Enzoway
Enzoway helps companies develop self-driven teams and leaders through autonomous leadership and better communication. With structured tools and reflection-driven development, it strengthens both individual and collective performance. This is because a focus on relationship and performance creates a more engaged work culture, improves collaboration and increases productivity. In collaboration with the project, the possibilities of an AI-driven coach are now being explored that will eventually offer personalized guidance to further strengthen leadership and team development.
Burde
Burde is a calendar manufacturer in Växjö. The annual production of calendars amounts to approximately 10 million products, with sales in large parts of Europe as well as the USA. Since calendars are inherently seasonal, careful planning is required to optimize sales and minimize environmental impact. Through collaboration with HPC, a pilot case is being created to test how AI and existing data can improve sales and production forecasts.
Nelson Garden
Nelson Garden is a leading company in the gardening industry. They work to make it easy for people to grow where they live, regardless of scale. In this pilot case, we will explore the possibilities of using machine learning to forecast the sales of a specific product based on historical data and external factors.
Physiologic
Physiologic is a startup company at Kalmar Science Park that contributes to the digital transformation of clinical physiological diagnostics by introducing automation and advanced decision support with significant elements of AI. Together with HPC for SMEs, algorithms for data processing are being developed with the aim of making real-time analyses and forecasts smarter, faster, and safer.
Cenrotech
Cenrotech has extensive experience working with Business Intelligence and has developed the powerful tool Verksamhetskoll. This tool has made it easier for companies to get started with Business Intelligence and benefit from its advantages. As part of our pilot case, we will explore the possibility of integrating the company's chatbot Neo, so that it can answer questions that users ask in the tool. For example, if questions about the budget arise during a budget meeting, Neo should be able to provide quick and accurate answers
AnswerOnline
AnswerOnline is a company in Kalmar that delivers personalized answering service by their communicators. To further evaluate and quality assure their support calls, we will explore the possibilities of implementing this with AI in the pilot case. This is mainly in retrospect for already recorded support calls, but possibly also in real time. As the calls may contain sensitive information, only local AI models will be used and developed. The hope is to find a fast and accurate workflow even though everything is done on our own servers, so that many calls can be analyzed.
Unsealed
Unsealed is the market's leading company for piece purchases of articles and works to help newspapers reach a larger target audience and contribute to their business. The company helps several players reach new readers and increase their subscription bases and is now taking the next step in processing by using AI to take advantage of — unique data — in a more valuable way. The project with Linnaeus University aims to drive AI-focused product development that provides a deeper understanding of reader behavior, enables better segmentation of customer groups and optimizes pricing to meet different reader needs.
JLT Mobile Computers
JLT Mobile Computers is a leading supplier of rugged, vehicle-mounted computers used in demanding environments. Together, we are focusing on developing a solution that turns off the screen when the vehicle is moving and turns it back on when it stops. This screen saver solution is one of several safety enhancements JLT is exploring to increase operational safety.
CGV
CGV (C. Gunnarsson Verkstads AB) is a machine supplier to the sawmill industry. We have started a collaboration with their automation department. In the pilot project, we are investigating how their machine can optimize the packing and unloading of boards in the most resource- and cost-efficient way possible.
Roadlake Analytics
Roadlake Analytics helps municipalities and businesses identify development opportunities by finding the right information, analyzing relevant data, and comparing schools in a simple format. The aim of the pilot case is to automate the selection of schools for comparison and to create an automatic summary text.
Virum Älgpark
Virum Älgpark offers moose safaris where visitors can pet and feed moose. The park also has other animals, a café, and a moose shop. In the pilot case, we will investigate whether it is possible to separate the moose Saga from the park's other moose. If successful, the goal is to use moose recognition to monitor the movements and well-being of all moose in the park.
Swepart
Swepart develops and manufactures advanced transmission solutions. The pilot case is a first step to investigate technical solutions that can automate and predict cutting wear using vibration measurements and computer vision. Both methods generate large amounts of data that are processed using AI models.
Completed pilot projects
Measure & Change
Measure & Change is a small Växjö-based company specializing in categorizing invoice lines to calculate the environmental impact of companies' purchases. The categorization is based on the Common Procurement Vocabulary (CPV), a comprehensive list of over 10,000 categories. Manual categorization, despite some automated procedures, is time-consuming, especially for experts. The pilot case explored different approaches, including the full use of large language models (LLM) as well as a combination of LLM and simpler NLP solutions. The proof of concept used LLM to translate and process invoice lines and partially interpret them.
OCCDEC
OCCDEC works internationally with AI and deep learning algorithms to contribute to a society where people, regardless of language and background, can feel safe. In the pilot case, an artificial neural network was trained to recognize different types of sounds in an urban environment. The aim was to detect movement patterns and abnormal sounds, mainly for crime prevention purposes.
Ölands Köksmejeri
Ölands Köksmejeri produces various types of cheese. During the spring, the business received guidance in using image recognition to read the labels of the cheeses as a step in further quality assurance of deliveries. The pilot resulted in a cheese label classification using a self-trained machine learning model. The model is based on a pre-trained artificial neural network that was fine-tuned with pre-classified cheese labels to ensure that the right cheeses are delivered.
Itfitsu
Itfitsu offers its customers a digital fitting room to simplify the sale of second-hand clothes online. They have developed an app with the digital fitting room in the form of an MVP. In the pilot project, we have collaborated to further develop and refine the model behind the digital fitting room.
Victory Pipes
Victory Pipes manufactures carbon fiber tools for remediation, maintenance, and inspection at high altitudes from the ground. The original plan was to explore the possibility of creating customized instruction manuals. However, during the project, the objective changed to investigating solutions for translating the manuals, particularly into German.
Qvanti
Qvanti makes restaurants, hotels, and cafés more efficient by combining purchasing, receiving, inventory, and recipe calculations in one platform. The pilot project started with developing a model to predict food waste at hotel breakfasts. Due to the complexity of the project, the focus changed to predicting food consumption in restaurants. We also examined the impact of external factors such as weather conditions, nearby events, timings, and special events.
