Right now, AI is applying for the same jobs you are.
Every cover letter mirrors the job description back at the hiring manager. Answers that sound good but say nothing. Generic applications that could belong to anyone.
Hiring managers are starting to recognize the pattern, and they're using it as a filter.
The candidates being chosen are using these tools differently. Not to do the work for them, but to see their own work more clearly.
Every tool we're covering today can write your cover letter, draft your interview answers, compose your follow-ups. But you shouldn't let them, at least not the final version.
These tools work best the way a good colleague works best. The one who reads your draft and says: "I don't know what this means, and I've known you for ten years." The one who listens to your answer and says: "You lost me after the first sentence." Use them that way. Let them push you. Then you write it (or rewrite it) yourself. In your voice. With your stories.
That's what gets you hired. You. Your stories, your voice.
Your application makes perfect sense. To you.
You sit down to write a cover letter, and you write something like: "Led multi-stakeholder coordination across five country offices." It feels specific. You can picture it all—the room, the calls, the forty-message email chain before anyone agreed on a workplan.
To a hiring manager who's never worked in your world, it's just more noise.
After fifteen or twenty years inside one system, it's normal to lose the ability to see your own experience from the outside. Not because you're not perceptive, because you've been living it too long. The language you developed to describe your work is precise. It just doesn't translate.
Claude Projects (free tier available, but honestly worth the investment to pay for) is where that translation happens. If you set up a project last week with your CV and career documents, upload the job description you're considering. Then ask: "Based on my documents, what are the three strongest points of alignment between my experience and this role? What's the biggest gap, and how should I address it?"
What comes back might surprise you. The strengths it highlights won't always be the ones you'd have led with. The gaps might not be what you expected. That's the point, it's showing you how your experience will be received and where to make the changes.
Then write the final version of your cover letter. Yourself. When you have a draft, paste it back into the LLM and ask: "Flag anything that sounds generic, vague, or like it could describe any candidate. Also flag anything that sounds AI-generated."
AI is good at catching its own patterns. And that says everything but means nothing tone? It's also what your writing defaults to at midnight when you're exhausted and just want the application done. (We've all submitted something at that hour and regretted it by morning.) Catching that before you send it is a small thing that can change everything.
One rule: never ask it to "improve" or "rewrite" your text. What comes back sounds like every other application in the pile. Ask for feedback. And then you make the decision about what to change. The difference matters more than you think.
You're better at interviews than you realize. Just not these interviews.
You're probably excellent at interviews at the UN or whatever industry you worked in. Competency-based questions, structured panels, scoring matrices, you knew the format, and you prepared for it. That's a real skill.
It also doesn't transfer the way you'd expect.
A private sector hiring manager wants to know two things in about thirty minutes: can you do the job, and would they want to work with you? It's less structured. More conversational. There's no initial matrix, no panel taking notes. And the openness that sounds easier is exactly what makes it harder, because you're used to knowing the rules, and here the rules change with every person you speak with.
ChatGPT Advanced Voice Mode closes this gap faster than almost anything else available. You have a real spoken conversation, natural voice, real-time responses. The free tier is limited; the paid plan is where it becomes genuinely useful.
Tell it the role, the company, the format you're expecting. Then: "Ask me interview questions one at a time. After each answer, give me specific feedback on what was strong and what I should improve. Be direct, even if it's uncomfortable."
Then answer out loud. Don't type it. Don't rehearse it in your head sitting on the sofa. Do it out loud, the way you'll have to when it actually counts.
Yes, sitting in your kitchen talking to your laptop may feel strange. (It will, for about ninety seconds. Then you forget you're not talking to a person, and something useful starts happening.) You hear yourself using jargon that means nothing outside your sector. You notice your answer ran four minutes when the question needed two sentences. You catch the moment you went abstract when what they wanted was a specific story about the last time you actually solved something.
Then ask it to switch styles. A casual screen with HR feels nothing like a technical conversation with a department head who wants to see how you think. Practice both. Practice until you're comfortable with the format. When you're finally across from a real person, you want the nerves to be about the stakes, not the lack of familiarity.
Free alternatives: Microsoft Copilot Voice is good. Gemini Live also works well. There are also a number of paid sites.
The financial conversation nobody taught us to have
In the UN, you had a grade and a step. Post adjustment, education grant, tax exemption. There was nothing to negotiate. The number was the number. You might not have loved it, but you never had to ask for it.
Outside, everything is different. The same job title pays dramatically different depending on the company, the city, the industry, and how the conversation goes. People with extraordinary experience and real skills accept offers well below what the role is worth. Not because they don't deserve more. Because they don't have data, they don't want to seem difficult, and they've simply never had to do this before.
If that sounds like it might be you, it's almost all of us.
Deep Research Models (Perplexity, Gemini, ChatGPT, or Claude) means you walk in with numbers instead of guesswork. Try this:
Research average salaries for [target role] at [type of organization] in [location]. Account for organization size, years of experience, and sector. Include base salary, typical bonus structure, and benefits context. Present as a table comparing entry, mid, and senior levels.
Run it for every role you're seriously considering. Compare across tools, they pull different results, and the differences are useful. You're building a range, so that when someone names a figure, you know whether it's fair or whether you're leaving something on the table.
Then practice. Go back to voice mode: "I've received an offer for [role] at [salary]. Based on my research, I believe the range is [X to Y]. Play the hiring manager. Push back so I can practice responding."
If you've never negotiated your salary before, it can be overwhelming. And like everything else we're working through, it gets less frightening with repetition. By the third round or so, something changes. You stop apologizing for asking. You start stating what you're worth like it's a fact.
Because it is.
None of these tools do the hard part. They won't write your story, build your network, or sit in the interview chair for you.
But they change how we look at things. The research that once felt impossible becomes a report in two hours. The interview format that used to bother you becomes something you've already done, quietly, in private, before anyone was watching.
That's a lot. Particularly when the weeks are starting to blur.
Better together,
Andrea
P.S. This piece is from The Career Reset, my newsletter for UN and INGO professionals figuring out what's next after funding cuts. If the job search part sounded like where you are right now, that's what the whole newsletter is about. When you sign up, you'll also get a free Career Diagnostic: a short tool that shows you where your experience is strongest, where the gaps are, and where to focus first.