Hiring ‘AI-native’ graduates? Rethink the job you’re giving them

Before adding “AI skills” to the next graduate job advertisement, be clear about what you expect that person to bring to the table, writes Bridie Rees.

“Help us automate this.” For a graduate who’s used AI throughout university to research assignments, test ideas and work through unfamiliar concepts, it can sound like a natural first project. Employers may see an “AI native” who’s ready to start building. Yet the hardest part of that brief may be understanding the work itself: why a request passes through three people, which checks serve a purpose and what happens when the process goes off script. That brief asks graduates to help redesign the ropes while they’re still learning them.

As tools make it easier for employees to build agents and automate workflows, graduates have an opportunity to contribute to how their teams operate at the earliest stages of their careers. They might help design an agent that coordinates onboarding, routes customer requests, or follows up missing information. To make it useful, they’ll need to talk to the people doing the work, trace the handovers and understand the exceptions. Those conversations can become part of how they learn the business, while giving the team a chance to question processes it has long taken for granted.

Our latest research of enterprise workers in Australia found that 71 per cent of workers in AI-active teams say AI meaningfully reduces manual work. For employers, that raises a practical consideration about the graduate role. Could juniors help identify and build those improvements as part of learning the business? Alongside knowledge of their field, hiring managers should assess whether candidates can investigate an unfamiliar problem, question assumptions and understand the people affected by a proposed automation. Relationship building will be a key skill in setting up effective agentic processes.

Find out how the work actually happens

I joined monday.com as the second member of our APJ People team. My colleague and I were doubling up on work and constantly requesting updates: Are you working on this? Where’s that up to? What’s happening? With just two of us, we still needed a shared way to organise our work and make ownership clear, so we set up a board where we could track our work. That experience shapes how I’d approach automation to this day. We had to recognise what was causing the back-and-forth before we could improve it. An agent built around unclear ownership would still need someone to allocate a job correctly.

For a graduate designing a workflow, the written procedure is a starting point. Think of it as a map: the people doing the work can explain the workings, shortcuts, and dead ends that never made it onto the page. Consider onboarding. A checklist might show that HR sends the paperwork, IT organises equipment, and the manager arranges induction. In practice, a start date changes, the equipment request contains the wrong information, or everyone assumes someone else has booked the first-day meetings. An unwritten step in the procedure may be work someone has been quietly doing for years. Before automating it, find out what that person knows.

Those details determine what to build. A fixed automation might send a reminder when a deadline passes; an AI agent could help interpret a request and route it to the right person. Both need clear ownership and reliable information. The team still needs to agree on who owns the decision before handing it to an agent. A graduate’s discovery can help reveal those gaps, while experienced colleagues explain which steps protect the business and which have simply become habit. This gives juniors a useful way to contribute as they build their understanding: they learn how the process works by helping the team examine it.

Hire for the questions the job requires

If a graduate will help build AI workflows, the interview should reveal how they approach an unfamiliar problem. Give them a fictional example of customer requests bouncing between two departments and let them question an interviewer playing a team member before recommending a solution. I’d pay attention to whether they ask who owns the request, what information each team needs and what happens in an unusual case. I’d also look at whether they revise their thinking when they hear something new. A confident demonstration of an AI tool can be impressive; the assessment needs to establish whether the candidate has identified a problem worth solving. Building, testing and improving an agent can help graduates refine their understanding, but asking the right questions at the outset saves time and gives that experimentation a clearer direction.

Foundational expertise remains essential. A finance graduate needs accounting knowledge to understand a reconciliation check. An HR graduate needs to recognise when an employee request involves circumstances that a standard workflow won’t accommodate. That knowledge helps them ask better questions and recognise the limits of their understanding. Employers should assess whether a candidate can explain their assumptions, identify what they still need to learn and seek advice from someone with the relevant experience.

Importantly, there’s a relationship skill needed. Asking someone to explain their work can feel like asking them to defend it, particularly if the conversation starts with a proposal to automate parts of their role. A graduate needs to listen carefully and understand the pressures that led to a workaround. People share more when they feel their experience is being respected and preparation has gone into a conversation. Curiosity becomes useful when someone can draw out those details, make sense of different accounts and explain what needs to change, in alignment with keeping the process running.

Make automation a shared responsibility

Employers can’t assume graduates will work all this out because they’ve used AI at university. Our research also found that only half (51 per cent) of workers believe their organisation helps them focus on key priorities. Before asking a graduate to automate a process, employers need to explain which problem matters and what improvement would look like. Start with one contained problem and pair the graduate with the person responsible for that process. Give them access to the colleagues who experience it, time to investigate and a clear outcome to work towards. The graduate can contribute to the design while an experienced owner remains accountable for how the process operates.

Platforms that let employees build agents around shared workflows, with instructions, access, and permissions are needed, to give teams a practical way to build agents that use shared information to carry out agreed tasks. They still need to decide what the agent can do, where there is a human in the loop and who will maintain it. Test the awkward cases as well as the straightforward ones: a changed start date, an incomplete request, or a conflicting instruction may tell you more than a demonstration where everything goes to plan. Judge the result by fewer misdirected requests, shorter waits or less time chasing information, and ask the graduate to explain what they changed to achieve that result.

Rewrite the graduate brief

Before adding “AI skills” to the next graduate job advertisement, be clear about what you expect that person to bring to the table. If the role includes building agents and automating workflows, assess whether candidates can investigate a problem and recognise where their understanding runs out. Then give them access to experienced colleagues and a real process to improve. Otherwise, employers risk hiring for confidence with a tool and mistaking it for readiness to redesign the work.

Bridie Rees is the senior people business partner at APJ at monday.com.

Want to see more stories from trusted news sources?
Make HR Leader a preferred news source on Google.