The headlines
- The fear question, "will AI take my job", is the wrong one. The useful question is "what work is AI making possible".
- On current projections AI is a net job creator: about 78 million more jobs worldwide by 2030, not fewer.
- Whole new roles have appeared in five years, clustered where AI meets data, security and human judgement, many paying well above average.
- The barrier is not robots, it is skills: 39% of workers' core skills will change by 2030.
- You cannot hire your way out of that gap. The winners build AI capability in-house, and in the UK that is levy-funded.
Look at the job titles filling UK hiring platforms in 2026: AI governance manager, responsible AI lead, AI ethicist, automation practitioner. Five years ago, none of these existed in any formal sense. A few years from now they may be as ordinary as "software developer" or "marketing analyst". We are living through one of the largest reshuffles of work in modern history, and most people are still looking at it through the wrong lens.
The headlines fixate on what AI takes away. That misses the more interesting, and more accurate, half of the story: the list of jobs AI is creating is growing fast, and the list of roles it is quietly making more valuable is longer still. This guide walks through both, with the numbers behind them, the salaries attached, and then gets to the part that actually matters for a UK business: what to do about it.
In a hurry? Book a free Role Mapping Call and we'll show you exactly which roles to build internally, which to augment, and which to hire for.
First, the question worth asking
"Will AI take my job" assumes work is a fixed pile that machines slice away. It never has been. Every general-purpose technology, electricity, the computer, the internet, destroyed categories of work and created larger ones, and the people who thrived were the ones who learned to use the new tool rather than compete with it. AI is following the same pattern, only faster.
So the better questions are these. What new roles is this technology creating? Which existing roles does it make more productive rather than redundant? And what is the human contribution that becomes more valuable, not less, as the machines get better? Answer those three and the future of work stops being a threat and starts being a plan.
Source: World Economic Forum, Future of Jobs Report 2025, drawing on more than 1,000 employers across 55 economies.
The three zones of work in the AI era
Most "best jobs of the future" lists are just a pile of job titles. That is not very useful, because it tells you nothing about why a role is on the list or what to do with it. A simpler map is to sort every job into three zones by how AI touches it.
The mistake is to obsess over the first zone and ignore the other two. The opportunity, for individuals and employers alike, lives in the second and third.
Book a free 20-minute Role Mapping Call. We'll show you exactly which roles to build internally, which to augment, and which to hire for.
Book a Role Mapping Call →Zone 3: brand new jobs that did not exist 5 years ago
Start with the most exciting zone, because it is the one the doom headlines never mention. These are real, advertised, well-paid UK roles that simply were not on the org chart in 2021. Almost all of them exist for the same reason: AI is powerful, but it needs people to govern it, secure it, point it at the right problem and check its work. Here are seven, with an indicative UK salary and the TESS route that builds the skills for it.
AI governance / responsible AI lead
Sits between the data team and legal and risk, translates AI risk into plain language, and makes sure systems match the rules and values the organisation signs up to.
AI & automation practitioner
The person inside a business who actually builds the workflows: automating reporting, connecting tools, turning "we have AI" into hours saved every week.
Business information security officer
Embeds security thinking into everyday business decisions, talking to engineers in the morning and the board in the afternoon, as AI makes attacks faster and harder to spot.
AI ethicist
Asks the "should we", not just the "can we": what happens to users and society if we deploy this here, this way, as AI spreads into hiring, lending and healthcare.
Data curation & evaluation specialist
Prepares, tags and verifies the data AI learns from, and grades model output for quality, accuracy and bias. AI does not learn on its own.
AI product manager
Decides what AI features to build, for whom, and how to ship them responsibly and measurably. Turning a clever model into a product people trust is a distinct discipline.
AI agent / automation manager
Manages the working relationship between staff and AI agents: designing workflows, training people, and stepping in when the human side needs support.
Salary figures are indicative UK ranges for 2026 and vary widely by sector, seniority and location. Sources include published UK salary guides; treat them as a guide, not a quote.
How Microsoft Copilot fits into these new AI roles
Almost every role above starts in the same place: someone getting genuinely good at an everyday AI tool, and for most UK businesses that tool is Microsoft Copilot. Copilot is where people first learn the skills these jobs are built on, writing a clear prompt, automating their own repetitive work, and checking AI output for accuracy and bias before they trust it.
That is the on-ramp. The person who learns to use Copilot well in their current job is the same person who grows into an automation practitioner, an AI agent manager, or the governance lead who can tell good output from bad. We start there from day one, then build towards the full apprenticeship. For the practical version, our Copilot guide for business teams walks through the exact prompting frameworks and guardrails we teach.
Notice what these roles have in common. None of them are "build the model". They are about directing, governing and applying AI inside a real organisation. That is good news, because those are skills you can build in people you already employ, not exotic talent you have to win a bidding war for.
