Predict why your CV/Resume will be interviewed or rejected, and create cover-letter

Summary

Writing a cover-letter is often a waste of time, and customising a CV/Resume is harder (but more valuable). But what do Recruiters ACTUALLY want to see? What do they read? What do they misunderstand -- and how do we change that before we contact them?

Useful/Productive

CUSTOM GPT:

Try it for free! (requires a FREE ChatGPT account)

https://chatgpt.com/g/g-VjCudyiXC-improve-your-job-application-and-cover-letter

Context

This is primarily for mid-to-senior roles in the tech industry – but the GPT isn’t specifically tied to them, and should extrapolate to other industries and seniorities relatively easily. NOTE however: I asked it to behave like a senior HR/Talent person, so it’s going to lean towards their typical approaches.

Approach

Looking back through my 15+ years experience as a hiring manager I saw that there’s an obvious difference between small companies and medium/large: the gatekeepers in Talent/HR. In larger companies these people – sometimes a chain of multiple teams/people – each have to ‘approve’ of your application before it ever arrives in the inbox of the hiring manager.

As a manager this can be frustrating: often you have no control or influence over these people or The Process, and any attempt to improve it – or just to be allowed to interview the best candidates for the role, in order to hire the best people – is rejected for internal political or budget reasons.

So then I went through my experiences of “what” the HR teams were selecting on, and “how/why” they evaluated them. Recurring themes were:

1. HR/Talent want a smooth logical “pipeline” (despite that never matching reality) where every candidate only “moves forwards” or “is rejected forever”
2. They run a suite of orthogonal qualitative metrics to attempt “unbiased” (although it’s never unbiased) evaluations
3. …but never more than 3-5 of them, since they don’t have time/interest/sufficient depth to do more
4. Each HR team/company has a unique subset of metrics they prefer to use – some use keyword matching, some try to ‘read between the lines’, some use quantitative checks. Unless you work in HR, you simply cannot know what they’re matching on.
5. Most HR teams, if you ask them to (and they have time to talk to you), are very happy to bump a candidate up/down based on any extra information that changes the scores.

I then attempted to recreate the process that an HR person is going through, and co-opt the LLM/AI to “think like my opponent”, and guide me on what they may be thinking.

Key techniques

1. Know what the technical terms are for the person you want the AI to pretend to be: in this case, the main one is “HR Business Partner” (catch-all term for high-skilled consulting/external recruiters hired to supplement HR teams).
2. Advise the AI on the expected process – words like “pipeline” and “shortlist or reject” – with a sequence of expected outcomes
3. (one of my favourite techniques in prompting) Mix “analyse” and “summarise”, and have the AI move back and forth between them *within one answer*. This requires some finesse, but once you get the hang of it you get much higher quality responses than if you ask for pure analysis, and much deeper responses than if you ask for pure summary.