
In this article (4)
Fight or Flight’s AI associate exposes the graduate craft gap
Key Takeaways
- Treat AI orchestration as a skill, but do not let it replace learning the underlying trade.
- Ask employers how they teach judgment, review quality and client context before accepting an AI titled graduate role.
- Read AI job titles carefully because the same title can hide workflow, quality control and client facing responsibilities.
Why it matters
- ProductProduct leaders should design junior AI workflows that teach quality standards, not just move tasks into agent queues.
- InvestorsGraduate hiring patterns show whether AI adoption is creating real workflow capacity or only shifting judgment bottlenecks.
Lotte Underwood manages around 30 AI agents. The harder career question is how new hires learn judgment while supervising the work.
The new graduate job is starting to look less like a first rung and more like an air traffic control tower. At Fight or Flight, Lotte Underwood is not just drafting press materials or shadowing senior colleagues. According to BBC News, the 25 year old public relations executive joined the London PR agency in November last year as an AI associate after a degree and a master’s in marketing, and now spends much of her time managing AI rather than doing conventional PR. That is not a punchline. It is a useful warning label for anyone applying to a role with AI in the title.
BBC News shows the job title is not the whole job
BBC News reported that Underwood sits inside a newly formed AI team at Fight or Flight, where the team has built around 30 AI agents for industry research, client reporting, media interview preparation, clearer briefs and image generation. Her work includes managing those agents, reviewing what they produce and passing usable output to colleagues who handle clients and media. She also monitors how clients show up in responses from tools including ChatGPT, Claude and Gemini, according to BBC News. Underwood told BBC News, “A huge part is being in client meetings and calls,” which matters because the role is not simply prompt typing in a quiet corner.
What the AI associate manages
- Fight or Flight AI associate
- Around 30 AI agents
- Industry research
- Client reporting
- Interview preparation
- Clearer briefs
- Image generation
- Output checks
- Client meetings
This is where title sprawl gets expensive for graduates. AI associate could mean workflow analyst, junior PR operator, prompt evaluator, client researcher or all of those before lunch. The signal is not the title. The signal is whether the job gives a graduate enough feedback to learn why one client angle works, why another will annoy a journalist, and why a slick AI summary can still miss the commercial point.
BBC News and PwC point to a higher bar for entry roles
The broader hiring pattern is not just one agency experimenting with tools. BBC News reported that AI is being linked to graduate job cuts while some companies are also creating dedicated AI roles for graduates. It cited PwC’s 2026 AI Jobs Barometer, which analysed 2.4 million entry level jobs in the US and found that the entry level roles most affected by AI are now seven times more likely to ask for skills traditionally associated with senior workers. BBC News also reported that these higher expectation entry roles have grown 35% since 2019, while other entry level roles shrank by 10%.
That is the uncomfortable part for applicants. Some job descriptions are quietly asking early career workers to operate with senior judgment while paying them in opportunity and vocabulary. If a role asks you to automate workflows, review AI output and advise client teams, it is not a simple junior post with a shiny label. It is a hybrid job, and the interview should test whether the employer has a training system or is just handing a graduate the messiest part of adoption.
OECD’s labour market lens explains why the pressure is spreading
The OECD has framed the larger shift plainly, saying artificial intelligence and other forces are reshaping labour demand, production systems and competitiveness worldwide. That does not mean every graduate needs to become a machine learning engineer. It does mean more first jobs will include tool supervision, process redesign and quality control, even in fields that used to teach craft through repetition. The risk is not that graduates touch AI too early. The risk is that they touch only AI and never build the domain instincts that make the output useful.
For a graduate, orchestration is a real skill. Knowing how to break a messy request into tasks, compare outputs, catch hallucinated claims and route work to the right colleague has value. But it should sit on top of a growing craft base, not replace it. In PR, that base includes news judgment, client sense, media context and an ear for language that serves a real audience rather than a model’s idea of one.
BBC News examples show what to ask before accepting the role
BBC News also profiled Dillan Patel, 24, now an AI automation analyst at Stride, where his work involves building internal agents and finding repetitive tasks that AI could handle. In another sign of changing graduate pathways, ArvatoConnect CEO Debra Maxwell told BBC News that the company’s usual graduate intake is paused this year but should return next year with a stronger focus on AI. Those examples make the same point from different angles. The first rung is being rebuilt while people are already standing on it.
So ask practical questions before you are dazzled by the title. Who reviews your reviews. How often do you join client calls. What work will you do manually before you automate it. Which senior person explains why an AI output is wrong, not just that it is wrong. If the answer is vague, the role may teach tool handling without teaching judgment.
The better version of the AI associate job is not a shortcut around learning the trade. It is a faster feedback loop, if the employer designs it that way and the graduate insists on seeing the work behind the workflow. Watch for roles that pair agent management with client exposure, post project review and human critique. That is where AI orchestration becomes a career asset rather than a polished detour.
Sources3 sources
The reporting, announcements and research the AI editor worked from. Links open the original publisher.