Liverpool John Moores University has enrolled 134 staff in an AI Academy run with training provider Multiverse, kicking off an ambitious push to embed generative AI into everyday work. The goal: reclaim up to 4.5 hours per week of administrative time per employee and redirect it to teaching, research, and student support. The program, reported today by EdTech Innovation Hub, marks one of the largest organized AI upskilling efforts in UK higher education.

The AI Academy: A Closer Look at the Training

The academy isn’t a single course. It’s a ladder of learning that starts with hands-on tool training and rises to strategic leadership. Staff from academic faculties and professional services are mixed into one cohort, a deliberate move to break down silos between research, administration, and support.

The initial intake spans three progressive levels:

  • Level 3: AI-Powered Productivity – Practical use of generative AI tools in daily work. Participants learn to work with Microsoft 365 Copilot and large-model assistants such as Google’s Gemini, focusing on automating routine tasks like drafting documents, summarizing meetings, and pulling data reports.
  • Additional Level 3 modules – AI for Business Value and an AI & Machine Learning Fellowship. These cover process redesign, ethical deployment, and turning data into actionable insights.
  • Level 5: AI Strategy and Leadership – Designed for managers and team leads who will oversee AI rollouts. The curriculum balances technical, operational, and ethical trade-offs, preparing them to govern adoption rather than just implement tools.

Multiverse, known for applied apprenticeships and corporate upskilling, structured the program around on-the-job learning and measurable outcomes. That matters because it pushes training beyond theory into workflows that real staff touch every day.

Breaking Down the 4.5-Hour Promise

The eye-catching projection is that staff who complete the academy and fully adopt AI-powered workflows could save an average of 4.5 hours a week. LJMU highlights four areas where this reclaimed time is expected to come from:

  • Automating reports and routine data pulls
  • Streamlining document preparation and drafting
  • Improving triage of enquiries and case work
  • Reducing manual note-taking through AI-assisted transcription and summarisation

It’s a compelling vision, but it’s also a projection, not a guarantee. The university and Multiverse haven’t published the full methodology behind the 4.5-hour figure—it draws on partner productivity estimates and early-use scenarios. For any institution planning a similar program, treating that number as a target rather than a factual baseline is prudent. Without a transparent, role-by-role baseline audit and controlled pilot results, the figure remains a well-informed estimate.

For everyday Windows users, the mechanics are familiar. Microsoft 365 Copilot already promises to summarize emails, generate PowerPoint slides, and analyze Excel data. What LJMU is attempting is a structured test of whether those promises hold at institutional scale.

What This Means for You: Lessons Across Disciplines

For Knowledge Workers

If you spend hours each week in Word, Outlook, or Teams, the LJMU experiment offers a preview of what’s possible. The 4.5-hour figure isn’t just an academic fantasy; it mirrors early productivity gains reported by organizations piloting Copilot. Automating report generation or document drafting isn’t about eliminating jobs—it’s about clearing the clutter so you can focus on higher-value work. The key takeaway: tool literacy matters as much as tool access. Without training, even the smartest AI assistant remains underused.

For IT Managers and Administrators

LJMU’s approach carries weighty governance implications. Rolling out generative AI across a workforce means confronting data protection (for UK institutions, UK GDPR applies), model selection, and vendor lock-in. The program’s explicit focus on “responsible AI” isn’t window dressing; it’s a necessity. If you’re planning a similar initiative, you’ll need to:

  • Classify data and decide whether models run on-premises, via enterprise cloud channels, or on public endpoints.
  • Negotiate contractual exit clauses and data portability with training partners.
  • Build human-in-the-loop review for any AI output used in decision-making.

The academy model also underscores a staffing shift: new roles such as AI ops and prompt engineers are emerging, while existing IT staff must upskill to manage AI deployments and audits.

For Educators

Generative AI changes assessment. When students can ask Copilot to draft an essay, traditional assignments become unreliable measures of learning. LJMU’s plan to redirect reclaimed time toward teaching and student support only works if assessments are redesigned. That means emphasizing process over product, integrating AI awareness into curricula, and training examiners to spot AI-generated artefacts—while using the same tools to offer consistent, evidence-based feedback.

The Bigger Picture: Why Universities Are Turning to AI Now

LJMU’s move didn’t happen in a vacuum. Three forces are converging to push higher education toward systemic AI adoption:

  • Financial pressure – Universities everywhere face tighter budgets. Efficiency gains that don’t harm core teaching are gold.
  • Student expectations – Learners want fast feedback, flexible contact, and modern digital tools. If AI can trim administrative overhead, more staff time can shift to direct student engagement.
  • Tool maturity – Two years ago, enterprise-ready generative AI was a concept. Now tools like Microsoft 365 Copilot are woven into the apps millions already use. The barrier to deployment has dropped sharply.

Rather than run isolated pilots, LJMU is treating AI as operational infrastructure. That signals a belief that benefits will come not from a few point solutions but from broad process changes across the institution.

A Practical Guide to Launching Your Own AI Upskilling Initiative

If LJMU’s experiment inspires you, don’t jump straight to enrollment. Here’s a step-by-step approach that prioritizes outcomes over optics:

  1. Start with a baseline audit. Measure how staff actually spend their time before any training begins. Without this, you can’t prove later savings.
  2. Define measurable goals that tie to student outcomes. Productivity metrics are only valuable if they translate into better teaching or support.
  3. Choose a training partner carefully. Vet vendor track records in higher education, financial stability, and ability to commit to multi-year delivery. Multiverse’s rapid growth is impressive, but institutions should insist on references and contractual guarantees around data handling and support SLAs.
  4. Build a cross-functional governance body. Include IT, legal, academic staff, student representatives, and data protection officers from day one.
  5. Run a controlled pilot. Start with a representative cohort and track outputs against a matched control group. Publish transparent results that show methodology and assumptions.
  6. Integrate tools into existing workflows. AI shouldn’t be a standalone stovepipe. If staff must switch between five apps to get value, they won’t.
  7. Create feedback loops. Collect staff and student input continuously, monitor usage patterns, and iterate on policy and training.
  8. Treat time savings as conditional. Require units to demonstrate how saved hours will be reallocated to teaching or student support before scaling deployment.

Throughout, keep governance front and center. Publish a clear acceptable-use policy, define escalation routes for misuse, and never allow unprotected student or staff data into public AI models.

Outlook: Beyond the First 134

LJMU’s initial cohort of 134 is just the start. If the academy delivers measurable results, the model will scale, and other institutions will follow. The true test won’t be whether staff report back hours—it will be whether those hours improve student outcomes. For IT leaders watching, the lesson is clear: AI adoption is organizational transformation, not a training exercise. The tools are ready; the challenge is weaving them into the fabric of daily work without breaking the trust that education depends on.