AI at Visma

Artificial intelligence is a valuable tool that doesn't just ensure business competitiveness, but also helps people and society thrive.

Learn about the incredible promise of AI and how we work with it at Visma.

We live in the age of AI—and it's here to stay

 

R._John_Wright_Tin_Man,_2015.png
Image licensed under the Creative Commons
Attribution-Share Alike 4.0 International license.

 

In the 1939 film The Wizard of Oz, the main characters are all searching for the Wizard. The Scarecrow is looking for a brain, the Cowardly Lion courage, and the Tin Man a heart—while Dorothy just wants to get home.

These human qualities, in particular intelligence (brain) and compassion (heart), are what we want AI to have too. We want AI to be intelligent enough to relieve us of tedious work, which it has become extremely good at. We also want AI to be transparent and compassionate, responding to people’s needs in line with what we expect. We are making headway on this aspect, continually incorporating ethical considerations into our practices and solutions.

Today, AI is such an integrated part of our daily existence, we hardly notice it. From the content we are shown in social media, to the music we get recommended, to the way our expense claims are handled, AI has a major (and growing) impact on our personal and professional lives.

This is just the beginning. Automation through AI is estimated to raise productivity growth globally by 0,8-1,4 % annually (McKinsey Global Institute, 2019). AI is unquestionably starting to deliver real-life business value. So how can organizations best take advantage of AI’s benefits?

Making the most of what you have

We are more aware than ever that we have access to a finite set of resources. There are only 24 hours in a day, with competing demands on our attention, energy, and wallets.


In business and society, we need to be efficient with the resources we have. We need to do more with less. However, organizations of all sizes are still missing out on enormous efficiency gains. These organizations are:

  • Less productive, with reduced time for important functions such as innovation and product quality
  • Less service-oriented, with a diminished ability to meet the growing expectations of their customers
  • Less able to satisfy their employees and are therefore a less attractive employer

Above all, they have a harder time reaching their goals—whether financial, social or sustainable. To reach those goals, they have to win the efficiency race and keep their business processes up to date. 

 

 

 

 

A look into an office where two colleagues sit at their desks focused on the task at hand.

What if machines could work more like humans instead of humans working like machines?

AI as we know it today is not about taking human jobs—it's about doing some of their tasks faster and better


One thing’s for sure: realizing AI’s full potential requires people and technology working hand in hand. It’s the people in your organization who drive the business, make up the culture, and deliver your end result. To stay competitive, you need to get the best out of your staff. They cannot do that when they are bogged down with busywork. 


AI greatly reduces busywork by first and foremost making the person it works for more efficient. Since it performs manual and repetitive tasks hundreds of times faster, it unlocks people’s time and creativity. Employees are freed up to focus on activities that bring more business and social value.


At the same time, AI is able to explore and find patterns in large amounts of data, a task that humans are simply not built to do. Humans can focus on what they do best: making decisions using the insights that AI brings.

The AI adoption journey

To adopt AI and fully utilize the business opportunities it brings, there are a few building blocks that companies must have in place. 

AI can be seen as a powerful extension of digitalization. Hence, you should begin with evaluating the state of your business’ digital transformation. Is digital technology integrated into all areas of your business, from how you operate to how you deliver value to customers? This building block is elementary to have in place before you embark on the AI adoption journey towards automation and optimization—and the incredible opportunities they bring.

An illustration of a man hiking up a mountain, with Digitalization, Automation, and Optimization representing the peaks he will reach in his business journey.

Digitalization

Adopting AI requires a solid digital foundation. It’s the highly digitalized companies that tend to invest more in AI and gain greater value from its use. 


But what makes a digital business? We can say it is the result of several digitalization processes, for example moving to a digital business model by adopting cloud technology solutions.

Automation

Automation is about letting machines take care of manual tasks as they have always been done, just automatically.


Automation is not a new idea; the very intention of automation is to remove human intervention. Just think about the evolution of our commute; from the early days where humans were hunting by foot to the use of horses and now a car that can drive on its own. 

Optimization

Optimization has an incredible ability to make the best of anything and can be described as a “next-level” version of automation.


When optimizing a process, it doesn’t just copy the task and do it the same way as done before (only faster)—it also finds the best solution.

 

The difference is clear

What’s at stake is the difference between surviving and leading, between reactivity and proactivity, between short-term thinking and long-term resiliency.

An illustration showing the diverging paths a business can take, with the AI path being clearly better than the non-AI path.

Ways in which AI empowers our customers

For decades, Visma has been at the forefront of making society more efficient through software, and we do not intend to stop. In fact, we are just getting started.


We have customers who struggle with some of the most challenging and complex mission-critical problems. With the power of AI, we are shaping the future of society through technology—helping companies fulfil their destiny.


