Most finance leaders have adopted AI, yet hours are still lost to manual data entry and rework. iplicit CFO Rob Steele explains why all-purpose AI fails without solid system foundations, and how UK finance teams can finally reclaim their evenings.
In Partnership With
“If AI is so clever, why am I still slogging away in spreadsheets at this time of night?”
You might have asked yourself a version of that question as another day’s work started to stretch into the evening.
I’m an AI enthusiast – but I can see why many of my fellow finance professionals are sceptical about its productivity-boosting potential. They might feel differently if they could see the technology giving them back the precious moments that long working hours can snatch from us.
That might mean being available for a small child at 6pm rather than 7.30 when an irreplaceable experience has passed.
It could mean finishing work in time to enjoy a film, show or meal out on a weeknight – instead of closing the laptop in exhaustion, ready to consume convenience food and whatever is fed to you by the streaming algorithm.
Or it could mean starting the weekend on Friday evening – rather than on Saturday, after you deal with those “few loose ends” that end up taking half the day.
My experience of using AI daily, for years, has shown me it really can help give you that elusive work-life balance. But it can disappoint unless you put the foundations in place to really benefit from it.
If you’re using AI to some extent but still seem to be putting in as many working hours as you did before, you’re not an outlier.
iplicit’s own research among 250 mid-market finance leaders last year uncovered a striking contradiction. On the one hand, 83% reported that they’d adopted AI, partly or fully. On the other, 42% cited manual data entry as a top drain on their time. Almost as many (40%) cited reconciliation of accounts and transactions, while 28% named invoice processing. On those figures, AI is not freeing up people’s time the way they might have hoped.
There’s other research to back up those concerns. A 2025 study by MIT Media Lab suggested that 95% of organisations running pilots were getting “zero return” from integrating generative AI with their existing systems.
One reason time savings can fall short of expectation is that AI’s work can require a lot of verification. A study by Workday and Hanover Research found that nearly 40% of time saved by all-purpose AI tools was then lost to “rework” – the process of correcting errors, rewriting text and verifying outputs. Naturally you’re going to be disappointed if you end up giving back 24 minutes of every hour you thought you’d saved.
If you’ve adopted AI without reducing your mountain of Stuff To Do, that isn’t necessarily a reflection on your personal efficiency. It may be that you need to address some fundamentals first.
The fact that many organisations in our research were still doing manual data entry, despite adopting AI, is revealing. In my experience, finance teams often take all-purpose AI tools and bolt them onto clunky systems and processes.
Before going all-in on AI, it’s worth asking where the inefficiencies lie in the work of your team or organisation. Is data being keyed manually from one system to another? Do processes rely on people answering email or signing paperwork? It’s worth addressing these inefficiencies first.
As for your finance system, is it giving you complete and accurate data without a lot of data entry and spreadsheet work? If not, it’s a brake on your productivity. While you can recoup some of that lost time by using AI in other areas, you’ll always be trying to mitigate inefficiencies that remain unaddressed. At iplicit, we’ve seen plenty of finance teams save several days of staff time each month by improving their core system. One organisation, Walking With the Wounded, saved 15 days every month in a small team.
By addressing processes and systems first, you’ll maximise the benefits of AI. You’ll also improve AI accuracy because you’ll have a sound system of record – a “deterministic” system, where the same input will always produce the same output. AI tools, such as those built into iplicit, will be able to produce reporting and analysis that’s traceable to data in the ledger, reducing that rework burden.
AI on its own is not the key to being at the school gates at 4pm occasionally. Precious moments like that might lie on the other side of a system upgrade you’ve been putting off for years.
For more resources on using AI within your team, click here.
