This blog article was first published on our Norwegian blog. You can read it here.
When summer rolls around, the rooms at our headquarters are filled with excited and energetic students. Our summer internship takes place over six weeks, where each student participates in real projects, uses technology that we work with daily, and gets to know other Visma employees.
“What surprised me the most when I started at Visma is a tie between the great working environment and how much freedom and trust we got during our project.”Julie
The students were divided into teams across several Visma companies and worked with various technologies, customers, and challenges. Here’s a little peek into the projects they worked on.
Tech Hub’s Cloud and Engineering: Using AI to improve software engineering
Aurora experimented and worked with a Slack app integrated with OpenAI (ChatGPT), which has been trained with specific information about technical debt and continuous deliveries. The purpose of the bot is to answer questions about these specific topics, and the goal is for this bot to be used as one of the aids in our software engineering assessments, which most teams in Visma use.
The result was a well-designed and functioning Slackbot, integrated with OpenAI, that can answer questions from both Slack and Confluence (another key goal for this project). Moving forward, the Architecture & Technology team will continue with development and make the bot available in various assessments.
Visma Software AS: Developing a new documentation platform
Amund, Herman, Julie, Tilde, and Joakim created a new documentation portal for Visma Software. They worked independently and were involved in every step of the process – from talking to internal customers and specifying requirements, to writing code and deploying it on Azure. They managed to convert all content and create a working solution for one of the company’s products. Unfortunately, they faced some technical challenges, which means further work must be done on it before it can support all products. All in all, it will be very useful and valuable to the company!
“I think the most challenging thing about this project was setting priorities. Our project was very extensive, so it was important that we used our time wisely and that we utilised the expertise of the five team members in the best way possible.”Herman
Tripletex: Exploring the viability of the Copilot AI assistant
Christine, Henning, and Marius – all students at NTNU – analysed the possibility for Tripletex to develop a generative AI assistant that can further simplify everyday life for their customers through a completely new and innovative service.
To see if a Tripletex Copilot could be a viable solution, they needed to answer the following questions:
- Does it solve real customer problems?
- Is it technically possible to create such a solution?
- Can such a concept be profitable?
The students explored various hypotheses and validated their assumptions along the way. Through testing different technologies and conducting user interviews, their proposed solution went through several iterations based on learnings and new data.
Tripletex is very satisfied with the work Christine, Henning, and Marius did this summer, with the way they took initiative to drive the project forward, and the recommendations they made. Tripletex got invaluable insights that they will take with them in the further work with this concept.
“Developing such an AI assistant is very new, so there were relatively few similar projects or sources to lean on. In addition, this was a project without a conclusion. Therefore, I found it most challenging to set a direction for the project, but it was also one of the most educational aspects of the project.”Christine
Visma Enterprise: Automating chat support for Enterprise Plus using generative AI
Visma Enterprise started 2023 with an ambitious goal: to explore the possibility of using Generative AI and Large Language Model (LLM) to automate chat in Enterprise Plus. This innovative solution aimed to help users find answers to simple questions and promote self-service. Interns Mathuran, Vegard, and Henrik learned the technology, identified essential components, and evaluated different suppliers all with the ultimate goal of creating a proof of concept that showed the AI bot in action.
And they succeeded!
The enormous effort put in during the summer months has given Visma Enterprise concrete measures and a clear plan for the way forward. This includes optimising the content for the bot, building the infrastructure behind the solution, and, perhaps most importantly, envisioning the future role of this technology in the Enterprise.
They are so grateful to Mathuran, Vegard, and Henrik for the outstanding cooperation and dedication that made this achievement possible. Their project has not only expanded Visma Enterprise’s technological horizons, but has also laid the foundation for a more efficient and customer-centric approach.
“The highlight of the project was when we were able to test the chat bot on various support matters. It was a very good feeling to get something working after trying countless times, and also that we could get quantitative results that could be used further.”Mathuran
Visma Resolve: Forecasting cash flow with PowerOffice (NO) and Rompslomp (NED)
During the summer, Visma Resolve had the pleasure of welcoming eight students from the Computer Technology and Industrial Economics departments at NTNU. As the spearhead of Visma’s increasingly large investment in AI, Visma Resolve is constantly looking for new tech talents to develop services that improve people’s work lives through the use of machine learning and optimisation. This year’s summer students delivered from the first line of code!
The group immediately embarked on three strategically important AI projects of high value for Visma, Visma Resolve, and our customers. Simen and Mathis delved into anomaly detection to help accountants identify anomalous transactions. Erik and Jarl explored advanced prediction models to predict time registrations. Marte, Tobias, Herman, and Heidi developed a forecasting model to predict the financial situation of small businesses.
Visma Resolve is so grateful to the interns for their efforts on the projects, and also for their social contributions in the office – in the Mario Kart room and at various dinners and karaoke bars!
“The most challenging, but also exciting, thing this summer was the size and depth of our project. Acquiring new domain knowledge within accounting, while at the same time having to familiarise yourself with large datasets before you can even start with machine learning, has been both complex and time-consuming.”Tobias
A journey of social and professional replenishment
In addition to the academic programme, the summer also consisted of a lot of fun outside the projects. At the start of the internship, all of the interns went on a cabin trip where they got to know each other well outside the office.
Each of them were also given their own buddy, who took good care of them during the programme and arranged social activities. They had joint training sessions at Barry’s Bootcamp, played paintball, dined at local restaurants, enjoyed stand-up shows, attended barbecues, and much more.
“The highlight was, without a doubt, the closing evening with dinner at Kverneriet and then karaoke at Syng. A perfect end to an eventful summer!”Henning
Interested in joining Visma as a Summer Intern or Management Trainee? Check out current opportunities on our Careers page.