iCONECT AI Predict and iCONECT Oversight are two new AI tools in Version 10. I don’t have an analogy from my life that I can use in this post because this is something I haven’t really seen before… like ANYWHERE.
Both AI Predict and Oversight start with the review of at least 50 documents by someone (typically a case expert) who knows the case well. These documents could be already “known” to the expert and bulk coded, or if they are not familiar, these documents can be coded individually. The individual coding is pretty awesome. The expert is shown one document at a time: they make a decision on the document and the system brings back the next best document to train the model. (no rounds or annoying time delays between documents) With each coding decision, the model is being refined in real time.
Once the model is trained, it can be checked as often as needed with a few clicks. The AI Model Refinement will show any documents the model isn’t sure about – giving the expert a chance to revisit them quickly before moving on to the next step.
Once the model is set, AI Predict or Oversight can start.
AI Predict applies a relevancy ranking for each record against each issue, and soft checks issues that are over the responsive threshold set for the project. This allows reviewers to concentrate on documents that are likely relevant first, speeding up the review of responsive documents.
Oversight starts tracking the linear review of the documents and provides up to date metrics on user and issue coding ‘potential errors’, which can help a review manager keep the case on track and identify issues that might required additional reviewer training – from the first few documents tagged.
These are “you’ve got to see it to believe it” features! Reach out for a demo, ‘cause this is next-level cool.
Want more? Book your demo below!