01/12/2025
Welcome back! In this series on the five actual truths about AI in the workplace, we’ve already looked at topics like trust and personalisation. Now, we’re talking about something every organisation feels as the year draws to a close: pressure. Budgets tighten. Performance targets stay high. Teams are short- staffed and still expected to keep the business running efficiently.
This year, a lot of organisations are exploring how AI can take some of that pressure away. Especially in high-pressure industries like banking, insurance, manufacturing and retail, companies are experimenting with AI tools that can automate routine tasks and help people spot issues earlier. But, like with any exciting, new technology, there is such a thing as overdoing it. Over-automation, rushed deployments or unwavering trust in a model can create more work than they remove.
So, how can organisations strike the right balance? Below, we’ll look at how you can use AI to boost efficiency with a common-sense approach, without losing good judgment, quality or trust along the way.
The endless quest for efficiency
Since the beginning of the Industrial Age, efficiency has become the holy grail in the business world. As the scale of mass production grew in the early 20th century, famous industrialists like Henry Ford became obsessed with finding a single, optimal way of doing things. Ford famously invented the moving assembly line, slashing production time for the Model T from 12 hours to just about 90 minutes.
The efficiency mindset gave us major breakthroughs in productivity, like standardised workflows and, eventually, tech-driven, automated processes. Still, no matter how ‘optimal’ a process is today, it can inevitably be made even more efficient later. Especially as new technologies like AI reach maturity.
Efficiency also means something different for each industry, and even for each organisation. For example, banks face rising operating costs, and some are dealing with old processes that slow everything down. Insurers face new regulations or claim volumes that grow faster than teams can handle. Retailers push through tight margins and ever-changing consumer preferences. Manufacturers work around supply chain interruptions and unpredictable energy costs.
In all these scenarios, there’s a strong use case for AI. But how do you make sure deployment actually improves efficiency? And how can you set the stage for lasting, sustainable efficiency, not just quick wins? It usually comes down to three things: data that reflects reality, workflows that are ready to adapt, and teams who understand what the AI is doing.
Are quick wins too good to be true?
AI success stories are starting to grab headlines, whether it’s faster mortgage processing at NatWest or accelerated insurance claims handling at Aviva. IT leaders see stories like these and expect their own initiatives to deliver fast, scalable results. We all want our AI investments to deliver quick wins. But the truth is: the companies that get the strongest, most sustainable efficiency boosts from AI are the ones who are keeping humans firmly in the loop and being realistic about what the technology can and cannot (yet) do.
In banking and insurance, for example, speeding up processes is only helpful when decisions remain fair and explainable. Regulators expect documented reasoning. Customers expect consistency. That’s why it’s crucial to prioritise transparency along with speed, and keep your teams well educated about what the tool does well, and what its limitations might be.
It’s also unrealistic to expect lasting value from AI if the data behind it is messy or incomplete. In retail, for example, an inventory management tool fed on limited data might perform well in the early days. But as soon as buying patterns shift, it will struggle. Rely on it too heavily, and you could wind up with stock that’s out of touch with actual demand.
In manufacturing, teams might also become overly reliant on predictive tools. Engineers might start skipping basic inspections because the new AI system hasn’t raised any alerts. When a fault appears that the model didn’t anticipate, the line stops and the team has to diagnose everything by hand. So much for improving efficiency.
The pattern is the same across industries. Quick wins are great, but only when the organisation maintains common sense and control, and keeps asking the basic questions: What is the model using? What is it assuming? What happens if it is wrong?
Making efficiency sustainable
Efficiency will always be a moving target. As soon as you improve one part of the business, pressure pops up somewhere else. AI can take some of the pressure off, but it only works when the right conditions are met. That means good data discipline, use cases that are ready for automation, and people who are trained on what it means to be the human in the loop.
Governance matters too. Map out who gets to use a tool, how often it gets checked and who’s keeping an eye on changes in the model. These things are easy to overlook when everyone is under pressure to deliver quick results, especially around the end of the year when budgets are stretched and teams are understaffed. But without proper governance from the outset, you’ll ultimately wind up losing efficiency because of confused customers or fixes that take longer than the original task.
Instead of racing for the quick win, allow your people time to learn what the models can do, and just as importantly, what they might get wrong. Encourage your teams to challenge the output and play an active part in improving the way it works. That’s really the most innovative thing about AI: its ability to receive direct input, learn from past mistakes and continually improve.
So, as the busy year-end period approaches, we can all look forward to a future in which AI will make our jobs more efficient. It may not be a cure-all, but it will definitely make real improvements in the way your teams work, especially if you build your AI strategy around accountability and human judgment from day one.
Get in touch with our dedicated team of experts, and start a conversation around your AI needs.
