On the Voice of Visma podcast, we sit down with leaders and colleagues from around Visma to hear their stories, learn from their expertise, and share the best lessons they’ve learned throughout their careers. These are the stories that shape us… and the reason Visma is unlike anywhere else. New episodes are released Wednesdays on Spotify, Apple Podcasts, Amazon Music, and YouTube.
The text in this article is from Episode 01 and has been edited for length and clarity.
Welcome, Jacob! Can you tell us a little about your role at Visma?
Yes, of course. I’m the AI director in Visma. Together with the AI strategy and enablement team, we’re driving two transformations in the organisation. The first one is that we are empowering all 15,000+ employees with generative AI tools so that everyone can do more of the things they are good at and have a better time at work. The most difficult thing that we are working on in this organisation is figuring out how we can use AI to create more value for our customers.
Where did your background and interest in AI come from?
My motivation for coming into Visma goes actually back to where I studied in Trondheim, which is a city in the middle of Norway. While I was writing my Master thesis there, I did one of the most advanced things I could do back then to solve a very complicated problem at the university hospital. I visited the room where they plan all the surgeries, and it’s eight or nine senior nurses planning the schedules. Our task back then was, “Okay, maybe these nurses could be better spent on actually helping out with the patients, helping out with the surgeries”. So we applied super advanced algorithms – I had an evolutionary, multi-objective scheduling algorithm that took into account the uncertainties of the surgeries. It was super advanced, and I was so proud. And I went to the nurses and said, “Hey, my laptop can do this, and you can go and help out with the patients and the surgeries”.
Whenever I go back to Trondheim, I ask my brother and his wife about the latest news from the hospital and if the nurses are now helping out with the surgeries. And do you know what the answer is? They’re still working in the same way. Because it takes so much more than just an algorithm to make a difference in an organisation.
“The realisation that I had back then is that we need to work as an organisation, as a team. We need to understand the complexity of doing something like that in a hospital, which is a complex organisation. It’s a lot about systems. It’s a lot about software. It’s a lot about people.”
What’s been the hardest part of our AI evolution so far?
I think there have been and continue to be a multitude of challenges. On the demand side, public and private sector customers need to know that this kind of technology is available and tell us what they expect in terms of quality and compliance. On the vendor side, we have to help them deliver. We have to show them that it’s possible.
“It’s not just a technological challenge. We need to understand the customers, understand the need, understand the value, and take the chance to develop something like that. Then communicate and work over a period of many years to prove that we can deliver.”
We’re an organisation that builds up a lot of experience with using artificial intelligence in a compliant way, in a responsible way, in an integrated way with systems that customers already have: accounting systems, ERP systems, HRM systems, workforce management systems. So, we have that experience; we have the systems in place. We understand the dynamics. Now we are integrating more and more artificial intelligence into the systems. And that gives us a headwind into this development. That’s why it’s so exciting to be in this organisation.
Where does Visma AI fit into helping the future of society?
Today, we have over 1.7 million customers that are not individuals but organisations or entrepreneurs or organisations in the public sector. So, we have a huge impact on society. These organisations employ a lot of employees and provide welfare services or services and products in the private market. We use technology to help them, and we have done that for decades through digitalisation. But, now, we can also do it through AI.
“AI is very much aligned with the vision of Visma. We have a vision of shaping the future of society through technology. And I think that’s a more powerful vision than it has ever been.”
I would like to paint a picture with the television series The Bear. It’s about a Michelin chef who takes over a New York restaurant because his brother is deceased – and it’s totally chaotic. In one of the episodes, he’s asked to provide the payroll for the last two years to the tax authorities. Well, he just started to revamp this restaurant like two months prior, so he doesn’t even know where the fire extinguisher is. So how can he help? It’s an impossible task. That is kind of why Visma is here. Because if he had put all of his stuff into the cloud, that would be available, it would be compliant, and it would be more productive. It would also allow them to make smarter decisions – and that is what we’re providing to entrepreneurs and to organisations.
Working smarter with AI agents
When you look at the future of AI, I think we’re moving from SaaS applications to agents. Even today, we’re building agents on top of our existing SaaS products that create a much more intuitive way to access data so that our clients can make better decisions. It’s similar to talking to a human, which is much more natural than interacting with an application and a computer screen and a keyboard.
It’s similar, I would say, to the iPhone moment. Before the iPhone, you had phones with 50 buttons on a very small phone. Nobody wants to use that. It’s much more natural to use your hands and scroll. It’s the same with apps. You have this QWERTY keyboard and a mouse, and we all adapted to that. We’re doing our best to circumvent the fact that we are interacting with data in this unnatural way with a mouse and keyboard.
With the rise of agents, we’re building a more intuitive user experience. Like I mentioned before, it’s more natural to have a conversation with something human-like, like an agent that understands you, your context, your intentions, and helps you to do what you should be able to do. Of course, there’s a long way from application to agent – we need to work with much more robust automation, much more integration between systems. That’s something that will really boost our compliance productivity and help us make smarter decisions.
