September 24th, 2021

Boosted Insights Product Update: Heat Maps Give Instant Feedback On Your AI Investment Management Models

Written By: Fraser Abe

We know that asset managers are incredibly strapped for time. From the morning meeting at 7:00 a.m. until well after markets close, an investment manager’s day is non-stop. We have heard from investment managers that they know AI is interesting, but they’re not sure how they might integrate it into their process – not only that, they doubt they have the time in the day to spend fussing over models. 

Faster, easier AI for institutional investors

We at believe that artificial intelligence can (and should) make an investment manager’s life easier. Our product updates, which are informed both by client requests and the former portfolio managers on our team, always seek to make asset management simpler and better. We recently shared how our portfolio one-pager reports can help asset managers with idea generation and act as an early signal of actionable AI insights for equity capital markets. Our most recent improvement on the portfolio one-pager reports is the addition of heat maps, so an investment manager can tell – within seconds – how well their model is performing across multiple timeframes. 

We now showcase Excess Return and Hit Rate as heat maps. With heat maps, a manager can see how the equities in their universe – as ranked by our machine learning platform Boosted Insights – performed across 1 month, 3 month, 6 month, 1 year and 2 year time horizons. The colour coded guide allows asset managers to simply glance at their portfolio one-pager report to gauge how their AI model is doing. It serves as another way to save the investment manager time and resources – if the heat map looks good for their investment style, they can have confidence their AI model works for their needs. If not, it can be an early flag to reach out to Customer Success to see how we can help improve their results. 

Excess return

The excess return table showcases the average excess return per time period per star rating bucket (1 – 5 stars) for the user’s model. This view highlights the excess return, as measured by (stock return – stock beta * benchmark return) so users can see how their AI portfolio is doing on multiple time horizons at a glance. 1 star rated stocks (those that the machine doesn’t like) are expected to have low excess return (negative) and 5 star rated stocks (those that the machine does like) are expected to have high excess return (positive). It serves as a fast and efficient way to gauge performance of their machine learning model. 

Hit rate

The hit rate table shows the percentage of times that the star rating bucket (1 – 5 stars) had positive excess return per time period. It enhances understanding of the excess returns highlighted above and again gives an easy overhead view for the asset manager to understand exactly how their ML model is performing. Similar to excess returns, 1 star rated stocks (those that the machine doesn’t like) are expected to have a lower hit rate (less excess return) and 5 star rated stocks (those that the machine does like) are expected to have a higher hit rate (higher excess return).


Adding heat maps to our portfolio one-pager reports is a way for institutional investors to see the value in their AI models instantaneously. When using the portfolio one-pager report as an equity ideas screen for their morning meetings (the portfolio one-pager report highlights the top 10 / bottom 10 movers for their rebalance period, giving the manager ideas for their portfolio), the heat map can provide further conviction for the investment manager and any other stakeholders. To learn more about how we are making artificial intelligence easier, more explainable and simpler to implement, please book a demo with us.

Leave a Reply

Back to Blog