Understanding how we estimate your potential savings and capacity uplift
The WorkSmart-AI Savings Calculator uses peer-reviewed and public research to estimate how much time, cost, and capacity your organisation could reclaim through staff training on AI.
The calculator has two modes:
- Standard Calculator – a quick, directional estimate for your whole department.
- Advanced Calculator – an optional detailed model that accounts for different academic and professional service roles.
Both are grounded in real evidence on productivity, task automation, and adoption patterns in higher education.
1. The Standard Calculator
What it shows
The standard calculator provides an overall picture of:
- Annual savings (£) – the estimated value of time saved through improved efficiency
- Productivity increase (%) – the percentage improvement in task completion across your department
- Capacity created (FTEs) – the equivalent number of full-time roles freed up through efficiency
- Five-year projection (£) – a forward estimate of cumulative savings
How it works
You enter three simple inputs:
- Number of employees
- Average annual salary (including typical on-costs such as NI and pensions)
- Target AI adoption level – basic, intermediate, or advanced (see Section 3)
The calculator then applies a realistic proportion of tasks that can be supported by AI tools.
It uses research on higher-education and public-sector productivity to model time saved at different levels of adoption.
Core assumptions
- Working hours: 1 FTE = 37.5 hours per week × 46 working weeks (allowing for annual leave)
- On-cost factor: 1.25 × salary (to include NI, pensions, and overheads)
- AI-eligible tasks: around 50–55% of total workload for most education and administrative roles
- Productivity uplift: 5–15% faster completion of eligible tasks depending on adoption level
These assumptions come from large-scale studies measuring real-world productivity gains from AI tools (see References section, below).
2. The Advanced Calculator
What it adds
The advanced calculator allows you to enter headcounts for different academic and professional service roles for a more precise result.
Each role has its own default ‘AI-eligible task share’, which reflects how much of its workload can realistically be supported or accelerated by AI.
What ‘AI-eligible task share’ means
An AI-eligible task share is the proportion of a person’s job made up of activities that AI can assist with directly, such as:
- Drafting or summarising text
- Writing communications or reports
- Preparing teaching materials
- Reviewing or analysing data and feedback
- Searching, referencing, or collating information
- Formatting or structuring content
It does not include work that requires in-person teaching, decision-making, or practical delivery that AI cannot (yet) perform reliably.
So, a role with a 60% AI-eligible task share spends roughly 60% of its time on activities that could be streamlined using AI tools.
These values are derived from research showing that roles involving writing, analysis, and communication see the greatest time savings from AI.
Role categories and data foundations
| Role type | Example departments or functions | Typical AI-eligible task share | Rationale |
|---|---|---|---|
| Educators | Teaching and learning | 55% | Lesson prep, resource creation, and feedback benefit strongly from AI support. |
| Researchers | Research staff, postdocs | 50% | Literature review, summarising, and drafting papers show measurable gains. |
| Educator-Researchers | Combined teaching and research roles | 53% | Balanced mix of academic and administrative tasks. |
| Library & Learning Resources | Librarians, learning developers | 50% | Metadata, summarisation, and research assistance tasks are moderately AI-amenable. |
| Student Services / Registry | Admissions, registry, student experience | 60% | High proportion of communications and form processing. |
| HR | Recruitment, onboarding, policy | 55% | Document creation, reporting, and communication tasks benefit significantly. |
| Finance | Accounts, payroll, procurement | 40% | Gains mainly in documentation and reporting rather than numerical analysis. |
| IT Services | Helpdesk, documentation | 35% | AI helps with communication and documentation but less with technical build work. |
| Marketing & Communications | Content, press, digital | 65% | High drafting and creative workload makes these roles highly AI-eligible. |
| Estates / Facilities | Operations, maintenance, planning | 15% | Limited desk-based content, lower automation potential. |
| Other Admin | Governance, secretariat, planning | 45% | Moderate written and reporting work. |
When you enter counts for these roles, the calculator replaces the single headcount figure and weights savings by role type, producing a blended productivity and capacity estimate.
3. Understanding Target Adoption Level
What it represents
The Target Adoption Level shows how widely and confidently staff are expected to use AI tools in their everyday work.
