Get Started
Log In
Get Started
Log In

4 Critical Questions to Ask Vendors When Shopping for a Screening and Vetting Tool

by Kristina Drye, on Aug 17, 2021 4:33:21 PM

Questions for Vendors

Choosing the right tool for your screening and vetting practice can be a daunting task. During the process, you’re likely to consider a variety of vendors. Knowing which vendors are trustworthy and which tools are helpful can be difficult when you’re surrounded by so many options. 

By using the four questions below, sifting through the noise to find the valuable tools can be easy. If your vendor can answer these questions, and they spend time with you to understand your data, you can be assured the team and the tool have legs as a solution to your screening needs. 

Vendors who can’t answer these questions or don’t spend time with you to understand your data are vendors you need to cut from your shortlist.

Below, we list the questions and an explanation of why they are important. We also answer the questions from the perspective of a member of the Giant Oak team.

Important Questions to Ask Your Vendors:

  • What data do you use to train your algorithms? 

A machine learning (ML) and artificial intelligence (AI) tool is only as good as the data it’s trained on. You should look for a vendor that can explain what data is used to train the algorithms and where it came from. 

GOST (Giant Oak Search Technology) Vendor: Algorithms for GOST are trained on datasets created by professional data scientists. The algorithms are trained on a set of identified “bad” - depending on the use case - and continue to improve as they are used by the customer. All algorithms are uniquely customizable to optimize the user’s results.

  • Can you explain AI and ML to me? 

Vendors who have a good understanding of these ideas will be able to explain them to you without resorting to buzzwords. Even if you know what AI and ML are - you can learn a lot about a vendor by listening to how they explain the concepts to you. 

GOST: In short, machine learning is the technical basis for data mining in big data. The process of machine learning is inductive, and it needs a lot of data to become precise. Artificial intelligence is the agent that perceives and takes action, building on the machine learning component. 

  • How do you test your tool for bias over time? 

ML systems don’t have human judgment and will reflect the bias of training data. Make sure you understand how your vendor tests and controls for errors. 

GOST: The team of scientists at Giant Oak is constantly testing GOST in its application. Our Customer Success specialists work with each client to monitor GOST’s performance, and our scientists perform checks and critical audits of the tool on a routine basis.

  • What tasks does your technology automate that most humans do now, and what tasks does your company believe must remain in the hands of humans? 

Understanding the role of the human in an AI world is critical. Your vendor should have a vision of what tasks should be handled by machines, now and in the future. Once they describe this to you, you should ask yourself: do you agree with their approach? 

GOST: GOST takes the manual screening and vetting process and renders it almost instantaneous. The time that a professional might use to search Google, Bing, or other internet providers; or scan static lists is replaced by GOST. GOST then provides a risk score to each entity, allowing for fast triage and prioritization. GOST does NOT adjudicate risk, nor does it make decisions about what is “good” or “bad”. The adjudication is left solely to the human professional using GOST as a tool to augment their processes.


Recent Posts