From a prompt to an agent that can carry a task through to the end.
We test how a model can plan, use tools and ask for human intervention exactly when it is needed.
Read in the AI LabAI LAB / IN PROGRESS
Artificial intelligence, applied.
Experiments, prototypes and conclusions about where artificial intelligence genuinely becomes useful.
AI JOURNAL
Ideas tested, not just discussed.
Articles, experiments and notes on what works, what does not, and what deserves to become a product.
Connected knowledge, with answers grounded in real sources.
A RAG prototype for an organisation's internal information, with verifiable sources and control over access.
Read in the AI LabInterfaces that adapt to intent, not just to the screen.
What changes in design when the interface can generate the right structure for the user's task.
Read in the AI LabHow we measure whether an AI prototype deserves to become a product.
Quality, speed, cost and control: four criteria before integrating it into a real process.
Read in the AI LabWe start from a clear use case and build the prototype that tells us whether it is worth continuing.