When IT budgets are cut, the same items usually go first: device refresh, maintenance contracts, training and the time of whoever runs IT. The saving shows up in the books straight away. The cost it causes is spread across working time, outages and risk, and never appears as a line of its own.
A worked example with open assumptions
The following values are assumptions. Replace them with your own:
- Cost per employee including payroll overhead and workplace: 80,000 euros a year, about 47 euros per hour at roughly 1,700 working hours.
- Ten minutes of waiting per day caused by slow devices, sign-ins and applications.
- 220 working days, 50 employees.
Ten minutes on 220 days is almost 37 hours per person per year, about 1,700 euros at 47 euros an hour. For 50 employees that comes to roughly 86,000 euros a year. A business laptop costs 1,000 to 1,500 euros depending on specification and lasts several years.
Whether it is five or fifteen minutes in your organisation can be measured: through sign-in times, application start times and tickets about performance. Only measured values turn the example into a basis for decisions.
Where the money is actually lost
- Devices past their planned lifetime: longer waits, more failures and operating systems without security updates. Since 14 October 2025, organisations only receive them for Windows 10 as a paid service through the Extended Security Updates programme.
- Licences nobody uses: Microsoft 365 is paid for, Outlook and Word are used. Teams, SharePoint and automation stay idle because nobody introduced them.
- Manual processes: data is exported, edited in Excel and typed in elsewhere. Every one of those steps costs time and introduces errors.
- An overloaded IT lead: whoever carries operations, support and projects alone does not get to documentation, testing and recovery drills. That only shows when something breaks.
Outages and security
Reduced maintenance and postponed updates lower running costs and raise the likelihood of an outage. What an hour of downtime costs depends on which systems are affected and whether customers notice. That figure should be known for the most important systems before their operation is cut back.
The same applies to security. Multi-factor sign-in, current systems, tested backups and one recovery drill a year are not large items. Without them, recovery and root-cause analysis after an attack take considerably longer.
AI requires order first
Microsoft 365 Copilot works with the data and permissions an organisation already has. If sharing is too broad or storage is disorganised, Copilot also surfaces content a user is allowed to open but should not see. Anyone who saved on permission design and information structure has to catch up before introducing AI.
What Copilot is worth in your own organisation is shown more reliably by a pilot with measured results than by any general percentage.
Cut selectively, not across the board
- Measure: evaluate sign-in times, device age, licence usage and ticket categories from the tools you already have.
- Quantify: record what an hour of downtime costs for the most important systems.
- Separate: cut spending without benefit, such as unused licences and orphaned cloud resources, and put the money where waiting time or risk is measurable.
- Repeat: review once a year with the same figures.
This way, part of the necessary investment can often be funded from the items that were cut.
Where an architecture review comes in
For organisations running their IT on Microsoft 365 and Azure, an architecture review derives these figures from existing data: device inventory, licence usage, cloud cost and security configuration. The result is a prioritised list of what to cut, what to postpone and where to invest. Implementation stays with your team or your service provider.