Launching today
Badge
AI agents collect peer reviews to generate proof of work
553 followers
AI agents collect peer reviews to generate proof of work
553 followers
Badge gives an AI agent that collects peer reviews to build proof of work. For job seekers - The agent connects to your contacts, requests anonymous reviews from your past colleagues, and builds a verified, portable trust score. score. For hiring managers - Badge gives you a 30 second reference checks on candidates, collecting authentic feedback from their actual coworkers to replace unreliable, AI-generated resumes and LinkedIn Recommendations.
How does the agent actually get past the awkwardness of asking former coworkers for reviews without feeling spammy, and is there a way to control who it reaches out to?
Badge
@orhanmentet8qe Great question - When you give the agent the phone book access, it finds the people who work in the same org as yours, so the natural boundary becomes the org.
How does it actually verify that the reviews come from real past coworkers and not just friends the candidate adds on a whim?
Badge
@sedagoncal THat's a great question Seda, we use work email authentication and every review gets a blue tick, here we check that the review giver and reviewe both work in the same organisation, however we are still working on features where we know that they have actually collaborated in the same projects, What we have figure3d out is that AI agents asks nuanced questions - which becomes difficult to answer when you have not worked with the person
termique
Congrats on the launch @lokesh_motwani1 ! Really like the framing here, portable proof-of-work instead of resume claims that anyone (or any AI) can write.
Curious how you're handling the cold-start problem: for a candidate's first few reviews, how do you make sure the peers reached out to actually give honest feedback and not just do a favor for a friend, especially since it's anonymous?
Badge
@rafihalilintar Rafi, Thanks for this question. For the cold start problem we have done pilots and for the honest feedback - this was the the most challenging part. and AI agent helps there, It probes. So first it asks binary questions and just like humans, it then probes into details and the review quality comes from the the details the reviewee gives in their answers. The agent here works just like any HRBP works in the org. it understands, it probes and thats how we can figure out which review is fake and which is not
Interesting concept. How do you handle verification of reviews and prevent fake or biased feedback from affecting someone’s Trust Score?
Badge
@vishalmehta8340
Vishal, Great question
So the first thing we have done is to add work email verification - verified reviews have an impact on trustscore more than unverified ones. We send OTP to work email to check if the reviwer works in the same org as of the reviewee. One can also argue that the fake reviews can come from people from the same org, but then org agent probes reviewer with sepcific questions, so the reviewer needs to know details to give a review. The quality of these review also feeds into the trust score calculation
In future, we are building these feautres - you can verify your prev org by giving us your old experience letters. I hope this answers the questions
GrowMeOrganic
How does Badge prevent coordinated positive reviews between close friends or former teammates??
Badge
@iamanantgupta Great question Anant. Our AI agent acts as a moderator and will ask really detailed instance based questions to the reviewer. The reviewer cannont post a review just like you do for Google places. Agent acts as a moderator and it will create frcition for the fake / cordinated reviews
Badge
@iamanantgupta Badge is designed to prevent cherry-picking. Rather than only allowing candidates to hand-pick who reviews them, Badge's AI agent works independently. It pulls data from your contacts to identify people you've actually worked with and reaches out to them for a review.This means the candidate has no control over who gets asked.
Congrats. How do candidates respond to critical anonymous feedback? Is there a way to provide context or growth updates?
Badge
@ragsyme Based on the "how toxic the review is" badge, filters out the super-toxic reviews and does not make them live. For the other reviews, typically the candidate can choose to hide up to 3 reviews but it will still affect the trust score.
Badge
@ragsyme Great question - We will take the growth updates part of the feature request. So far the agent does not provide this to the user but it can definitely do this. Thanks for this Raghav!
Making references machine-collected instead of candidate-picked fixes the cherry-picking problem, but it moves the weight onto people who never signed up: the agent reaches them through a contact book, and their honest take on an ex-colleague becomes stored data. Can a reviewer later see, edit, or delete what they contributed, or does it live in the candidate's trust score for good?
Badge
@vollos They definitely can do all of the above but there is a 90 day cooldown period before they can make any changes in order to prevent abuse.
@raaghav_naraayan_m_v Makes sense, a cooldown is a smart way to stop people gaming their own record. Appreciate the straight answer, good luck with the rest of the launch.
Badge
@vollos Thank you!
Badge
@vollos TO give a review the colleague need to sign up on the app. Reviewer can delete the reviews but the trustscore as of now does not change when you remove the review.
@lokesh_motwani1 Fair enough on the sign-up point, I had that part wrong. And being upfront that deleting a review doesn't roll the score back is more transparency than most teams would give this early, so credit for that. Good luck with the launch.