It captures both reach (how many people use AI) and depth (how effectively they use it).
High adoption doesn’t just mean more licences – it means AI is built into daily workflows.
| Typical organisational stage | Description |
| Early stage | A few pilot teams or individuals are experimenting with AI. Use is limited and training is minimal. |
| Scaling up | Large or multiple teams use AI for communication, reporting, and planning. Training and guidance are in place. |
| Mature adoption | Large-scale integration. Staff use AI routinely, supported by clear policies and embedded workflows. |
Why higher adoption levels lead to greater productivity
Each step up the scale increases productivity for three reasons:
- More people benefit – adoption spreads across teams, so small time savings multiply.
- Staff use AI more effectively – training improves tool use and reduces friction.
- AI becomes part of the workflow – efficiency gains are built into systems, templates, and processes.
| Adoption level | Approximate adoption rate | Typical efficiency uplift on eligible tasks | Example overall productivity gain (with 55% AI-eligible work) |
|---|---|---|---|
| Basic | 35% | 8% faster completion | 0.55 × 0.35 × 0.08 = 1.5% overall |
| Intermediate | 60% | 12% faster completion | 0.55 × 0.60 × 0.12 = 4.0% overall |
| Advanced | 80% | 15% faster completion | 0.55 × 0.80 × 0.15 = 6.6% overall |
These figures align with multiple large-scale studies that measured real-world improvements across different levels of AI maturity.
Research behind the adoption levels
| Source | What it found | How it informs the model |
|---|---|---|
| St. Louis Fed (2025) – Measuring the Impact of Generative AI on Work Productivity | Employees using AI reported a 5.4% reduction in weekly work hours. | Forms the conservative baseline for Basic adoption. |
| BCG & Harvard Business School (2023) – Navigating the Jagged Technological Frontier | Consultants completed 12% more tasks and were 25% faster using GPT-4. | Supports the Intermediate range. |
| Microsoft Work Trend Index (2024–2025) | Organisations saw 10–20% faster completion of drafting and summarising tasks once adoption scaled. | Reinforces Intermediate to Advanced assumptions. |
| OECD (2025) – Unlocking Productivity with Generative AI | Found consistent 5–15% productivity gains in knowledge work. | Validates the overall 5–15% range. |
| Fitzpatrick et al. (2025) – Assessing Gen AI Value in Public Sector Context | Civil-service teams were 34% faster on document tasks and 17% higher quality after full AI integration. | Illustrates the upper bound of Advanced adoption. |
| McKinsey (2025) – AI in the Workplace | Organisations with structured AI training achieved up to three times higher productivity improvements. | Confirms that training and adoption quality drive impact. |
How this applies to universities
- Academic staff often experiment individually but need structured support to use AI confidently in teaching and research.
- Professional services teams (registry, HR, marketing) adopt faster when guided by workflow-specific training and governance.
- Technical and compliance teams (IT, finance, estates) require secure systems and policy frameworks before wide adoption.
Choosing a Target Adoption Level lets you model different stages of readiness – from cautious pilots to confident, organisation-wide adoption – and see how training and engagement affect potential savings.
4. KPIs and How the Calculator Works Them Out
The calculator produces four key results:
- Annual cost savings (£)
- Productivity increase (%)
- Capacity created (FTEs)
- Five-year projection (£)
Here’s what each represents and how it’s calculated.
1) Annual cost savings (£)
This estimates the total value of time saved through improved efficiency.
Formula (plain version):
Annual savings = Number of staff × Average loaded salary × Productivity improvement
Example:
If 50 staff each earn £50,000, with an on-cost factor of 1.25 and a 4% productivity improvement:
50 × (£50,000 × 1.25) × 0.04 = £125,000 annual savings
2) Productivity increase (%)
This measures the overall improvement in working time.
It combines three factors:
- The proportion of work that can be supported by AI (AI-eligible task share)
- The proportion of staff who will use AI (adoption rate)
- The average efficiency improvement on those tasks (uplift)
Formula:
Productivity increase = Eligible work × Adoption rate × Efficiency gain
Example:
0.55 × 0.60 × 0.12 = 0.0396, or 4% total improvement.