Zone 2: roles AI makes bigger (not redundant)
This is where most people actually work, and where most of the change will land. These jobs are not disappearing. The repetitive part is being handed to AI, and the human is moving up the value chain to judgement, exceptions and relationships.
| Role | What AI takes over | What the human moves towards |
|---|---|---|
| Data analyst | Pulling, cleaning and charting data | Framing the right question and interpreting what the numbers mean for the decision |
| Marketer | First drafts, variants, scheduling, reporting | Strategy, brand judgement, knowing what is actually worth saying |
| Software developer | Boilerplate code, test scaffolding, documentation | System design, reviewing AI output, deciding what to build |
| Customer service | Routine, repetitive queries | Complex cases, upset customers, the moments that need a human |
| HR & recruitment | CV sifting, drafting adverts and policies, scheduling | Judgement on people, culture, motivation and difficult conversations |
| Finance & bookkeeping | Data entry, reconciliation, routine reporting | Analysis, advice, spotting what the numbers are really saying |
The through-line is the same in every row: AI handles the predictable, the person handles everything else. The people who win in these roles are not the ones who resist the tool, nor the ones who blindly trust it, but the ones who learn to direct it and check it well.
Zone 1: the human skills that become more valuable
It is worth saying plainly, because the fear is real. There is a core of work that AI does not replace, and counter-intuitively, the more capable AI becomes, the more valuable that core gets. Leadership and the ability to take responsibility for a decision. Complex judgement where the factors do not fit neatly into a formula. Care work: teaching, nursing, therapy, the jobs that are fundamentally about one human being present for another. Original creative work. Negotiation, persuasion and trust.
The World Economic Forum calls the rising category "judgement work", and it captures the point well. As routine output becomes abundant and cheap, the scarce, prized skill is knowing what is worth doing, whether the output is any good, and what it means for real people. That is a human contribution, and it is growing.
The more powerful AI becomes, the more valuable distinctly human judgement turns out to be. The new jobs are not replacements for human effort. They are expressions of it.— Rod Doyle, Director, TESS Group
Which jobs are most at risk from AI by 2030?
The roles most exposed to automation are the ones built on predictable, repeatable tasks: routine data entry, straightforward bookkeeping, first-line customer queries, chunks of clerical and administrative work, and simple reporting. Forecasters from PwC to Goldman Sachs put large numbers on this exposure, and it is real.
But exposed does not mean extinct. In almost every case it is specific tasks that get automated, not the whole job, and the role is rebuilt around the parts that need a person: handling exceptions, exercising judgement, and looking after relationships. The risk concentrates in people who do not re-skill, not in the job titles themselves. Which is exactly why the response that works is training, not fear.
The real problem: the AI skills gap (2026)
Here is the uncomfortable middle of the story. AI being a net job creator does not mean it is painless, because the new and reshaped jobs need different skills from the ones they replace. The same research that projects 78 million net new jobs also finds that 39% of workers' core skills will change by 2030, and that the skills gap, not the technology, is the single biggest barrier employers report.
In other words, the jobs are coming. Whether your people, or your business, are ready to do them is the open question. This is the hinge on which the whole future-of-work debate turns, and it is the part most "jobs of the future" articles quietly skip, because it is the hard bit.
Estimates of how many jobs AI exposes vary widely (Goldman Sachs has put the global figure at the equivalent of 300 million full-time roles, while also forecasting a meaningful lift to global GDP). The number everyone agrees on matters more: most of today's workforce will need to learn new skills this decade. That is a training problem, and training problems are solvable.
If you are planning your own career, the takeaway is simple: treat any qualification you have as a foundation, not a finished house. A degree still teaches you how to think, but on its own it no longer carries a whole career. Build two layers on top: the applied AI layer (prompting, automating your own work, checking output) and the durable human layer (communication, judgement, leadership, the willingness to keep learning). You do not need to become a machine-learning engineer. For most people, the highest-return move is to become genuinely good at using AI in the job they already do, which is what a structured AI apprenticeship is built to deliver.
What this means for employers in the UK
For a business, the temptation is to treat all this as a hiring problem: go and recruit the AI governance lead, the automation practitioner, the data specialist. Good luck. Those people are scarce, expensive (the projected average UK AI salary is now around £81,000), and every competitor wants them too. And even if you win the hire, you cannot recruit your way out of a 39% shift in the skills of your whole workforce. The maths does not work.
The UK context makes the build option unusually attractive right now. The Growth and Skills Levy already pays for accredited AI apprenticeships, and from August 2026 the rules get more generous: free training for under-25s at smaller employers and a £2,000 hiring grant. We see the sharpest demand in the sectors we work in most, finance, the public sector and manufacturing, where AI governance, automation and data skills are moving from "nice to have" to "cannot operate without".
Team Leader Level 3 loses funding in September 2026. If you have run leadership cohorts through it, that budget does not have to disappear, it can move to AI-enhanced leadership routes that build the new skills your managers actually need. See which standards are defunded and how to transition a Team Leader L3 cohort to AI L4.
