LenzIndependent, multi-model fact-checking API for AI workflows
Promoted
VIRA — Find influencers who actually fit
AI ranks influencer fit so brands stop guessing
4 followers
AI ranks influencer fit so brands stop guessing
4 followers
VIRA searches scraped Instagram profiles and uses AI to score each creator against your specific campaign. No cold outreach. No guesswork. Localized to Melbourne.
How does the AI scoring actually work when my campaign brief is pretty niche, like a sustainable fashion drop with a tight aesthetic vibe, rather than just basic follower counts and engagement rates?
@ayfermerca8493 Great question, thank you! For now it works better than follower counts, but it's not magic for niche aesthetic briefs yet.
The model sees your full campaign description, the influencer's bio, their category, followers, and engagement rate. For a niche brief , it does simple semantic matching. If you write "sustainable drop, slow fashion values, minimal aesthetic", it'll score higher on influencers whose bios contain that language. Someone who writes "slow fashion advocate | ethical wardrobe" will rank above "fashion | lifestyle | streetwear" even if follower counts are similar.
The honest limitation that I'm still working on is matching actual content aesthetic. An influencer with perfect visual fit but a generic bio might score lower than someone who writes the right words but posts differently. The match reasons in the UI show you exactly what it matched on, definitely worth reading to calibrate how much weight to give each score and leaves the ultimate decision on the hands of our users.
Some practical tips I found for specific briefs is to write a detailed prompts. "Sustainable drop, neutral tones, considered minimal aesthetic, no fast fashion" gives the model real signal. The more niche language you use in the brief, the better the matching.
Richer data (post captions, hashtags, visual content) would make this significantly better for aesthetic-driven briefs, that's on the roadmap in our future updates! For now the score is most useful for filtering out obviously bad fits quickly, then you make the final aesthetic call yourself.