That’s why the calculator shows “Based on 12% efficiency @ 60% adoption”.
3) Capacity created (FTEs)
This converts time saved into full-time-equivalent staff.
Formula:
Capacity created = Number of staff × Productivity increase
Example:
60 staff × 0.04 = 2.4 FTEs
That means the time saved equals roughly 2.4 extra full-time roles.
4) Five-year projection (£)
This projects annual savings over five years.
If savings stay constant:
Five-year savings = Annual savings × 5
If you add annual growth (for scaling or improvement), the calculator compounds the savings each year to reflect that change.
Worked example
Using these inputs:
- Efficiency on eligible tasks: 12%
- Adoption rate: 60%
- Eligible task share: 55%
- Number of staff: 60
- Average salary: £45,000
- On-cost factor: 1.25
Productivity improvement = 0.55 × 0.60 × 0.12 = 4%
Annual cost savings = 60 × (£45,000 × 1.25) × 0.04 = £133,650
Capacity created = 60 × 0.04 = 2.4 FTEs
Five-year projection = £133,650 × 5 = £668,250
| Metric | Result | Description |
|---|---|---|
| Annual cost savings | £133,650 | Based on 12% efficiency @ 60% adoption |
| Productivity increase | 4% | Average across all entered staff |
| Capacity created (FTEs) | 2.4 | Equivalent full-time roles of capacity |
| Five-year projection | £668,250 | Total projected savings over five years |
| Assumptions | 1 FTE = 37.5 hrs/week | 46 working weeks per year |
5. Data and Evidence Sources
| Area | Evidence base |
|---|---|
| Productivity uplift (5–15%) | St. Louis Fed (2025); OECD (2025); Fitzpatrick et al. (2025). |
| Adoption levels (35–80%) | BCG (2025); McKinsey (2025). |
| Task eligibility (~50%) | Henseke et al. (2025); McKinsey (2025). |
| Higher education adoption | Microsoft (2025a); HEPI (2025); UNESCO (2024). |
| Cost avoidance and back-office efficiency | Harvard / Wharton (2023). |
| Error reduction and quality control | Harvard / Wharton (2023); PwC (2024). |
| Salary and working patterns | ONS ASHE (2024). |
6. Interpreting your results
The calculator’s figures are directional, not definitive.
They illustrate the scale of potential benefit if AI is adopted effectively, supported by training and governance.
Even a conservative 5% productivity gain can reclaim about 3,000 staff hours per year in a 40-person department – roughly one full-time role’s worth of additional capacity.
7. Why we focus on opportunity, not ROI?
Every organisation is different.
Rather than publish ROI figures, the calculator shows the potential scale of savings and capacity uplift, so you can identify where the biggest impact lies and discuss tailored investment levels during consultation.
References
- BCG (2025). AI at Work: Momentum Builds but Gaps Remain. Link
- BCG & Harvard Business School (2023) – Navigating the Jagged Technological Frontier. Link
- Boston Consulting Group and Harvard Business School (2023). Navigating the Jagged Technological Frontier. Link
- Fitzpatrick, N. et al. (2025). Assessing Generative AI Value in Public Sector Context. arXiv:2502.09479. Link
- Harvard / Wharton (2023). ChatGPT and Quality of Work. Link
- Henseke, G. et al. (2025). Generative AI Susceptibility Index (GAISI). arXiv:2507.22748. Link
- HEPI (2025). Student Generative AI Survey. Link
- McKinsey (2025). AI in the Workplace. Link
- Microsoft (2025a). AI in Education Report. Link
- Microsoft (2025b). Microsoft Work Trend Index. Link
- Noy, S. and Zhang, W. (2023). Experimental Evidence on the Productivity Effects of Generative AI. Science. Link
- ONS ASHE (2024). Annual Survey of Hours and Earnings. Link
- OECD (2025). Unlocking Productivity with Generative AI. Link
- PwC (2024). Risk Assurance Insight Series: AI and Quality Control. Link
- St Louis Fed (2025). Measuring the Impact of Generative AI on Work Productivity. Link
- UNESCO (2024). AI and Digital Education Programme. Link