Will AI Change Graduate Jobs? The Skills Employers Say Still Matter

AI is already changing some of the work described by employers in these recordings. For a student, the useful question is what you should learn and demonstrate when applying for an internship, trainee programme or first job. A confident prediction about your entire career is less useful than knowing how teams expect you to work.

We reviewed seven CareerFairy Live Streams, including UBS’s AI discussion, its Group Functions conversation, EY-Parthenon’s technology strategy session, EY application and trainee sessions, and CareerFairy’s panel on the first three working years. One additional AI recording listed in the backlog had no available transcript and is excluded from this guide.

The quick answer: the speakers pointed towards a combination of relevant subject knowledge, practical AI understanding, critical judgement, communication and continuous learning. They described AI changing tasks and workflows. These conversations do not establish how many graduate jobs will disappear, or guarantee that any role will remain unchanged.

How do employers say AI is changing early-career work?

They described tools becoming part of existing work, alongside specialist teams building and checking AI systems. That creates several different ways a graduate might encounter AI.

In its September 2025 session, UBS said its ambition was to operate as an AI-enabled institution and that AI concerned employees across the bank. The speakers discussed general tools as well as applications linked to the bank’s own knowledge and advisory work.

One UBS speaker expected many tasks to involve collaboration between people and AI. His emphasis was on having domain expertise as well as knowing how to use AI, recognise risks and respond to them. That is a view expressed in the recording, rather than a proven description of every future job.

UBS’s April 2026 Group Functions session made the discussion more concrete. The teams described data and technology used in compliance, reviewing clients and detecting suspicious activity. They also discussed model risk and testing whether models were reliable in the decisions where they were used.

In EY-Parthenon’s November 2025 technology strategy conversation, speakers described internal AI tools helping with documents and extracting insights. They said tools could speed up parts of the work, while stressing that outputs needed checking and that consulting involved understanding people and clients.

The practical question for your next interview is therefore specific: which tasks in this team involve AI, and what responsibility stays with the person doing the work? That tells you more than asking whether the whole industry is safe from automation.

Will AI replace graduate jobs?

The transcripts cannot answer that for the job market as a whole. They contain employer plans, current examples and individual opinions, not a forecast of graduate employment.

UBS speakers expected substantial changes to roles and described the speed of technological adoption as a reason to keep adapting. EY-Parthenon speakers were more reassuring about the continued need for consultants and interns, particularly because of judgement and relationships with clients.

Those positions are compatible in some respects: work can change while people remain involved. But it would be misleading to turn them into a universal promise that every graduate job will remain available.

It would also be misleading to say that interpersonal work cannot change. The EY-Parthenon point was grounded in the speakers’ own consulting experience: understanding client concerns and working through human situations formed an important part of their roles.

For your preparation, separate the claims you can act on from the predictions you cannot verify. You can learn how to evaluate an output, understand a business problem and explain your reasoning. You cannot choose a permanently future-proof job from these conversations.

Which skills matter if everyone can use an AI tool?

The employer evidence points to knowing what a good result looks like. AI familiarity becomes more valuable when you can connect it to the subject and task.

Relevant subject knowledge

UBS’s AI speakers linked effective use to domain expertise. Its Group Functions teams also wanted a foundation in technology or the relevant business area that they could build on. A tool does not remove the need to understand the problem you are solving.

EY’s March 2026 trainee discussion made a related point in another setting. The speakers described extensive learning after joining, but said a certain amount of relevant basic knowledge was still important. Structured training was not presented as a substitute for every foundation from study.

For an application, name the foundation you actually have. It could be accounting knowledge, software development, statistics or an understanding of a business process. Then explain how you have applied it and where you are still learning.

Critical judgement

EY-Parthenon explicitly discussed checking whether AI-generated output was reliable and made sense. UBS placed responsible use, governance and risk within its AI discussion.

Our practical synthesis is to prepare an example where you evaluated a result rather than accepted it. What information did you compare it with? What uncertainty remained? What did you correct? Use an actual academic or personal project and avoid confidential material.

This is stronger evidence than saying that you use an AI assistant every day. The frequency of tool use says little about the quality of your decisions.

Communication and understanding people

EY-Parthenon’s speakers described consulting as a business involving people: understanding what matters to clients and where their difficulties lie. UBS’s technical careers session also encouraged dialogue during interviews, including asking questions about the problem.

You can practise this through group projects. Explain a technical choice to someone with a different background, check what they need and respond to their questions. In an interview, describe the situation and your contribution rather than simply claiming excellent communication skills.

Learning through change

The CareerFairy first-three-years panel stressed that much learning happens after university. Speakers also pointed out that useful skills can come from work, moving to a new environment and other experiences outside the degree itself.

