AImSearch AIm
YOU ARE HEREAIm→03 Apply→Career Compass
Back to ApplyDeveloper future

AIM CAREER COMPASS

Choose a skill.
Build proof you can use it.

Your practical career plan: explore a pathway, try a small project and keep evidence of your work. Jobs research and occupation comparisons live in Future of Work.

Future of WorkExplore AI in law, design, programming and other roles—plus the separate robot dataset and recent work news.
START WITH YOUR QUESTION

What do you need to understand?

Choose one pathway and one small project. You do not need to predict the job market before practising a useful skill.

Find my career lens →

WEEKLY AI CAREER RADAR

Ten roles worth investigating now.

Ranked for opportunity signal, accessibility and relevance—not salary promises. Every path includes something you can learn and something you can prove.

01People + process94

AI adoption specialist

Helps ordinary teams choose useful tools and change how work gets done.

Study path
AI fluency, process mapping, change management and responsible adoption.
Experience proof
Document one before-and-after workflow with time saved, risks and human checks.
Open the full career pathway +
Foundation you need
Useful experience in a real workplace, clear communication and curiosity about how work actually happens.
What to learn
AI capabilities and limits, process mapping, change management, privacy, responsible use and measuring adoption.
Formal study
A degree is helpful but not essential. Business analysis, change management or industry qualifications can strengthen credibility.
Tools to practise
A mainstream AI assistant, process-mapping software, spreadsheets, survey tools and a simple project board.
Possible entry route
Begin as the AI champion in a department you already understand, then lead one supervised adoption project.
Your first practical plan
Map one workflow, establish a safe-use rule, train five users and publish a before-and-after result with limitations.
02Governance92

AI governance and risk analyst

Organisations need policies, inventories, impact checks and evidence.

Study path
AI governance frameworks, privacy, risk, audit and sector regulation.
Experience proof
Create a sample AI register, risk assessment and practical use policy.
Open the full career pathway +
Foundation you need
Comfort with policies, evidence, structured analysis and asking difficult questions about impact and accountability.
What to learn
The NIST AI RMF, privacy principles, risk registers, impact assessments, model documentation, audit and relevant local regulation.
Formal study
Risk, law, compliance, audit, information security or data-governance study is valuable. Coding is usually not the first requirement.
Tools to practise
Risk registers, policy templates, data-flow diagrams, control libraries and documentation or audit platforms.
Possible entry route
Move from compliance, audit, privacy, security, legal operations or responsible technology work.
Your first practical plan
Create a sample AI inventory, classify three use cases, complete one impact assessment and draft a practical use policy.
03People + process91

AI product manager

Connects customer problems, business value, design and technical teams.

Study path
Product discovery, metrics, AI limitations, UX and responsible release.
Experience proof
Write a product brief and test a small prototype with five real users.
Open the full career pathway +
Foundation you need
Customer empathy, prioritisation, commercial judgement and the ability to coordinate design, engineering and business teams.
What to learn
Product discovery, user research, AI failure modes, evaluation metrics, experimentation, data responsibility and release controls.
Formal study
Product-management training helps, but a portfolio of evidence can matter more than a particular degree.
Tools to practise
Prototyping software, analytics, experiment tracking, issue management and an AI assistant for safe concept testing.
Possible entry route
Move from product, operations, customer success, business analysis or domain-specialist roles.
Your first practical plan
Interview five users, define one problem and success metric, test a low-risk prototype and document why it should or should not proceed.
7 more career pathways are available with free membership.

Create a free AIm account to save your direction and unlock more checked guidance.

Create a free account →

Career radar reviewed 4 October 2026. Regional demand varies; verify live vacancies, entry requirements and right-to-work rules before investing in training.

PREPARE PRACTICALLY

Turn uncertainty into one useful next step

Open practical guides for skills, evidence, role redesign and safer career decisions.

How to use these guides: choose one question, try one step and keep one piece of evidence. Open the deeper explanation only when it helps.

01

AI for your current job

List five repetitive tasks. Test one low-risk task with sample data. Keep the human decision and measure whether quality improves.

Anyone who wants a safer starting point without changing career.One week · 30 minutes a day
Open the practical guide +

Why this matters

The quickest value often comes from improving a task you already understand. Your experience helps you spot weak AI answers.

What to do

  1. List five tasks that repeat every week and circle one that uses no confidential data.
  2. Write down the current time, quality problems and the decision that must stay human.
  3. Run the task three times with sample information and compare it with your normal method.
  4. Keep the experiment only if it saves time without lowering accuracy or accountability.

Evidence to keep

A before-and-after example, time saved, errors found and the final human check.

Watch out for

Do not paste customer, employee, financial, health or unpublished business information into an unapproved tool.

Starter prompt

Help me redesign this task safely. Ask what the goal is, what information is sensitive, what a good result looks like and which decision must remain human.

02

Your first ten useful prompts

State the goal, audience, context, constraints and desired format. Ask the tool to identify uncertainty and show what needs checking.

Beginners who get vague, generic or overconfident answers.45 minutes to build · reuse every week
Open the practical guide +

Why this matters

A reusable prompt library removes the blank-page feeling and makes results easier to compare and improve.

What to do

  1. Choose ten real tasks: explain, summarise, brainstorm, compare, rewrite, plan, question, check, practise and reflect.
  2. For each prompt include the goal, audience, useful context, limits and desired format.
  3. Add: identify uncertainty, do not invent facts and tell me what I should verify.
  4. Save only prompts that worked on more than one example and record what you changed.

Evidence to keep

Ten named prompt templates with an example input, useful output and a verification note.

