Have data that could support AI?
- Collection
- Eligibility
- Result
What kind of work do you do?
Supplier guide
Tell us what you have and how it was created. Sapien1 can help data owners explore whether a collection could support AI training, evaluation or research.
Start with a short description. It can cover an existing dataset, records from a business workflow or a collection you could create.
Describe the collection and its context
Useful starting points include documents, code, structured records, audio, images, video and examples of how a task is completed. The context around a record can matter as much as the record itself: what happened, which instructions applied and how the result was assessed.
For example, a software collection might include source files, change history and tests. A workflow collection might include a request, the relevant documents, the actions taken and the resulting work. Describe what is present and what is missing.
What to include in your first email
- What the data represents and how it was collected.
- Its approximate size, format, language and time period, where known.
- Your relationship to the data and any known limits on sharing or use.
- What you would like to explore, such as licensing an existing collection or creating data for a specific task.
“Not sure” is a useful answer when a detail needs investigation. Keep the first email to a description; we can discuss samples and an appropriate access method after reviewing the fit.
For help preparing the description and identifying questions about permission, see what to prepare before licensing your company's data.
What happens next
The first conversation establishes the possible use and the questions that need answering. If there is a fit, the next step is to agree on an assessment: which sample to review, what permissions need confirming and what quality checks matter.
The scope and terms of any further work are agreed for the project.
Looking for data for your own AI project? See data sourcing for training and evaluation.