Does Relevance AI learn from what I do?

Relevance AI allows you to modify values in your datasets. The modification is not limited to the original fields in your dataset (i.e. fields in the uploaded CSV). Through the Editor you can modify all the existing values including initial fields and the AI processing outputs (e.g. AI Tagging and AI Clustering).

Do my modifications affect the AI model and its next decisions?

The short answer, for now, is no. Training an AI model needs:

1. To save your data and keeping track of your modifications.
2. To provide every single client with their own specific AI model, trained on both their own data and modifications. Whilst this is on our potential roadmap, it will require infrastructure changes and time to implement at scale.

Note: Using similarity detection, our models are capable of suggesting similar items based on your modifications. But your actions or changes are not stored for training purposes and are only available within one browser session.

When tagging, will the tags and learnings from my first project apply to my subsequent projects?

As explained above, the AI model's behaviour and output are not changed based on your modifications. And under the current layout, datasets are independent of each other. So, when you prefer to use a pre-existing code frame (as opposed to an AI-generated code frame), simply copy-paste the comma-separated list as explained here.