We're running into issues where people search returns a job title that is incorrect based on reviewing their actual linkedin profile. Any tips on reducing these inaccuracies? We're using an ai column leverage gpt 4.0 to confirm the job title by consulting the linkedin profile but looking for a better method.
Hey Nicolas B. the best way is to pull the current role from linkedin instead of just the headline, then cross-check with company and department info i usually add an AI check to flag mismatches and assign a confidence score only override titles when confidence is low.. this keeps errors minimal and saves a lot of guesswork.
Yo Nicolas B. The most reliable way is to use Clay LinkedIn Profile Scraper first and pull the title directly from the profile then use GPT only to clean or normalize it If you scrape the profile first and treat AI as a cleanup step instead of a fact checker the job titles stay accurate.
We use people search -> obtain linkedin profiles -> use ai to verify job title of linkedin profiles generated by previous step.
I'll look into apify Bharat D., beyond that it seems like using ai columns to verify the linkedin profile associated with the people search is the way to go.
Cheers 🙂
Nicolas B. Any issues asking a Claygent to double check the profile? Thats where I have challenges. LinkedIn is good at stopping it.
Laz G. find people itself doesn’t guarantee department or headline because it only returns what the underlying source exposes If you want those fields consistently, the correct approach is to chain an enrich person step after find people and feed it the linkedIn URL enrich Person is built for structured profile data, so it returns department, headline, seniority and team info even when find people can’t
