The realistic guide to using AI in your 2026 job search

AI has changed the mechanics of job hunting far more than it has changed the outcome. Applications take minutes instead of evenings, which means everyone applies to more roles, which means recruiters see more noise. The candidates who win are not the ones generating the most documents — they are the ones using the time they save to do the things that never scaled: research, referrals, and follow-up.

What AI is genuinely good at

Rewriting is the killer use case. Turning a messy description of your work into tight, measurable resume bullets is mechanical labour that a model does well and humans do slowly. The same is true for adapting a stable cover-letter core to a specific posting, or drafting a follow-up note you have been avoiding for three days.

It is also excellent at pre-interview preparation: paste a job description and ask for the ten questions most likely to come up, then rehearse answers out loud. That is free, unlimited practice against a realistic brief.

What it is bad at

AI cannot know your actual numbers, and it will happily produce confident-sounding placeholders that read as lies in an interview. It also has a strong pull toward generic phrasing — the exact quality that makes an application forgettable. Every generated document needs a pass where you replace at least one abstraction with a fact only you could supply.

A weekly workflow that works

Block ninety minutes twice a week. In the first half, shortlist roles and generate tailored bullets and letters for each — five applications is a realistic target. In the second half, do the unscalable work: find one human at each company, send a short message, and follow up on anything from last week.

Track everything in one sheet: company, role, date applied, contact, follow-up date. The tracker matters more than any single document, because job searches are lost to forgetting, not to writing quality.