Where it gets tricky is with some of the limitations with the technology. Still today, these technologies – these large language models, generative AI – they’re probabilistic. It’s like throwing a dice. You don’t really know what you’re going to get. And that makes complete automation quite difficult. So, we have some challenges to overcome. I would expect for agents to be kind of part of business applications or mainstream consumer applications to take at least two or three, four years maybe.
Improving digital accessibility
Beyond that, and more importantly, agents also help create more availability to these kinds of systems. When we build swimming halls, for example, we build a ramp so that people in wheelchairs can get into the swimming hall. As a society, we accommodate that. But it’s not the same level of accessibility yet if you’re working in a web application or if you’re searching for information on the internet. There was a case in Norway last year where a guy with poor eyesight wasn’t able to find information on how to vote because the municipality didn’t have a website that accommodated people with poor eyesight. It should go without saying that, if you have a disability, that shouldn’t stop you from becoming an entrepreneur or making a difference in the public sector.
“By making software more available through having a conversational interface, we’re really making a difference in society. And I think that’s where we’re headed. I think it’s not just an opportunity but a responsibility.”
What’s Visma Group’s approach to AI?
So, I would say we are driving two transformations. The first is that we empower everyone at Visma to use these generative AI tools in a way that is relevant for them to do their job. And we’re also using that technology to figure out how we can create more value for our customers in our products.
Is it priority number one for Visma Group? I wouldn’t say so. I would say that we’re trying to figure out, as an individual, what is important to me. What are my strengths? What are my contributions to Visma and to our customers? What am I good at? What do I like to do? And then the question is: How can generative AI help you do your job even better and to have even more fun at work? That’s the level of urgency that is in Visma and the entire industry.
Similarly, on the product side, it’s not that the strategy is that everyone should do AI in their products. You should be asking yourself: How can generative AI and AI help us to create the most value for our customers?
“It all comes down to aligning generative AI and AI with our aspirations as individuals and our strategies with our products. Each of us must find a way to use these tools. But I think there’s a positive way to do it.”
Why should employees actively use and be excited about AI?
I encourage everyone in Visma, all employees, to apply a positive mindset to how you can use these tools. As a developer, you can use GitHub Copilot to stay in the flow instead of going out from your coding environment to your browser and figuring out how to go about solving an issue that you’ve stumbled upon. You can do it in your application. That is kind of the primary use case for these tools today, and more than 50% of Visma developers are using GitHub Copilot.
You can also do it to produce code faster from scratch, to automate the testing, or to create tests in an easier way. Thomas Dohmke, CEO of GitHub, expects AI to be a part of every step of the software development cycle. And I believe that, too. So it’s not going to be just about writing code. It’s about figuring out what to develop. It’s planning how you, as a team, are going to develop that. It’s doing the implementation, but also doing the deployment, the pull requests, security and, of course, monitoring the application, creating architecture. I mean, AI will influence all these steps over the next few years.
For people in customer support, it’s super relevant to look into this technology, as well. How can we feed large language models with our best knowledge so that these models can help us with providing more and better support to our customers, and help our customers with the most challenging questions? I think these models can help users to log into the application if they have lost their password or do some of the easier stuff, and then we can help them with the real problems; the tough ones, the unique ones that require our support.
“For anyone in Visma, there are so many capabilities with these technologies. Start today for the reasons it provides value today, but also because it’s a step into the future.”
And introducing… VismaGPT!
I’m so happy to say that we recently launched VismaGPT, which is our own kind of ChatGPT-like experience. That will allow us to have an offering that is more compliant and where we can integrate the models and capabilities that we believe in and that Visma employees want. Already today there’s a wide menu of such APIs and services that we can integrate. But there will be more in the future. So, we’re very excited about that and eager to learn about how Visma employees see that response.
Last question: What does being “Champions of business software” mean to you?
I’d actually like to use the Queen song, “We Are the Champions”, as an example to answer this question. Many people think that song is about winning, but I actually think that song is not about winning. I think it’s about the hardships prior to winning. “I paid my sentence… We keep on fighting till the end… I’ve had my share of sand kicked in my face, but I’ve come through… I made mistakes.” And I think the word “Champion” says more about what you have done than what you are. It means you have competition behind you. You had to fight to get something. And that’s what we’re doing as Visma, as well, as a team.
We are continuously fighting to make the most impact. We don’t feel satisfied in this ever-changing environment. There are hardships underway, and with artificial intelligence, there are a lot of ethical dilemmas. There are a lot of regulations. There are a lot of our customers that are afraid of the technology – in many ways for good reason. But we are here to understand the technology, understand our customers, understand the regulations, and deliver this technology in a responsible way.
Listen to the Voice of Visma podcast on Spotify, Apple Podcasts, and Amazon Music. Or watch every episode on YouTube.