Opinion · the twin moment
The digital twin wave has a trust problem
Digital twins for job seekers are having their moment. Fortune covered Eightfold's Ashu Garg describing TwinCafe, where job seekers upload a résumé, a LinkedIn profile, and work samples to create a digital version of themselves. Recruiting thought leaders are asking, in earnest, whether recruiter agents will soon prescreen job-seeker agents. The effort asymmetry that idea attacks is real: candidates pour hours into applications that companies screen in seconds.
I build in this space, and I think the twin framing gets one thing importantly wrong.
An agent that answers as you inherits every deception problem
A twin is designed to be mistaken for you. Even when everything it says is accurate, the recruiter on the other side has to wonder: am I talking to the person or the machine? Did the candidate really say this? The moment that question exists, trust leaks out of the conversation — and hiring runs on trust.
An agent answering about you, grounded in your real material, earns trust that one answering as you never will.
The opposite shape works. An openly-AI representative that answers about the person — grounded only in material they actually provided — is something a recruiter will happily interrogate, because nobody is being fooled. There is no uncanny moment, no disclosure question, no "was that really them?" It is an assistant, it says so, and it knows the person's work cold.
What this looks like in production
This isn't a thought experiment. A proxy is an AI that answers questions about you, so people ask instead of skim. It learns about you from the web and the materials you upload, then interviews you so your story goes broader than any document. Visitors ask it anything. Recruiters paste a job listing and it analyzes the fit against the person's real background — honest gaps included, because a representative that hides gaps is just a twin with better manners.
Live example: mine answers for me every day at askproxy.ai/christopher. Ask it something hard, or paste a job listing at it.
The agent-to-agent future is half here
Will recruiting agents prescreen candidate agents? In one direction it already happens: AI crawlers and assistants read proxies today (they're served structured, open context), and the job-listing fit analysis is exactly a prescreening conversation — just one where the candidate's side is honest about what it is. The missing half is protocol, not possibility. When recruiter agents start asking, the candidates who win will be the ones whose agents were built to be questioned, not to pass as human.
Build yours free — live in about five minutes: askproxy.ai.