core competence for accounting software providers.

The second reason is that we think you will lose your customers if you don’t do it because the other features (the core value you’re trying to deliver) may become a hygiene product that is very easy to build. And you will lose it to three different kinds of competitors that already exist in the market:

1. Software developers who have started with (or whose core competence has been all the way) AI automation and have created other financial management features on top of that
2. Established financial software providers who have later added AI to their software
3. The “faceless” competitor: all kinds of software cocktails where your customers deal with their P2P process by uniting different kinds of software, for example via API integrations between different software by themselves

Now or later?

By now we have hopefully justified why it’s useful for you to offer purchase invoice automation as a built-in feature in your software. The next question: Is this urgent?

There are many reasons why it is urgent, and the first one was displayed earlier in this chapter: Your customers will want it pretty soon after beating the first bottleneck: old-school invoice formats. So although all automation is not widely accessible in many countries right now, it will be sooner than you might think. Many of your customers are still battling with the first bottleneck but the digital transformation is moving those customers rather quickly toward the second bottleneck. By the time your customers are demanding those solutions more and more, it’s too late to start building or buying your own solution.

The second reason is the fact that your competitors are already making moves.

Many accounting software providers have taken invoice automation to their roadmap or have already started creating these automation solutions. For example, Norway’s leading accounting software-provider Tripletex began addressing this problem in 2019. After buying and implementing the solution, their opinion of the urgency is as follows:

“We found that invoice automation was a must-have in order to be in the front of technology and to have a system that the customers will still use in 3 years’ time. The investment was definitely worthwhile. People are only just starting to use it but in a few years’ time people will use AI more and more and then it’s really important to have it working already. We also learned a lot during the process and gained valuable insights. Our advice to others considering buying invoice automation is to start early and take enough time to make the solution as efficient as possible.” Jorid Anne Trandem, Product Owner at Visma Tripletex

We’ve decided to buy, but what should we buy?

The next challenge is the fact that the invoice automation market is quite flooded with different options. How could you know what to buy?

Buying a solution like this is a big decision. When you are planning to purchase an AI solution to incorporate to your invoice process, it’s good to keep the following things in mind:

The checklist

AI capabilities

Find out what does the solution actually offer. Like discussed in chapter x, the AI solutions on the market vary a lot. When comparing different providers have an idea of are you looking for and what elements should the AI solution include.

• Does the provided automation only automate one part of the entire invoicing process or the entire process?
• What different elements are included in the solution?

What does the automation performance actually mean?

Almost all purchase invoice automation providers talk about automation percentages or levels of automation. However, it is important to know that most of these levels or criteria for percentages are set by the automation providers and do not follow a set criteria. Because of this it is not possible to compare providers only looking at the improvement percentages that they promise – since they all tell their own story.

When comparing solutions take a deeper look at what do these percentages actually mean and how are they being measured. This way you can get better understanding of what does the AI solution do and how efficient they actually are.

How is the service built?

It is also important to take a closer look at how is the automation solution provided

• Is the solution built in REST API and thus possible to change if needed?
• Or, is the solution built on some custom or proprietary technology that can create a vendor lock-in situation?

The ultimate goal of course is that you don’t need to change service providers or transition to another solution, but it is good to keep options open or to at least acknowledge the boundary conditions. We recommend that the solution follows these guidelines:

• RESTful API
• Open documentation available
• make sure the built integration will be properly reviewed and commented

Pricing models

What does the price consist of and how is it formed? During the price negotiation process make sure to get a deeper understanding of the offered pricing model.

• does it have low transaction costs but big annual expenses
• what is included in the price and what is not
• are there any additional costs or hidden fees
• is the total cost easy to understand

Understand what the pricing means for your and end-client’s business

Provided level of support

What is the provided level of support that the solution provider is offering?

Building a solution like this takes time and effort from both parties. Together set guidelines for what kind of commitment is expected from both sides so that the mutually agreed goals are to be achieved.

Do they understand your strategic goals

Finally, one of the important things is to be clear that the automation provider understands your company’s strategic goals and is confident that this new automation solution helps achieve them. This requires a lot of discussions, brainstorming and trust among the people involved.

It’s crucial to be on the same page so that the best possible outcome is achieved.

Different AI providers on the market are for example FabricAI, Visma AutoSuggest, Medius and Rillion.