In four weeks we show you where AI can be implemented in a chosen department, with results in hours and cost saved, value created and risk reduced per month, plus a 90-day plan to make it happen. Your people move onto higher-value work, not more work.
We modelled a fictional UK airport operator with £1bn turnover, 15 departments and about 3,500 staff, using cautious assumptions. These are illustrative figures to show the scale, not benchmarks or results from a real client.
Back-office teams with lots of routine, rules-based work tend to show the biggest share of time freed.
| Department | People | Share of time freed | Hours a year | Value of staff time a year |
|---|---|---|---|---|
| Finance | 95 | 31% | 49,280 | £1.69m |
| IT, Digital and Cyber | 150 | 22% | 54,400 | £2.23m |
| Commercial (Retail, Parking and Property) | 95 | 23% | 36,780 | £1.26m |
| People and HR | 75 | 25% | 30,640 | £1.01m |
| Procurement and Contracts | 45 | 25% | 18,890 | £748k |
Illustrative estimate. Annual figures at full potential, before AI running costs.
Every organisation is different. These six things make the biggest difference, and the four-week plan checks each one.
The example assumes little AI in use today. If your teams already use it well, some of the savings are already banked.
Copilot or AI built into your existing systems may already cover some tasks. We count each saving once.
Clear, repeatable processes are quick to improve. Messy or undocumented ones need sorting out first.
AI works best on data that's in one place and trusted. Scattered or poor-quality data slows things down.
Freed time is only real if people use the tools. Training and trust matter as much as the technology.
Security, regulation and your own governance can limit where AI is allowed. Clear sponsorship from the top speeds everything up.
Every board meeting someone asks what we're doing about AI, and there's still no straight answer.
Our best people spend most of their week copying numbers from one system into another.
We paid for a pilot last year and got a deck and a demo nobody uses.
Most AI pilots stall for the same reason. They start with a tool and go looking for a problem. The work gets added on top of people's day jobs, so they end up under more pressure, not less.
The Department AI Plan starts the other way round. We look at how one department really works, task by task, and only then decide where AI fits and where it doesn't.
The same four types drive the estimator above and the plan you get at the end.
Judgement, relationships and decisions that need a person. AI stays out.
AI does the first pass and a person reviews it and owns the result.
Routine, rules-based work that AI runs, with a person picking up the exceptions.
Work you couldn't afford to do before, plus the oversight that keeps AI honest.
Move the sliders to match one of your departments. The numbers update as you go. This is a rough, illustrative estimate built on cautious assumptions you can see and change. It isn't a quote or a promise. The four-week plan replaces it with your real figures.
Keep it human, AI drafts and a person checks, or AI handles it. Agreed with the people who do the work.
What each change would free up, with the assumptions written down so finance can check them.
What to switch on first, what it will cost to put in place, who owns it and how you'll know it's working. Built on a process model of how the department really runs.
Moving staff onto higher-value work, not handing them more of it, with the risks and running costs shown in plain sight.
A half-day kick-off with you and the department head. We map the department's processes and talk to the people who do the work.
Each regular task goes into Type 1, 2, 3 or 4. Where we can, we time a few tasks with and without AI rather than guess.
Hours and cost a month for each change, the one-off cost to set it up, the AI running costs, software you could drop and the risks you'd reduce.
We walk you and your team through the findings, test them together, then hand over the final plan.
No day rates and no creeping scope. We agree one fixed price for your department after a 30-minute call.
Hours a year = people × productive hours (1,650 unless you change it).
Hours freed = hours × (Type 2 share × 25% + Type 3 share × 70%), then minus 15% for the oversight AI needs.
Cost of an hour = salary × (1 + 30% on-costs) ÷ productive hours.
Value of staff time freed = hours freed × cost of an hour.
Lower headcount = value of staff time freed × the share you expect to take as fewer posts. The rest is people reassigned to higher-value work.
Added value = rework avoided (2.5% of Type 2 and 3 time) + new work at £45 an hour.
Net value a year = staff time freed + added value + software cancelled − AI running costs. Risk is shown on its own.
All figures are full potential. Plan on about a quarter in year one, as teams adopt each change. Monthly figures are the annual estimate divided by 12. Real savings and incidents won't arrive evenly month by month.
These are cautious starting assumptions, not industry benchmarks. Results vary with how much AI you already use, the software you already pay for, how settled your processes are, the state of your data, your people and your rules. The plan swaps every one of them for your own figures.
I'm Warren Bradley CEng MICE, founder of Ambrad Consulting. I've spent 28 years working in and alongside large client organisations, so I know how they really run and I've felt the same frustrations. I use AI every day to run my own business. The plan starts from how your work actually happens, not from a vendor's demo.
No. It's an illustrative estimate from the assumptions shown, to help you decide if a closer look is worth it. The four-week plan replaces it with your own figures.
Good, that's a head start. Tools are only half of it. The plan looks at the work itself, department by department, so the licences you already pay for actually take time off people's desks.
It doesn't have to. A single fixed-price department is usually small enough to buy without a full tender.
That's one of the things the plan tells you. Every change comes with its one-off set-up cost (connecting systems, training and support) and its running cost, so you can see how quickly each one pays back before you commit. The quickest wins often use software you already pay for. If you want us to build it with you, we quote a fixed price from the plan.
That isn't the aim. The plan is built to move people onto higher-value work, slow hiring or cut overtime. What you do with the freed time is your decision.
Bring your estimate. We'll talk through your department's biggest time drains and tell you honestly whether the plan is worth it for you.
Book a 30-minute call calendly.com/thecurve-wanbridge/30min