Paper Number
1072
Paper Type
Complete Research Paper
Abstract
The venture capital (VC) industry has been experiencing a substantive transformation in recent years due to the rise of online sources for startup data and their utilization in identifying promis-ing startup founders. The resulting significant increase in leads has proportionally elevated the workload required for screening them along the investment process. Together with a well-known German VC firm, we have designed and implemented a Machine Learning pipeline trained on a large dataset of 39,114 LinkedIn profiles with their respective screening decisions aiming to en-hance the efficiency of the process. Our method boosts screening productivity by auto-rejecting 23% of founders, missing at most 1% relevant ones, yielding a 663:1 accurate rejection ratio. Venture capitalists can adjust auto-rejection up to 57% at 10% miss rate. 84% of users prefer the AI-augmented workflow, while 16% pre-filter profiles for reduced workload. Its success convinced the VC firm to promptly implement the system in production.
Recommended Citation
Maurer, Marc; Buz, Tolga; Dremel, Christian; and de Melo, Gerard, "Design and Evaluation of an AI-Augmented Screening System for Venture Capitalists" (2024). ECIS 2024 Proceedings. 7.
https://aisel.aisnet.org/ecis2024/track03_ai/track03_ai/7
Design and Evaluation of an AI-Augmented Screening System for Venture Capitalists
The venture capital (VC) industry has been experiencing a substantive transformation in recent years due to the rise of online sources for startup data and their utilization in identifying promis-ing startup founders. The resulting significant increase in leads has proportionally elevated the workload required for screening them along the investment process. Together with a well-known German VC firm, we have designed and implemented a Machine Learning pipeline trained on a large dataset of 39,114 LinkedIn profiles with their respective screening decisions aiming to en-hance the efficiency of the process. Our method boosts screening productivity by auto-rejecting 23% of founders, missing at most 1% relevant ones, yielding a 663:1 accurate rejection ratio. Venture capitalists can adjust auto-rejection up to 57% at 10% miss rate. 84% of users prefer the AI-augmented workflow, while 16% pre-filter profiles for reduced workload. Its success convinced the VC firm to promptly implement the system in production.
When commenting on articles, please be friendly, welcoming, respectful and abide by the AIS eLibrary Discussion Thread Code of Conduct posted here.