Where practitioners
shape the future of work

Practitioner: Caroline Tolond
Role: Head of Careers, Employability, Enterprise & Alumni
Institution: Arden University

Problem: Arden runs a lean careers team and has no physical front desk, with a distributed student population that does not fit the traditional campus model. Demand for support arrives at all hours, including substantial out-of-hours traffic, and there is no realistic way to staff a human-only service to meet it. At the same time, pointing students at the free version of ChatGPT is neither equitable nor effective. The challenge is to provide tailored first-line support at volume, protect student data, and avoid widening the digital divide between confident AI users and others.

Solution: The Careers, Employability and Alumni team implemented a specialist careers chatbot, built in partnership with a provider following review of their academic and ethical approach. Arden considered building an in-house bot but concluded the team did not have the time or capacity. The licensing model offered a scalable way to expand first-line support capacity within existing resource constraints.

Over a two-year period, the chatbot handled more than 4,000 individual conversations across multiple universities. The contract has now been renewed for a further three years, and a beta version using conversational large language models is in testing. Crucially, the bot’s footprint is expanding. It will move beyond the public site into Arden’s virtual learning environment and the Target Connect careers management system. This puts first-line support exactly where students are already working rather than asking them to come and find it.

At Arden, the chatbot sits alongside complementary AI products. CareerSet provides automated CV and interview feedback, and Graduates First supports preparation for online testing and psychometrics. Where the chatbot cannot resolve a query, it can refer students into the advising team to book an appointment or signpost to other sources. This speaks to a wider introduction of AI technologies at Arden, where the team is encouraged to explore appropriate uses of AI within clear professional and ethical boundaries.

Considerations: A shiny product on its own does not generate engagement. The bot delivers first-line support at volume, but uptake depends on staff being visible in the curriculum and on the ground actively directing students toward it. The value of human guidance is preserved, with the chatbot helping to free adviser capacity for more complex and developmental conversations.

Equity of access is central to the case for paying for products rather than defaulting to free tools. The licence costs for curated products worked out cheaper than buying every student a month of premium ChatGPT, and the curated route avoids leaving the least-resourced students to fend for themselves. Students at every kind of institution, including highly selective ones, struggle to use general-purpose AI well without first knowing what good output looks like. This is a careers education problem that bots alone cannot solve and must be considered.

Want to know more? Contact Caroline

Practitioners: Professor Mirjam Hauck & Dr Kristen Reid
Roles: Academic Staff
Institution: The Open University

Problem: The dominant conversation about AI in higher education has focused on two things: academic integrity and productivity. Both treat AI as something to be managed rather than understood. What has been largely absent is a framework that takes seriously the social, cultural, political, and equity dimensions of AI. This is particularly true for the students most likely to be disadvantaged by the uncritical adoption of AI. For an institution like the Open University, where a third of students come from widening participation backgrounds, and where a significant population study in environments without reliable internet access, the stakes of getting this right are especially high.

Solution: Professor Mirjam Hauck has developed a Critical AI Literacy (CAIL) Skills Framework grounded explicitly in equity, diversity, inclusion, and access. In turn, the framework moves beyond asking what AI can do and into who it serves, who it excludes, and what it demands of people whose experience it was not built around. The CAIL framework was developed in collaboration with university colleagues from the OU Gen AI Learning Design Team, the OU’s Careers and Employability Service and Associate Lecturers. Professor Emeritus Mike Sharples, from the OU’s Institute of Educational Technology, was the advisor.

One of the framework dimensions is dedicated specifically to AI Careers. Here, the framework provides concrete examples of how career-relevant critical AI literacy can be embedded into learning and teaching, and it is designed to be transferable across disciplines rather than tied to any particular subject area. Indeed, the framework has been adopted by Dr Kristen Reid for a new introductory module in the Business undergraduate programme. Dr Reid is using the framework as a way of having students engage with AI in their studies from the outset by thinking about how new tools can be navigated critically and collaboratively.

The broader argument of the framework is captured in a shift the OU team describes as moving from policing to pedagogy to partnership. Where policing focuses on what students cannot do with AI, and pedagogy focuses on how to teach AI use, partnership means developing AI literacy with students. This process centres student voice and co-creates understanding of what AI can and cannot do.

The framework is now explicitly referenced in a white paper that recently went to the OU’s Senate, outlining ten principles for AI at the institution. That it took time for the OU to formally adopt its own researcher’s work reflects the challenges inherent in large institutions. Nonetheless, the framework has been adopted elsewhere and is actively being used by many higher education institutions in the UK and internationally.

