I build Forte, so I have a bias. I'll flag it up front, then try to be fair about the landscape.
What works
- Matching as a ranker, not a filter. Tools that surface the order in which to apply — rather than deciding whether a role is "a match" — respect the fact that the applicant is the ultimate judge. Ranking compounds; filtering just hides.
- Résumé tailoring with constraints. Tailoring absolutely works when the model is constrained to your real experience. It doesn't work when it's free to invent. Tools that advertise "10x more interviews" without explaining how they avoid hallucination are not serious.
- Skill self-assessment. Short diagnostic quizzes (not marathon tests) calibrated against the specific stack you claim — these are surprisingly useful both as a pre-interview warm-up and as input to a smarter matcher.
What's hype
- Auto-apply at scale. You are not beating the recruiter's spam filter; you are training it to mark your domain. I have watched friends apply to 600 roles via auto-apply tools and get 2 callbacks. That's a worse conversion than sending 20 by hand.
- "AI-optimised for ATS." ATS systems are keyword grep and TF-IDF. You don't need GPT-4 to pass them; you need to not shoot yourself in the foot with a creative PDF layout. Most "ATS-optimised" claims are boilerplate marketing.
- Interview bots. The category exists because first-time-seekers are anxious, not because the bots are useful. A mock interview with a friend who works in your target role is worth ten of these.
What I'd build if I weren't building Forte
A weekly digest that tracks the five companies I'd most want to work at, and tells me when a role opens up that matches me above some threshold. I suspect it'd have a 5x response rate over generic daily scans — because the signal comes from the target list, not the feed.
(Forte's "company focus" feature is exactly this. It's in the roadmap for the next release.)