Better Models, Better Content

Our red flag modelling and ranking performance has increased by approximately 20%.


This means higher quality content across the board, including…

  1. More red flags that matter

  2. Fewer red flags that don’t (i.e. false positives)

  3. More accurate Risk Scores

Better red flag ranking:

Red flags are now coded red, orange or yellow depending on severity. Red flags are high impact, recent flags like restatements, new investigations and CFO departures. Orange flags are red flags that are less recent, lower severity or flags for which the model is less confident like remediated control weaknesses or prior period regulatory investigations. Yellow flags are lower severity and/or lower confidence. Yellow flags tend to include items like early adoption of accounting policies and geographic risks. Learn more about the new model and our process here.


What hasn’t changed:

  • Categorization of red flags: We categorize red flags using a separate process based on keywords only. This process has not changed and still contains some errors. We are working on a more robust solution coming soon!

  • Artificial Intelligence is never perfect: Our processes are fully automated so there will always be some false positives, no matter how good we get. In the current version of the model, false positives occur more commonly in 20-Fs. The most common false positives relate to financial experts, governance matters and cost of goods sold.


PLUS: Screening (idea gen) just got MUCH faster


Find your next great stock pick or short idea with instantaneous screening results.


Missing an active Bedrock AI license?


Request a trial at bedrock-ai.com/contact-us.

101 views0 comments