Businesses lose a quarter of the week to admin

More than a quarter of every working week is being lost to admin that should not exist. That is the finding of a study examining 680 processes across 128 companies in 10 sectors, which found that businesses lose 26% of working time to wasted effort. For a typical firm in the study, 424 weekly hours contained 110 hours of fully recoverable time. The technology to reclaim most of that time already exists. The harder question is which processes to target first.

Where the working week disappears

The research identified five process categories responsible for most of the waste: reporting, onboarding, invoicing, approvals, and CRM clean-up. These are not edge cases or niche departmental quirks. They are the daily workload of almost every finance manager and operations manager in a UK SME.

Each category shares the same root cause: manual effort applied to tasks that follow predictable rules. Someone exports a spreadsheet and re-keys the figures into a different system. Someone chases an invoice that is sitting in a shared inbox. Someone waits for an approval that could be routed automatically. The work is real, but the human involvement is largely unnecessary.

At a fully-loaded rate of $100 per hour across a 46-week year, the study estimates these inefficiencies cost an average business around $510,000 annually. That figure is denominated in US dollars as presented in the source research, but the underlying process waste is directly comparable to what UK finance and operations teams report.

Why this is an 80/20 people problem, not a software problem

It is tempting to treat admin waste as a tooling failure. Buy better software, the thinking goes, and the hours come back. But the research pattern tells a different story. Most of the companies examined already had the platforms to automate these tasks. The waste persisted because the processes had never been mapped, the rules had never been codified, and nobody had made adoption a priority.

This is gecco's defining observation about AI adoption: it is 80% people and culture, and 20% technology. A new automation installed into a team that does not understand it, trust it, or know how to maintain it will be bypassed within weeks. The five process categories identified in the research are not difficult to automate technically. The challenge is getting the right people to define the rules, test the outputs, and own the results.

For UK finance managers and operations managers, this reframes the question. The issue is not whether your business has access to automation tools. It almost certainly does. The issue is whether your team has a clear picture of where the waste sits and a practical method for addressing it.

What AI automation actually does in these five areas

Targeted automation approaches each category differently.

In reporting, AI can pull data from multiple sources, apply consistent formatting, and distribute outputs on a schedule. The finance manager who spends two hours each Monday assembling a cash-flow summary can receive a draft that needs only a ten-minute review.

In onboarding, structured workflows route new-joiner tasks to the right people automatically, track completion, and send reminders without anyone managing a checklist manually.

In invoicing, AI can match purchase orders to receipts, flag discrepancies, and route approvals based on value thresholds. The operations manager chasing three outstanding invoices across two inboxes is replaced by a system that escalates automatically after a defined period.

In approvals, rules-based routing removes the bottleneck of a single decision-maker. Requests below a certain value are approved instantly. Those above it go to the right person with the supporting context already attached.

In CRM clean-up, AI identifies duplicate records, flags stale contacts, and prompts updates based on recent activity. The data quality work that was skipped because nobody had time for it happens continuously in the background.

None of this requires a wholesale platform replacement. These improvements sit on top of tools most SMEs already use.

Making this work for your business

The practical starting point is a process audit focused on the five categories above. Finance managers and operations managers are best placed to identify where manual effort is highest, where errors occur most often, and where delays create downstream problems.

Once the target processes are identified, automation can be introduced incrementally. A single invoicing workflow or approval chain can be live within days. The team learns by doing, builds confidence in the outputs, and extends coverage when it is ready.

As part of its AI Assistants service, gecco builds specialist assistants that sit inside the platforms your team already uses, such as ChatGPT or Claude, and connects them to the workflows that move data between your systems. This approach addresses the admin categories the research identifies without requiring your team to switch platforms or learn new tools from scratch.

The 110 recoverable hours the study identifies per week are not theoretical. They represent real tasks that real team members are completing manually right now. The question for most businesses is not whether automation would help, but where to begin.

Implementation safeguards

Automating reporting, invoicing, and approvals involves processing data that may include personal information. Under UK GDPR and the Data Protection Act 2018, businesses need a lawful basis for that processing, clear records of what is automated and why, and appropriate safeguards when automated decisions affect individuals.

This does not make automation harder. It makes the design process more deliberate. Mapping a process carefully enough to automate it is also mapping it carefully enough to document it for compliance purposes. The two activities reinforce each other.

Finance managers should confirm that any automation touching payroll, supplier data, or customer records is reviewed against their data protection obligations before going live. In most cases the compliance work is straightforward. But skipping it creates exposure that undermines the efficiency gains.

A further consideration is error handling. Automated processes need defined routes for exceptions. An invoicing workflow that fails silently on a mismatched purchase order creates more work than the manual process it replaced. Build in alerts, audit trails, and human review points for edge cases from the start.

Starting with the right process

The research covers 680 processes across 128 companies. The consistent finding is that waste clusters in predictable places. That is useful information for any UK operations manager or finance manager trying to justify an automation project internally.

The business case does not require a custom study. The data already exists. The task is mapping your own version of those five categories, quantifying the hours involved, and identifying which one carries the highest return for the least complexity.

For most SMEs, that is invoicing or approvals. Both follow clear rules. Both have defined inputs and outputs. Both produce measurable results within a short time of going live.

Where to go from here

If your finance or operations function is carrying admin weight that your team has learned to treat as normal, this research makes a useful prompt for a structured review. The AI Readiness survey is designed to surface exactly where those inefficiencies sit and what level of automation readiness your business currently has.

Take the AI Readiness survey. You will receive 65+ free resources and a custom AI Readiness report based on your answers. From there you can book a free 30-minute showcase to walk through the results with a gecco adviser.

---

Website · LinkedIn · Case Studies · Newsletter