Clear Talent, LLC was acquired by Patreon in 2023. Before then, they worked with us to scale their operation team with automated decision making. Read the case study to understand how to leverage AI to automate your business.
Before being acquired by Patreon in 2023, Clear Talent, LLC was a recruiting companyspecialized in recruiting software engineers. The CEO and founder had a team of about 40people. The work was broken into sourcing, and hiring. Sourcers would find potential goodcandidates for a role, and a recruiter would message that person. In 2022, they hired Build TestRepeat to automate their company.
Automated decision making is when your product makes decisions a human would usuallymake in a business. This is a form of AI usually called machine learning but can be simpleralgorithms as well.
ADM in a nutshell automates the work of a business analyst. It allows us to make an analysisand decision at a high scale - maybe multiple times a day as opposed to a weekly or monthly oreven quarterly decision stemming from an analyst’s work.
A good example of ADM is a product recommendation system. An analyst would look at allcombined purchases per user to identify “usually bought together” products. Then theseproducts can be recommended under every product listing. How often should that list update? Isit once a month? Once a week? Maybe multiple times a day? The level of automation heredepends on the scale at which users make purchases. If you are amazon, then it’s within 5minutes and they are well known for that. If you have only a few products and customers, thencomputing a list of best sellers once a quarter or year will be the right level of work
In recruiting, the work of an analyst is to best match the type of candidate experience to an openposition. They develop a strategy for an open role which is crucial to drive your sourcing. InADM we call that strategy a “model”.
Instead of going at it manually, we came with a way to automatically build these models. Webuilt a chrome extension which would have a drop down to select the position you are sourcingfor: say “iOS engineer at XYZ company”. When a sourcer finds an appropriate candidate, theyclick a button to upload the candidate’s profile and attach it on our system to that dropdown roleselected. We would then parse out all the titles that candidate has, the year of experience, andmore proprietary signals. And these signals would be associated with a type of job like the onethe source selected. Thus we are creating models automatically based on the sourcer’sexpertise without impacting their workflow.
After human review of all the signals, we would approve the model, and then apply it to everycandidate that we have ever encountered to quickly find all the eligible candidates.
Once you find success in automated decision making, it’s about creating more of it. If somethingworked, we kept it and improved our targeting, if it didn’t we moved on to the next. Everytime wemade a hire, that would be used to confirm the model’s signals.
If business analysts, or product managers do critical work to kick off a process or operation inyour company on a day-to-day basis, then that is a great hint that you could use automateddecision making software. Their talent can be 10Xed when used in conjunction with software.Book a consultation with us, and we can explore solutions together.
