Don’t Automate Yet

Automation has been all the hype for over a decade now. Millions of jobs are predicted to be automated away thanks to advanced robotics, more capable software, and more computing power becoming available every minute. Automation, though, is not a solution to every problem. There are downsides to it that should be weighed carefully. Before choosing to automate any task, a significant effort should be applied towards figuring out why it’s important to automate the task and what the risks are. If a job task, process, or other reporting) can be eliminated instead of being automated, that is the better alternative. If the job would benefit from the flexibility and intellect of a human being, automation might not be a good idea in that case either. Like any tool applied incorrectly, there can be downsides to automating:

1. Automating a bad process does not lead to cost savings or gains in efficiency
2. Automation requires continuous investment in upgrades and maintenance
3. Robots and software are not as flexible as people
4. Automated tools do not have ideas — they can only do exactly what they are told to do
5. Support for existing tech tools wanes as new technologies are invented
6. We lose control of processes when rely too heavily on technology we don’t fully grasp, created by outside vendors we don’t control

Automating repetitive tasks with cheap software is a quick way to save money and a choice that is made every day. Inevitably, there will be corner case scenarios that automated processes cannot handle properly. Before choosing to commit to automation, the process needs to be understood in great detail and simplified to its necessary components. The way Frederick Taylor and the Gilbreths we read about earlier broke work tasks down into basic elements to better understand how long a given workload should take, anything being automated needs to be understood at that level of detail.

Assembly line work has benefited greatly from robotics in recent years. Videos of any car manufacturing facility, for example, will show dozens of robots being used to move doors onto vehicle frames, robots painting cars or welding the frame together and even robots (computers really) driving the vehicles. Note, though, that those are all separate robots which are programmed to do a very specific task. The only way the robots on the assembly line work is by assigning them a small piece of the whole puzzle. I.E. the robot that puts doors on the car frame cannot also paint the vehicle (unless it is reprogrammed and repositioned). To go from an entirely human assembly line to a fully robotic one will take years; only small bits and pieces are moved to automation at a time. Jumping right into a fully automated process would cause more headaches than anything, despite the perceived benefits.

When it comes time to consider automating tasks or work in your field, before jumping right into it, it’s important to narrow down a very specific process to automate before looking at automating the entire thing. A few examples of where this often goes wrong;

1. Automated outbound messaging for sales, recruiting, account management, business development, or anything similar often relies on variable placeholders like a person’s name, company, or job title. Most tools will pull that information from a database or profile somewhere, e.g. LinkedIn. Most automated tools can’t handle any non-standard format, though, and will result in errors, devaluing the outbound message’s intent. The solution most teams implement for this kind of thing is human quality control, which often defeats the purpose of automation.
2. Automated reporting on KPIs, sales numbers, warm leads, accounting / finance numbers or social media followers often results in too many reports that no one reads. Automating too many reports is a waste of effort if they are not being used for anything. Instead of trying to use more automating to make the reports look more interesting, eliminating them might be a better alternative.

The idea of automation, namely to get work done without having to do any of the work, is an appealing one. The risks to automating need to be understood before committing to expensive software, consultants, or other technology products that are outside the control of your organization. So, before automating, seek to understand.