Some questions deserve an experiment.
DKG Labs gives AI exploration and new ideas a place to become something we can test.
For teams, builders and specialists with a practical question about AI, a workflow or a new way to use information.
01 / The question
Begin with something worth finding out.
A useful experiment starts with a specific question. Define:
- The person doing the work and the task they need to complete.
- What is difficult today.
- What a useful result would look like.
That gives the prototype a clear boundary and a result to examine.
02 / The experiment
Make the question tangible.
Build a focused prototype around the task, its inputs and how a person will review the result. Trying it can reveal useful behaviour, limitations and requirements that discussion alone could miss.
Areas of exploration
- AI and human review
- Test where a person needs to check an AI-generated summary, comparison or response.
- Handoffs between teams
- Explore what the next person needs to know, when they need it and how that information reaches them.
- Finding information across documents
- Try ways to locate and compare relevant passages while keeping the evidence behind an answer visible.
- New interfaces
- Try a different way to present information or guide a task, so people can show what helps and what gets in their way.
03 / The learning
What did the experiment show?
Compare the result with the original question, then record three things.
- What helped.Where the result was useful to the person doing the task.
- What fell short.Inconsistent responses, unclear information or practical constraints.
- What remains unknown.The questions this experiment could not answer.
The next conversation
What would you like to find out?
Tell us about the task, the people involved and the question you want to explore. An early idea is a useful place to begin.