Considerations: Implementing a framework of this kind requires institutional conditions that go beyond individual enthusiasm. Staff development, protected curriculum time, and senior backing must all be considered. The OU’s experience suggests that uptake outside the originating institution can sometimes outpace internal adoption. For careers practitioners interested in applying the framework, the AI Careers dimension offers the most direct entry point and can be introduced incrementally rather than requiring wholesale curriculum redesign.

Want to know more? Contact Mirjam

Practitioner: Maria Olariu
Role: Employability and Careers Practitioner
Institution: Middlesex University

Problem: Careers teams are under pressure to give students meaningful AI skills, but the discourse on campus tends to default to either policing academic integrity or generic digital skills training. Neither approach helps students understand how AI is actually used in the workplace, nor gives them the hands-on confidence they need to enter a sector where AI knowledge is increasingly expected.

Solution: Maria co-designed the Emerging Professionals Programme with a partner AI startup, creating a structured two-week experience that moved students from conceptual understanding into production work.

The first week was delivered online and brought together a range of professional speakers covering how AI intersects with career development across different industry sectors. The week concluded with an online wrap-up session to consolidate learning.

Weeks two through four shifted to the company's premises for three full days of practical work. Students were placed into interdisciplinary groups drawn from across the university, combining different skills and subject backgrounds by design. Their task was to develop an original idea for a video podcast and then produce it. The company provided studio space, camera equipment, editing software, and skilled staff to support them. When it came to the editing stage, students worked collaboratively and used AI tools to complete the work.

Students who performed strongly were fast-tracked into the Horizons programme, a separate initiative that placed them into real-world project work with the company, giving them a direct pathway from the programme to substantive work experience. One student went on to work with the startup's creative team directly. Those who completed the programme also went on to participate in a Google AI training, thereby continuing the learning beyond the programme itself.

A showcase event hosted during a university conference allowed students to present their work to senior figures from both higher education and industry, providing a high-profile platform for the cohort's output. Several students have continued developing their podcasts independently, with episodes visible on LinkedIn. All have maintained the networks they built during the programme.

The entire programme ran at no cost to the university. All speakers, the company's premises, the equipment, the follow-up placement opportunities, and the Horizons places were offered on a voluntary basis. Maria led the project in collaboration with a representative from the startup, with the wider team taking turns on delivery, chairing online sessions, and attending on-site days.

Considerations: This is a high-effort project that requires a strong industry partner willing to commit time and resources on a voluntary basis. The replicability depends significantly on the relationships you bring to it. Building in clear impact measurement from the outset, rather than relying on student feedback and participation data after the fact, would strengthen the case for future iterations. Ensure any follow-on work experience or placement activity complies with your institution's employment and GDPR policies.

Want to know more? Contact Maria

Practitioner: Chris Webb
Role: Careers and Employability Practitioner
Institution: University of Huddersfield

Problem: Chris has been running careers sessions for student groups from diverse courses, and the preparation demands were significant. Producing personalised, cohort-specific resources — such as mock assessment scenarios, sample CVs, and practice interview questions — consumed a disproportionate amount of time. The institutional AI provision was Copilot Basic, which he found useful but limited. Three years ago he took out a personal paid ChatGPT subscription, using it entirely for his own practice on non-proprietary work.

Solution: Chris uses ChatGPT in three distinct ways: for resource creation, for building custom tools he uses in guidance sessions, and for coaching students to use AI as a practical research tool themselves.

For resource creation, ChatGPT has transformed preparation time. What used to take hours can now be turned around quickly and edited to fit the group in front of him. Running a mock assessment centre for postgraduate business students one week and computing engineers the next no longer means building from scratch each time. Chris estimates this saves him around four hours per resource cycle. He labels all AI-generated materials to say which tool was used, and treats this as an opening for conversations with students about transparency and what responsible AI use looks like in practice.

For one-to-one guidance, Chris has built two custom GPTs he uses alongside students in sessions. The first is a career exploration bot that deliberately avoids showing job titles. Students often fixate on a specific role, which closes down thinking rather than opening it up. Instead of showing job titles, the bot takes the skills, interests, and values that a student shares during the session and generates broad areas of work, example industries, and organisations. Chris uses this as a launching point, then typically continues building out the picture with the student. The second custom GPT generates practice interview questions across six thematic areas drawn from years of running mock interviews. He uses it both for his own preparation ahead of practice interview sessions and directly with students to open up a conversation about which questions feel most relevant to their particular background and situation.