How to build these skills internally (without hiring)
The employers who come out ahead do the opposite of a hiring scramble. They build capability in the people they already have, at scale, and hire selectively on top. It is cheaper, it is faster across a team, it lifts retention because people can see you investing in them, and it spreads AI fluency through the organisation instead of bottling it up in two new starters.
| The old playbook | The new playbook |
|---|---|
| Try to hire scarce AI specialists | Build AI skills across the team you already have |
| One or two experts, AI fluency bottled up | Whole departments confident with AI day to day |
| Expensive, slow, easily poached | Funded, scalable, and it improves retention |
| Capability walks out when someone leaves | Capability stays in the business and compounds |
In practice it is a simple four-step loop.
Map your roles. Work out which are exposed, which are being augmented, and which are brand-new AI positions you will need.
Match each to a funded route. Map every role to an apprenticeship or unit that builds the right skills (the table of roles above is your starting key).
Train your existing team. Put current employees through the programme on levy funding, so capability is built across the team, not bought in one or two hires.
Embed guardrails and measure. Build responsible-AI habits into how the team works, and measure the productivity gain so the investment is visible and safe.
Tell us the roles you need to fill and we'll map them to the right levy-funded route, so the training cost is usually covered.
Map my roles →How TESS Group can help you map & fill these roles
We are an AI apprenticeship specialist, so this is the part we are built for. The roles in Zone 3 and the augmented jobs in Zone 2 are not abstractions to us; they are what our programmes produce. Across Levels 3 to 6, every one is delivered on levy funding, with no coding background required to start. Here is the direct mapping.
| The role you need | The funded TESS route |
|---|---|
| AI & automation practitioner, AI agent manager, data curation specialist | AI & Automation Practitioner Level 4 (flagship) |
| AI governance lead, AI ethicist, AI product manager, the AI-risk side of security | AI for Leaders & Managers Level 4 + AI Adoption & Governance unit |
| HR and people teams adopting AI responsibly | AI for People Leaders Level 4 |
| A fast start for a whole team, before a full apprenticeship | Short AI Apprenticeship Units (2 to 4 weeks each) |
What this looks like in practice
Your ops team is drowning in manual reporting
Rather than hire an automation specialist, put one or two existing people through the AI & Automation Practitioner Level 4. They learn to build the automations on your actual processes, and the capability stays in the team. Fully levy-funded.
Your managers do not know how to lead AI adoption
Adoption stalls when leaders cannot model it. AI for Leaders & Managers (Level 4) gives managers the governance, judgement and adoption playbook to lead it properly, not just sign off the licence.
Your HR or people team is rolling out Copilot
A licence is not a capability. AI for People Leaders (Level 4) turns an HR team into confident, responsible AI users, the gap we see again and again, including in the Copilot workshop we ran with a client HR team.
Tell us the roles you are most worried about, or most excited about. We will map them to the right funded route and send you a short recommendation report, no commitment. We handle the funding detail; the AI & Automation Practitioner Level 4 and our leadership routes are fully levy-funded.
Map my roles & send the report →Frequently asked questions.
What new jobs is AI creating in 2026?
Roles that barely existed five years ago are now among the fastest-growing in business: AI governance and responsible-AI leads, business information security officers, AI ethicists, AI and automation practitioners who build workflows inside a business, data curation and evaluation specialists, AI product managers, and people who manage the working relationship between staff and AI agents. They cluster where AI meets data, security and human judgement.
Which jobs are most at risk from AI?
Roles built mainly on repetitive, predictable tasks are the most exposed: routine data entry, basic bookkeeping, first-line customer queries, parts of administrative and clerical work, and simple reporting. Most of these jobs are reshaped rather than deleted outright, as the predictable parts are automated and the human keeps the judgement, exceptions and relationships.
Will AI create more jobs than it destroys?
On the largest current estimate, yes. The World Economic Forum's Future of Jobs Report 2025 projects 170 million new roles created and 92 million displaced by 2030, a net gain of about 78 million jobs. The catch is the skills gap: the report also finds 39% of workers' core skills will change by 2030, so the new jobs go to people and teams who have re-skilled.
Do you still need a degree to work in AI?
A degree still helps, because analytical thinking and depth of knowledge still matter, but it is no longer enough on its own and it is not the only route in. Many of the new AI roles reward applied skill: the ability to use AI tools well, judge their output, and redesign how work gets done. In the UK, apprenticeships such as the AI and Automation Practitioner Level 4 build exactly those applied skills, with no degree required to start.
What skills should my team build for the future of work?
Two layers. First, applied AI skills: prompting, working alongside AI without blindly trusting it, automating repetitive workflows, and judging output for accuracy and bias. Many of these new roles start with practical Microsoft Copilot skills. Second, the durable human skills AI does not replace: analytical judgement, leadership, communication, and adaptability. The highest-value people pair both, which is what a structured AI apprenticeship is designed to develop.
How can a UK employer fund AI skills training?
Through the Growth and Skills Levy (the successor to the Apprenticeship Levy). Levy-paying employers draw down from their levy pot, and smaller non-levy employers have most or all of the cost met by government, with full funding for many under-25 starts from August 2026. TESS Group's AI apprenticeships, from Level 3 to Level 6, are delivered on this funding, so the training cost is usually covered.