Just like the Tin Man, we want to use our power for good. With the guidance of humans, AI can be the compassionate brain that gives people more time and opportunities.

 

At Visma, our AI efforts are based on the needs of our customers. We have therefore created AI competence areas—teams working with different AI methodologies—to overcome customer challenges, share knowledge across teams, and strengthen Visma's collective AI expertise. 


By empowering customers with AI technology, we can radically transform their manual processes into competitive advantages. Below are two examples of how AI has streamlined business processes for our customers. 

Helping home nursing care optimize its travel routes

In Norway, more and more elderly are receiving health treatment in their homes, creating a burgeoning in-home care industry. Until 2019, routing nurses to the right house at the right time was done manually—an extremely difficult and complex task. With just 3 nurses and 30 patients, the number of possible routes exceeds the number of atoms in the universe.


Visma Flyt Route Planner uses techniques from AI and Operations Research to help home nursing care optimize its travel routes. The app allocates patients to nurses and determines the best route for each nurse to take. Nurses can now spend 10% more of their time with patients—the equivalent of 2 extra cups of coffee per day.

See more examples here.

Delegating expense claim validation to a digital worker

Have you ever been writing a travel claim and waited for days or maybe weeks to get it approved? In claims verification, a digital worker, or "expense-bot", will do the job as a controller, but faster and with higher quality. The expense-bot verifies that amounts, dates, passenger names, and other details are filled out according to the attached receipt. The digital worker will either approve or reject the claim, and send it in the correct workflow with an actionable message.


For each claim the expense-bot processes, it frees up time. In fact, since 2018 the expense-bot has saved customers from 40%-90% of all time spent on claims validation.

See more examples here.

A teenager sits at the kitchen table eating breakfast and looking at an iPad.

Using AI responsibly

AI is among the most promising solutions for creating unprecedented value growth for our customers, shareholders and society. AI can solve many of society's sustainability and efficiency challenges, but will also lead to widespread changes in many dimensions that need to be addressed in ethical and respectful ways.

Visma follows these principles for design, development and deployment of ethical and sustainable AI:

Human centric
Our mission for AI is to extend human abilities. We aspire to act ethically with care and explore opportunities, while also being proactively aware of potential risks. We supervise our AI solutions and are ready to intervene, improve and mitigate issues where needed.

Transparent, explainable, fair and unbiased
We aim to proactively explain the development, use and results of our AI solutions to customers and employees in a user-friendly way. AI decisions regarding individuals are continuously tested for skewed results or biases that might have existed in the service-relevant data that are used to train the AI algorithms. An AI system that makes determinations related to individuals should be explainable to a non-technical person of why it arrived at its conclusions.

Privacy and data protective
Our AI development is based on privacy by design. We respect user privacy and data rights, and inform users about how data are collected, stored, used and protected through clear privacy statements. We strive to use only the minimum data necessary for every step of development, and encrypt or anonymise data where possible.

Accountable
We take responsibility for our AI solutions to give our users confidence, and provide a point of contact for questions, directly or through our trusted partner network. User feedback is continuously incorporated into our AI solutions, and any issues are addressed promptly.

Safe, secure and sustainable
Our AI solutions are designed and deployed securely and robustly, defended from adversarial attacks, to prevent misuse and reduce the risk of being compromised or causing harm. This also applies to any data used to train, develop or test the AI. Finally, for environmental sustainability, Visma conducts the majority of its computationally intensive AI development in data centres powered by renewable energy.

How we see the future of AI

The potential of AI is boundless. The technology is so versatile, it can be used in nearly all industries and business areas, and will undoubtedly continue to impact our jobs and ways of working. 


The analysts who study AI are confident it will not be a technological job thief. Instead, they believe that AI will bring with it a net job growth. 


AI's larger impact will be in complementing and augmenting human capabilities, not replacing them. Realizing AI's full potential requires people and technology to join forces—working hand in hand. 

An illustration of a woman sitting with a robot.

Latest from our blog

Colleagues sit in the office and discuss machine learning in front of the whiteboard.

Democratising Machine Learning through Hugging Face

🤗 Hugging Face is a company that democratises AI by making models available to everyone. Find out how Visma's Machine Learning Assets team is contributing to the cause by giving back to the community, in the field of Document AI.

A diagram showing a closed loop for AI evaluation.

Garbage in, Gospel Out – The Unrealistic Expectations of AI

AI has been a buzzword for quite some time. And as a professional in the field, you may have gotten the question: “Can’t we just use some AI?” In this article, you’ll get the needed support to explain why AI is or is not a good solution, through a framework to anchor your arguments. We call it the AI Evaluation Loop.