I've never used AI before. Where do I start?

Open ChatGPT or Claude in a browser today. Both have free tiers, no card required. Ask it to help with one real task you'd normally do yourself — a personal statement, a job application answer, a summary of your lecture notes. That single 20-minute session teaches you more than any intro video.

Is it too late to learn AI in 2026?

It's the opposite. Two years in, the tools are stable and the interfaces are graduate-friendly. Employers are only just starting to filter for AI-literate hires, so the gap between 'people who claim AI skills' and 'people who can prove them' is at its widest. That gap is your opportunity.

Is AI safe to use for my job applications?

Yes — with judgement. Use AI to draft, restructure and pressure-test your answers, not to submit unedited output. Every application system now has AI-detection built in; anything that reads like uncut AI output gets flagged. The winning move is human voice + AI leverage, not one or the other.

How much does it cost to get started?

Nothing. ChatGPT, Claude, Gemini, NotebookLM and Lovable all have free tiers that are enough to build a real portfolio project. You only need to pay (~£15–£20/mo) if you want longer context windows, image generation or faster models. Start free; upgrade only when you hit a specific limit.

What's the first useful thing to build?

A custom AI helper for your own job hunt. Feed it your CV, the target job description, and the company's about page. Ask it to draft tailored cover letter openings, predict likely interview questions, and mock-interview you. It's useful, it's low-stakes, and it becomes the demo you show recruiters.

I feel behind everyone else. Am I?

No. Surveys consistently show that under 30% of UK graduates use AI systematically for job applications, and under 10% have built anything with it. The 'everyone else' feeling is a Twitter/LinkedIn distortion — the highly online users are 1% of the applicant pool. You're closer to the front than you think.

New to AI? The first-steps guide for UK students and graduates

The 20-minute first session

Open one of these in a browser, right now. Free, no card needed:

chat.openai.com — ChatGPT. The default. Best all-rounder.
claude.ai — Claude. Better for long documents and careful writing.
gemini.google.com — Gemini. Best if you live in Google Docs and Gmail.

Pick one. Paste in a real thing you're working on — a personal statement, a job application question, a lecture summary you need to write. Ask it to help. Ask it follow-ups. Notice what's good and what isn't. That's your first session done.

The four-week ramp

Week 1
Use it every day, for anything
Meal planning, essay drafts, code you don't understand, career-choice pros/cons. Volume matters more than quality here — you're building intuition for what AI is good and bad at.
Week 2
Learn the four-part prompt
Role, task, context, format. Read one short prompt-engineering guide (Anthropic's or DeepLearning.AI's — both free, under an hour). Rewrite three of last week's prompts using the structure. Notice the jump in quality.
Week 3
Try a Custom GPT or Claude Project
Build a persistent AI helper for one specific task — job-application drafter, essay feedback bot, revision coach. This is your first real "thing you built." Total time: a couple of evenings.
Week 4
Build a public tool
Move up from a Custom GPT to a Lovable app or an n8n workflow. Something with its own URL. Record a two-minute demo. Now you have something to link on a CV.

The five prompts every graduate should own

The tailored cover letter opener. Feed CV + job spec, ask for three distinct opening paragraphs, pick one, edit heavily.
The interview pressure-test. Paste the job description. Ask AI to be a hostile interviewer and grill you on the toughest 10 questions.
The application-answer rewriter. Paste your first draft + word limit. Ask for a version that leads with impact and cuts filler.
The company research brief. Ask AI to summarise a company's last annual report, three strategic priorities, and likely interview themes.
The essay/report structurer. Paste your notes; ask for an argument-first outline before you start writing.

Save each as a Custom GPT or Claude Project. You'll use them dozens of times per application cycle.

Frequently asked

I've never used AI before. Where do I start?: Open ChatGPT or Claude in a browser today. Both have free tiers, no card required. Ask it to help with one real task you'd normally do yourself — a personal statement, a job application answer, a summary of your lecture notes. That single 20-minute session teaches you more than any intro video.
Is it too late to learn AI in 2026?: It's the opposite. Two years in, the tools are stable and the interfaces are graduate-friendly. Employers are only just starting to filter for AI-literate hires, so the gap between 'people who claim AI skills' and 'people who can prove them' is at its widest. That gap is your opportunity.
Is AI safe to use for my job applications?: Yes — with judgement. Use AI to draft, restructure and pressure-test your answers, not to submit unedited output. Every application system now has AI-detection built in; anything that reads like uncut AI output gets flagged. The winning move is human voice + AI leverage, not one or the other.
How much does it cost to get started?: Nothing. ChatGPT, Claude, Gemini, NotebookLM and Lovable all have free tiers that are enough to build a real portfolio project. You only need to pay (~£15–£20/mo) if you want longer context windows, image generation or faster models. Start free; upgrade only when you hit a specific limit.
What's the first useful thing to build?: A custom AI helper for your own job hunt. Feed it your CV, the target job description, and the company's about page. Ask it to draft tailored cover letter openings, predict likely interview questions, and mock-interview you. It's useful, it's low-stakes, and it becomes the demo you show recruiters.
I feel behind everyone else. Am I?: No. Surveys consistently show that under 30% of UK graduates use AI systematically for job applications, and under 10% have built anything with it. The 'everyone else' feeling is a Twitter/LinkedIn distortion — the highly online users are 1% of the applicant pool. You're closer to the front than you think.

Next step

You are not behind. You're on time.

The CareerGlowUp Build is the four-week ramp above, condensed into two weeks with support and accountability. You leave with a working AI tool and the story that gets interviews.

Join the waiting list