Watch out for

A detailed prompt can still produce a confident mistake. Important claims need reliable sources outside the AI answer.

Starter prompt

Act as a patient coach. Ask me one question at a time until you understand my goal, audience, context, constraints and preferred output. Then draft a reusable prompt and a verification checklist.

03

Build a credible portfolio

Show the problem, your process, a safe experiment, the measurable result, what failed and where human review remains necessary.

Career changers, consultants and employees proving practical value.Two weekends for the first case study
Open the practical guide +

Why this matters

A portfolio demonstrates judgement and results. Tool certificates alone rarely show how you solve a real problem responsibly.

What to do

  1. Pick a small problem you have permission to work on and remove private information.
  2. Describe the old process, users, success measure and risks before mentioning a tool.
  3. Show your prompt or workflow, iterations, failed attempts and verification method.
  4. Finish with the measurable outcome, limitation and what a qualified human still decides.

Evidence to keep

A two-page case study with redacted examples, measures, limitations and permission to share.

Watch out for

Never use employer or client material without written permission. Do not invent results, testimonials or customers.

Starter prompt

Interview me about this project and turn my answers into an honest case study: problem, users, baseline, experiment, evidence, failures, safety controls, result and limitation.

04

Move into AI without coding

Start with adoption, governance, training, product, workflow analysis, evaluation or industry-specialist roles.

People with domain, communication, operations, teaching or customer skills.Eight to twelve weeks for credible evidence
Open the practical guide +

Why this matters

Many AI roles need problem framing, trust, evaluation, change and industry knowledge more than software development.

What to do

  1. Choose one route: adoption, governance, training, workflow design, quality evaluation or industry expertise.
  2. Read ten current job descriptions and highlight repeated outcomes and skills—not tool names alone.
  3. Learn one foundation, complete one credible free course and build one role-relevant example.
  4. Speak with three practitioners and revise your plan using what employers actually ask for.

Evidence to keep

A skills-gap map, one relevant course, one portfolio example and three practitioner conversations.

Watch out for

Avoid expensive bootcamps promising a job or guaranteed income. Confirm employer demand before paying for training.

Starter prompt

Compare my existing experience with these AI-adjacent career routes. Show transferable strengths, important gaps, one free learning source and a small portfolio project for each suitable route.

05

Prepare for an exposed role

Identify which tasks are routine, strengthen relationship and judgement work, learn AI-assisted workflows and collect evidence of value.

People worried that routine parts of their role may be automated.A 30-day resilience plan
Open the practical guide +

Why this matters

Roles usually change task by task. Separating exposed tasks from human strengths creates a more useful plan than predicting whether a whole job disappears.

What to do

  1. Break your role into tasks and mark each as routine, judgement-heavy, relationship-based or physical/contextual.
  2. Learn to supervise AI on one exposed task while improving a distinctly human part of the role.
  3. Ask your manager which outcomes and responsibilities will matter more over the next year.
  4. Record quality, customer trust and decisions you improved—not merely faster output.

Evidence to keep

A task map, one supervised AI workflow, one strengthened human capability and a manager conversation.

Watch out for

Exposure scores are signals, not predictions. Do not make a major career decision from one ranking or headline.

Starter prompt

Help me map my role by tasks. For each task estimate why AI may assist, change or struggle with it, what evidence is missing and which human capability I should strengthen.

06

Check an AI job advert

Confirm the employer, role scope, location and pay. Never pay to apply, share credentials or install unknown interview software.

Anyone evaluating AI jobs, freelance work or training-linked offers.15-minute check before applying
Open the practical guide +

Why this matters

Fast-growing fields attract impersonation, fake interviews and unrealistic earnings claims. A short check protects your identity, money and devices.

What to do

  1. Find the role independently on the employer's official careers domain.
  2. Check the recruiter identity, email domain, location, duties, salary language and publication date.
  3. Compare the role with the employer's real products and two similar adverts.
  4. Stop if asked for payment, cryptocurrency, passwords, identity documents too early or unknown software.

Evidence to keep

The official listing URL, verified recruiter channel and a saved copy of the role description.

Watch out for

An AI-generated video interview or polished company page is not proof of identity. Contact the employer through an independently verified channel.

Starter prompt

Create a verification checklist for this job advert. Separate facts I can confirm, claims needing evidence, suspicious requests and safe questions to ask the employer. Do not decide that it is legitimate for me.

CHECK THE EVIDENCE

How AIm judges a career signal

Read the AIm source notes first, then open original research only when you need the full methodology.

HOW AIm JUDGES THE SIGNAL

Evidence first. Automation for facts—not forecasts.

The weekly Career Scout compares employer demand, credible labour research, skills requirements, location, recency and scam signals. Verified low-risk source and date corrections can publish automatically. Rankings, disputed evidence and consequential career claims remain human-reviewed. AIm never promises outcomes.

ILO · Global GenAI exposure index+
AIm source note

Compares occupations by the tasks generative AI could affect. Exposure means potential task change—not a prediction that the job disappears.

Optional: open the full source ↗
OECD · AI exposure measure+
AIm source note

Adds evidence about where AI capabilities overlap with occupational tasks. It is useful for comparison, but local demand and employer choices still matter.

Optional: open the full source ↗
World Economic Forum · Future of Jobs+
AIm source note

Summarises employer expectations about growing and declining roles and skills. Treat these as surveyed expectations, not guaranteed outcomes.

Optional: open the full source ↗
US BLS · Occupational projections+
AIm source note

Provides occupation-level employment projections based on US labour data. It is a useful demand signal, but it does not represent every country or measure AI alone.

Optional: open the full source ↗