That does not mean every experience is equally relevant. It means you should explain what it taught you. Adapting to an unfamiliar situation, responding to feedback or learning a new tool can be meaningful evidence when it connects to the role.

Do you need to become an AI engineer?

No single technical path was presented as necessary for everyone. UBS’s September 2025 discussion distinguished different levels of AI learning, including non-technical users, builders and people with more specialist expertise.

The speakers described graduate learning pathways covering foundations, risks, data, deeper topics and practical exercises. That was the approach discussed at UBS at the time; it is not a promise about onboarding at another employer.

If you want a technical AI role, the relevant technical requirements still matter. UBS’s Group Functions discussion described teams at the intersection of software, data and business problems, and acknowledged that matching the whole technical stack and business knowledge was difficult for a new graduate.

If you want a business role, your starting point may instead be understanding appropriate use, evaluating information and knowing when to ask someone with specialist expertise. Learn what the vacancy requires before collecting tools or courses indiscriminately.

The distinction helps you set a realistic aim. You do not need to pretend that trying a chatbot makes you an AI specialist. You can show that you understand how a tool fits your work and what you need to learn next.

How should you show these skills on your CV?

Use evidence tied to the job description. EY Switzerland’s May 2025 interview session recommended comparing your academic and professional experience with the role’s requirements, including both technical and interpersonal skills.

The recruiters explained that a CV can emphasise different responsibilities for different roles while remaining truthful. The same project might demonstrate analysis in one application and coordination in another, depending on what you actually did.

EY Germany’s February 2026 application training encouraged students to reflect on their skills and experiences broadly. It also discussed preparing examples of success, difficulties, team roles and feedback, including examples from university when professional experience was limited.

Our synthesis is to write a project bullet around the task, your action and the result. If AI was involved, explain your contribution: choosing an approach, checking information, comparing results or communicating a decision. Avoid making the tool sound like the only contributor.

You do not need a made-up percentage improvement. A clear description of what you completed or learned is better than an impressive number you cannot defend.

For more help choosing evidence, read what recruiters look for in graduates. If you are applying for technical roles, our guide to junior software engineering expectations develops the project and interview angle.

What should you ask an employer about AI and training?

Ask about how the team works, rather than expecting a recruiter to predict the future labour market. The UBS and EY recordings support questions about tools, responsibility, relevant foundations and learning.

For example, ask which tasks a new intern would do personally and which tools support them. Ask how work is reviewed, what training is available and how someone learns the business context. For a technical team, ask how junior colleagues contribute and when they work with experienced specialists.

EY-Parthenon described training and an academy alongside gradual introduction to projects. EY’s trainee session explained that trainees received the same functional learning as direct entrants, while the programme offered exposure to more areas. That distinction is useful: rotations and training are related benefits, but they answer different needs.

A programme with lots of rotations is not automatically the one with the best AI learning. Check both the work and the support. Our guide to trainee programmes versus direct entry can help you compare those routes.

What can you do now to prepare for changing graduate work?

Build a small, relevant example rather than trying to master every new tool. This sequence is our synthesis of the employer discussions.

Choose a role you might apply for and identify a realistic problem within it.
Use your existing subject knowledge to plan an approach before reaching for a tool.
If you use AI, work with suitable non-confidential material and check the result against evidence you understand.
Record what you changed, what you could not verify and what you would ask an experienced colleague.
Explain the work to another person and use their questions to improve your explanation.

The exercise gives you something concrete to discuss. It also exposes gaps you can work on: a weak foundation, an unchecked assumption or an explanation that makes sense only to you.

Frequently asked questions about AI and graduate jobs

Will AI eliminate all entry-level jobs?

These recordings do not support that conclusion. UBS discussed changing roles and human–AI collaboration; EY-Parthenon speakers expected continued demand for people in their work. Neither establishes a forecast for all employers.

Are communication skills enough on their own?

No. The discussions also stressed relevant foundations and judgement. The useful combination is understanding the task, evaluating information and communicating with the people involved.

Do I need coding skills for every AI-related job?

No universal requirement emerged. UBS distinguished non-technical AI users from builders and specialists. The technical requirements depend on the role, so check the current vacancy.

Can university projects demonstrate useful skills?

Yes. EY’s application sessions explicitly included academic experience, teamwork and learning from feedback. Explain your own contribution and its relevance to the position.

Is using an AI tool enough to stand out?

Tool use alone says little about your judgement. The UBS and EY-Parthenon discussions suggest showing how you understood the problem, checked the result and took responsibility for the work.

Want to ask how a team is changing? Join a CareerFairy Live Stream and ask employers about the tasks, tools and learning they expect from their next interns and graduates.