AI can save considerable amounts of your time. However you have to be careful and not forget that Large Language Models are nondeterministic by nature, ie. asking the same question twice yields two differing answers. We use AI in two ways: To support and speed up development work, and to enhance customer solutions.

Our development process benefits from AI through shortened development time, extended test cases and improved documentation. We use AI to generate plans, layouts, project structure, and test cases. We then leverage AI tools to accelerate coding and testing. Human approval is always required before any code is submitted for delivery.

There are limitless possibilities to enhance customer solutions. Many organisations start with an inhouse knowledge base that staff personnel could query. This structure is particularily useful with nonstructured data. Or, the internet application monitoring and logging could be analyzed by AI, to find usage patterns or bottlenecks - or potential security breaches.

We believe that the most secure way of running AI is to lock it inside your own firewalls. Improved language models and algorithms make this feasible while keeping costs on a sane level.