ERP and the A.I. Factor

Artificial Intelligence (AI) is here. Once a topic of conversation, news, and science fiction, AI has finally entered our technology landscape. We are only beginning to see the impact it will have on our businesses, our jobs, and our enterprise software, but we already know that impact is profound and growing.

Until very recently advanced AI functionality was extremely expensive due to the limited and immature tooling available. Constructing the algorithms required highly skilled and very specialized employees, not available to most software vendors. In addition, these new and complex algorithms required immense amounts of additional data storage for pattern recognition and required CPU-intensive computing power for pattern recognition.

All this has changed. We will see it often. Current ERP software will incorporate AI capabilities more and more, and here is why.

Today, all major cloud vendors offer both PaaS and IaaS that specifically address computing and storage issues. Fierce competition between major cloud players has resulted in decreasing prices for both cloud storage and computing power, literally democratized machine learning functionality and AI capabilities. It is allowing software vendors to incorporate more-complex algorithms that crunch bigger and bigger datasets.

The result is ERP with AI for all.  It is leading to ever-more-advanced solutions. It has given birth to a whole new range of systems capable of making decisions based on historic data. Data collection is becoming pervasive, automatic and non-intrusive, instead of spotty, manual, and requiring high levels of interaction. Natural language input will soon decrease the need for manually intensive UI.  Increasingly, computers will proactively support decision making.

We recently announced our new digital assistant, Wanda. Wanda and its functional agents epitomize the evolution to semi-intelligent and self-driving solutions. They liberate people from their tedious manual interactions with enterprise software and allow them to focus on running their business and serving their customers.

Should we fear AI?

Human beings all share an important trait – the ability to devise tooling that simplifies tasks and drives our species forward. Our tools liberate us from mundane day-to-day chores, so we can focus on more abstract challenges, solving more problems – problems we solve, ironically, by devising even more tools. This has made us the most successful and adaptable animal on the planet.

Every new technology invented gave rise to a backlash. Change is difficult. Many people will prefer the status quo.

Despite resistance to change, advances in technology always prevail. In time we learn to trust and use our new tools with enthusiasm. The introduction of AI into enterprise applications will follow the same path. It’s only human. Initially there will be lots of resistance because AI will fundamentally change how users interact with software.  Large amounts of structured and unstructured new data will enter the ERP system invisibly, through mobile, UI improvements and IoT. Knowledge systems will employ agents – computer programs that decide and act independently on behalf of an employee.  Many users will feel threatened. Despite this, technology will prevail. It always does because companies will always strive to operate more efficiently and focus on their primary objective, serving their customers.

Should we then fear it? No, not at all. New technology brings us better living standards, safer working conditions, more goods, better services, and endless benefits. AI will do likewise. It will remove tedious, repetitious work.  It will empower employees to improve their performance. It will enable companies to provide more goods and better services. A new era is upon us. Embrace it. The impact of AI on ERP will be enormous and wonderful.

Self-driving enterprise software

By using advanced technologies within pattern recognition, machine learning and computer aided decision support systems, as well as new IoT devices, software can make automated decisions on behalf of the user, or provide guidance to allow users to make intelligent decisions based on smart patterns. This is achieved through self-learning software that assumes actions based on previous or similar scenarios and uses that to pre-populate and self-drive.

Interaction with the software is minimized to only involve people when human intelligence and experience is critical to resolving a business issue or to drive business processes in the right direction. If user interaction is required, a UI will be provided in its simplest form to accommodate that, on the users preferred device of choice, intelligently adapted to the situation.

Incredible value can be derived by limiting the amount of manual input required by users so they can focus on serving customers. This is the future of business software.

What is self-driving enterprise software really about?

In a nut shell self-driving enterprise software is literally about creating software with which the user does not need to interact, or at least interaction is minimized. The reality is that the majority of the interactions users have with their enterprise software is administrative. It’s not about doing their job well or running their business but simply maintaining the mechanics of sending invoices, doing bookkeeping or filing expense reports.

All these mechanic interactions can be tracked, analyzed and fed into algorithms so the software can learn typical usage patterns within a specific context whether that’s procurement, invoicing, bookkeeping or expenses. Through pattern recognition and statistical analysis the software can “learn” typical user actions in specific situations and automate the expected action in the future.

Analyzing past interactions will allow us to construct more semi-intelligent solutions. Solutions that dynamically alter behavior based on previous interactions, decisions and user actions. There are a number of different aspects to self-driving software.

Adaptive UI

The fundamental idea of adaptive UI is, as the name suggests, UI that intelligently adapts to the user based on previous interactions. In today’s enterprise software, forms or screens typically contain many fields that need to be filled out by the user. Some fields are linked with other fields, and there are even hard dependencies meaning that if a certain field has a specific value, others need to be populated exactly right for the system to accept the data as a whole. As a causal user simply entering a purchase order it can be a horrifying experience – regardless of how nice the UI looks. The adaptive UI records everything the user is doing, when he or she is doing it and in what order. If a user always populates a specific number of fields with the same values it can do that automatically. Even better, if some fields are always left untouched – they can be removed dynamically from the UI to make it less ‘noisy’ or confusing. By removing un-used fields, and populating everything automatically based on previous interactions – the system can complete the task on the user’s behalf. For example, if someone orders A4 binders on every second Monday from the same vendor, the system might as well just do it.


Normally business processes require user interaction for approvals. A manager needs to approve a request for a new laptop from one of his reports for example. If the manager in question always approves requests within certain criteria, most commonly the price, the system can give that approval and notify the manager it’s been done.

Historical and Current Data

What I’ve discussed above is all based on historical data; however, using aggregation of historical and current data – like your present location software has the ability to automatically populate travel expenses, field service work reports, and even suggest additional maintenance work do be done as part of a field service visit.

Some examples of where self-driving ERP will add significant value:

Expense Reports

Expense reporting is a great example of an administrative task that lends itself to self-driving software. In the future systems will automatically create line items in an expense report based on the scanning of a receipt. By extracting the address the system can gauge what the item is and based on the time know if it’s breakfast, lunch or dinner. Essentially populating the complete line. Rather than having to store the receipt, go into the system and enter expense type, date, amount, all the user has to do is photograph the receipt.

Field Service

Within field service, systems can schedule maintenance visits for equipment based on the previous maintenance of similar equipment at other customers. Also, if a visit is already planned, it may prove more economical to – pro-actively – service other equipment while already on-site.

The future is here

Previously analyzing massive amounts of data was expensive due to the amount of data storage and computational power required. However, due to ever decreasing prices on storage and CPU, running advanced algorithms becomes more accessible enabling inclusion of semi-intelligent agents into a vast majority of the functionality within enterprise software. Customers will benefit from being able to focus on running their businesses and serving their clients rather than the tedious mechanics of maintaining the ERP.

This is why we focus on developing enterprise software that allows customers to focus on their business instead of focusing on the software, by making it completely self-driving. We’re already starting to deliver this in our software.