When this new era of AI hit the headlines – and, by that, we generally mean LLMs and creative coding engines that made AI a far more user-friendly and useable tool for the average business – it was accompanied by plenty of stern warnings that AI would soon replace all of us. Years on, that hasn’t happened – and shows no real signs of happening – and the alternative is actually far more beneficial to those businesses.
The alternative we speak of? The ability to utilise AI strategically in order to minimise repetitive, mundane, laborious tasks through automations. Things like generating reports, logging into multiple platforms throughout the day just to check data, and scraping information from competitors for hours on end – tasks that seem robotic, in other words, turned out to be ideal for…well…for the robots.
But AI still feels like a topic that is as long as a piece of string for many, which means businesses that are flooded with the everyday tasks have little time for figuring out exactly how to deploy it. Here’s what you need to know about the use cases for automation in your back office.
The Roles Ready to Evolve
Ops coordinators and junior analysts used to spend many hours of the day – if not entire days of the week – logging in and out of different platforms, scraping information and generating reports on information that, while fundamental to business continuity, didn’t require a huge amount of expert input.
Now, fewer job listings mention those familiar old terms like ‘data entry’. Instead, it’s about overseeing automated workflowers – essentially, being the human in the loop who ensures that the higher quality aspects of the work meet the right standard, which is far more impactful for the business than the repetitive tasks these roles used to be eclipsed by.
High frequency tasks with a low requirement for judgement and clear right/wrong parameters are ideally suited to automation.
Some businesses are, of course, proving slower to catch up than others. A large part of this stems from a fundamental misunderstanding of the way automation runs, and how it can be integrated seamlessly into the existing framework, since most of this doesn’t run through some new company-wide platform, but through web browser automation. It’s software that operates a real browser to click, type, and navigate the way a person would have in the past. The internal portals, vendor dashboards, and legacy systems this work touches were mostly never built with an API in mind, which is exactly why scripting or API integration alone hasn’t been able to touch them.
What’s Not Ready to Evolve?
There’s a huge amount of hype and faith being generated around AI right now, but there’s also a lot of skepticism, since most of us can reel off a few instances where we interacted with an AI – whether an agent, an LLM, or a chatbot – that demonstrated a profound lack of understanding or comprehension. Hallucinations and goofs mean that caution is needed, but only for certain roles.
For that reason, tasks that require novel judgment mid-task aren’t yet suited to automation in the same way that repetitive, low-judgement tasks are. We simply need to take a step back, review what’s draining most of our time and tying-up human expertise in mundane tasks, and set the automations there.
