PitchBook launches new AI-enabled tool to predict success for a start-up
The VC Exit Predictor was trained on 46,000 observations from start-ups with known outcomes that was followed by testing the algorithm on more than 11,000 out-of-sample observations.
Can artificial intelligence (AI) do a better job than humans in predicting a successful exit for start-ups?
PitchBook, the venture capital and private equity database whose articles we frequently post in teknovation.biz, believes AI is up to the task and has launched something called the VC Exit Predictor. Obviously, VC stands for venture capital.
“The primary goal of the VC Exit Predictor is to objectively score companies based on all relevant and available information in the PitchBook database,” the company wrote in unveiling its latest innovation. “This release comes at a time when the incredible growth in capital flowing to start-ups has expanded the investment opportunity set, making it more time-consuming for investors to sift through all of the potential options. In the last 10 years, the number of actively VC-backed companies in the PitchBook database has tripled to nearly 120,000 at the end of 2022.”
According to PitchBook’s news release announcing the new tool, it is based on “training the model on 46,000 observations from start-ups with known outcomes” that was followed by testing “the algorithm on more than 11,000 out-of-sample observations.” That accurately predicted success (M&A and IPO) versus no exit outcomes at a rate of 75 percent.
In addition to the news release, PitchBook posted this less-than-four-minute video.