AI Lab: AI solutions for businesses

We help your organisation put AI to work as a colleague that handles repetitive or time-consuming tasks. We develop AI agents for these tasks and help you deploy them. We also help you use AI when working with sensitive information.

We plan the implementation based on what the agent needs to do, the information and specialist knowledge it requires, and how much it will be used. We also consider the sensitivity of the information it handles. We help you deploy agent solutions that use locally hosted language models, as well as other types of agent solutions. These solutions can incorporate retrieval-augmented generation (RAG), which allows the agent to retrieve information from selected sources to support its responses.

We start with a needs assessment workshop and pilot testing. We set up the necessary operating environment and implement the first agents for testing. Our ongoing services include maintaining the operating environment in accordance with the service agreement and implementing agents according to a plan agreed with your organisation.

We develop machine learning algorithms for integration into your company’s products and solutions. Your company owns the intellectual property rights (IPR) to the algorithms, so the know-how and competitive advantage stay with you.

1. Initial experiments with the available data.
2. A workshop to agree on objectives and testing arrangements.
3. Data collection, including measurement arrangements and data preparation.
4. Model development in the AI Lab environment. Developing the model may require additional data collection, in which case we return to stage 3.
5. Review of the results and planning the next steps based on feedback.
6. Optimisation and implementation in your company’s system.
7. Prototype testing and fine-tuning until the requirements are met.

When product requirements or user needs are not yet clear, we work with you to clarify them and build a working prototype. This support service lays the foundation for the subsequent development of a machine learning algorithm.

We use design thinking to deepen customer insight, define requirements and build a prototype. The method includes the following stages:

1. Understanding: Who are the users, and what matters in their work?
2. Defining: What problem are we solving, and what requirements must the solution meet?
3. Generating ideas: What kinds of solutions could meet the need?
4. Prototyping: What could the first version of the solution look like?
5. Testing: Does the solution work for its users?