Finally, for student research, Chris has started being more explicit about how students can use ChatGPT as a starting point for employer and industry research. A recent example involved a student who wanted to stay in Greater Manchester and focus on interior design. Rather than sending them away to search manually, Chris worked with them in session to use ChatGPT to scrape the internet and to build an initial employer spreadsheet. The student then supplemented that base with their own industry knowledge and connections. The same approach is being explored for the placements team, where a student currently spends time each week manually trawling job boards to compile a vacancy email list — a task that could be structured as an automated daily output using the same method.

Chris is careful about digital equity when demonstrating these tools with students: the model features he uses require a paid subscription, and he actively explores with students what is achievable on free-tier access and what alternatives exist.

Considerations: Chris uses ChatGPT on a personal subscription and has made a deliberate decision not to use it for anything involving student data, which stays within the institutional Copilot environment. Any practitioner replicating this approach should be clear about the same boundary. When demonstrating paid ChatGPT features to students, be transparent about what requires a subscription and what is available on free accounts.

Software: ChatGPT

Want to know more? Contact Chris

Practitioner: Matt Lewis
Role: Careers Consultant
Institution: Middlesex University

Problem: Middlesex University has a small careers team that relies on student assistants to triage students seeking support to the correct services. Without specialist knowledge, those assistants often struggled to find the correct Careers Consultant or service, especially because many Careers Consultants work across multiple schools. When a student seeking support asked about job searching in their specific industry, or needed to know which consultant covered their course, the assistants often struggled.

Solution: Matt built a searchable knowledge base in Google's free NotebookLM tool. Student assistants can now type any question into the notebook during a triage conversation and get an instant, accurate answer drawn only from those documents.

Matt built the tool using documents he had already developed. The knowledge base brings together three types of content. First, industry guides that cover job boards, professional bodies, and career pathways for every subject area the team supports. Second, a staff allocation spreadsheet showing which consultant covers which programme at undergraduate and postgraduate level. Third, a catch-all reference document for edge cases, such as ensuring veterinary nursing students are directed to the right sector rather than towards NHS roles. Because NotebookLM only draws on the documents uploaded to it, answers stay relevant and institution-specific. The response length can also be adjusted to suit the pace of the conversation, with shorter answers for quick queries and more detailed responses when needed.

Matt has also built a personal version for his own specialism in the arts, loading in detailed programme guides that give him on-demand access to the depth of knowledge he has built up over years. The entire project runs on the free version of NotebookLM and required no technical background to set up.

Considerations: When using software like NotebookLM, consider data privacy and GDPR, making sure that institutional guidelines are met and that private data is not shared.

Software: Google NotebookLM

Want to know more? Contact Matt

Practitioner: Séan Richardson
Role: Careers and Digital Learning
Institution: London South Bank University

Problem: LSBU has an exceptionally diverse student body, with higher-than-average proportions of part-time students, students with caring responsibilities, and those working alongside their studies. Due to this, students needed flexible, self-paced access to careers support that traditional one-off workshops could not provide.

Solution: The team built the LSBU Career Skills Award, a structured digital employability platform that students work through at their own pace, receiving a digital certificate for each level they complete.

The award is divided into three levels: Core, Advanced, and Expert. Core covers the foundational skills every student needs to apply for any role, including CV writing, cover letters, and basic interview technique. Advanced targets the skills needed to secure a graduate-level position. Expert prepares the most motivated students for the full demands of the modern hiring process.

Importantly, the Skills Award automatically offers personalised resources. With no action required from the student, the platform detects their course and serves a tailored version of the award. A Drama student sees guidance on the audition process, whereas a Civil Engineering student receives sector-specific advice on technical skills. Academics can also request a customised version for classroom use, embedding employability directly into the curriculum.

AI tools are integrated throughout, including CareerSet, which requires students to score at least 70% on their CV before they can book a consultant appointment, ensuring students arrive prepared and consultants can focus on deeper work. The platform also integrates AI-powered custom ChatGPTs that guide students through CV writing and simulate job interviews at the point of need.

Using Symplicity CareerHub, the team further tracks which students have not engaged with careers services and triangulates this against demographic data to identify those least likely to secure a graduate-level role, enabling targeted outreach before it is too late.

The platform was co-created with students on work-based learning placements in the Careers Hub, who road-tested it and shaped the final product through their feedback.

Considerations: This project requires a significant investment in an EdTech platform as well as ongoing staff time to build, maintain, and update content. Course-specific personalisation requires coordination with academic departments across the institution.

Software: Symplicity CareerHub, CareerSet

Want to know more? Contact Séan

Read the full report: LSBU Career Skills Award Report