You know how they say an apple doesn’t fall far from the tree when referring to children and parents? Automation is kind of like that too.
Automation works best when you can think from the outside and work your way inwards, rather than taking your existing process and just mimicking it. The problem that I see is that most businesses have a bad process to begin with. Automating it only makes it run faster. It doesn’t fix the underlying problem.
You know, there’s a lot of excitement around AI and automation today, so everyone’s natural inclination is to take what they’re doing, put a bot or a process on it, slap it together, and call it automation, which is technically right. It is automation. But I think step one would be looking at the business process itself and asking, why do we do certain things that we do? If we have a process that takes eleven steps, maybe we could trim half of them anyway. Are they there because it’s something that’s just been there forever, or something someone came up with months and months ago that we’ve just continued to do just because?
If you think about it, whenever you build a broken process and then automate it, that broken process happens faster, but it’s still a broken process.
What I’m actually seeing
Usually, I come to you here with a story about how I’ve seen a broken process being automated. But automation is so new that I haven’t walked into many businesses doing it. In fact, I haven’t walked into any business doing automation yet.
I will say that I’ve worked on a couple. You could call them prospects. They’ve explained how their automations are working, and let me tell you, it’s quite scary. They’re relying on these broken processes and potentially broken automations to run their business now.
A lot of them rely on the vendor they’re trying to automate against, whether it be their ERP system or some other system, staying the same. For instance, if a vendor moves a button from one page to another page through a normal update, their automation process is gone. It only takes them finding out that maybe orders didn’t ship, or money didn’t clear in the bank, that something broke. Then they have to figure out why.
Why this happens
So then we ask, why does this happen? AI makes automation super, super easy because it can find a way to get it done. Although we look at AI as something that thinks, it doesn’t really put in the effort you’d want it to. When you say, “Hey, I need this process automated,” it just says, “Yes, I’ll automate this process for you.” Not many times does AI come back and say, “well, should we be automating this, or should we be fixing that? Hey, by the way, this is going to break.” These are the things that I don’t commonly see with automation software, because their goal is to help you move faster. So the obvious answer is yes, I’ll automate this. In reality, they’re just setting you up for failure, in my opinion.
Then what happens? We review the plan and notice there are workarounds in there. A workaround someone might have invented three years ago, five years ago, or twenty years ago. No one knows why. A lot of people are afraid to make changes, so they leave it.
Another common issue is that the person who built the process isn’t the person who’s going to automate it. The person automating it is just taking the process and putting it into technology instead of using a critical mindset and asking, “Why did we do it this way or that way?” A lot of them won’t even involve the person who made it, for a handful of reasons. One reason might be resistance to change. Another might be that they no longer work at the company, or they think that they can do it and call it a day. But by involving the people who build the processes, you can ask them questions like, “Does this step need to be here? Could it be accomplished just by getting the data from this other place?” Right?
It kind of ties into the way that we build different platforms, software, or automations. Often we go back to pencil and paper. Forget the process. What’s the end result and what do we want done? Maybe the end result takes only three steps, even though you take a hundred. We can sit with the line-of-business owner, or the owner themselves, and say, “This was your end result. Here’s how I got to the end result. Now that I’ve done this, tell me, does this make sense? What am I missing?” They might be able to fill you in on details. Maybe some of the things that you think were workarounds aren’t actually workarounds. That’s okay.
But oftentimes we sit there, and we can trim out five of the steps. Maybe they’re going through this system to get the data, but they didn’t realize that the data actually originated from another system. So why not go to the source system instead of going to the middleman system? We might be asking someone questions on decisions that are obvious yes/no’s. We can build logic into the automation. (If it’s this threshold, we’re going to do this. If it’s that threshold, we’re going to do that.) To be quite frank with you, the error rate is probably going to be very little to none. We can account for those things too. We can have anomaly detection to stop a process and ask a question. That doesn’t have to be the daily frequent mode of automation.
What a broken process actually looks like
Let’s talk about what a broken process actually looks like. One easy way to spot it is by drawing it out on paper. You’re going to notice that there are steps that you can’t make sense of. There are steps you might not even be able to write out because you don’t know how they all work.
Some steps exist because something failed once, and you kept them in there because you never want that failure to happen again, even though it probably won’t ever happen again, since your business has changed altogether.
We’ve talked about data duplication. You’re probably entering data in two places by hand.
Another one is that you might have exceptions in there that span out more than the normal path itself. If that’s the case, that means your process is built around exceptions, not the process itself.
Then there’s the biggest tell: “so-and-so does it because they just know how to do it.” People shouldn’t have to rely on one person to know how to do a process, because if they’re gone, it takes a lot of training to rebuild. Instead, cross-train or set up the process so the right people can share the information.
How to fix it before you automate it
Here’s what you can do to fix a broken process before you automate it. Draw the process on a whiteboard or a piece of paper, and take a look at it as best as possible with a fresh set of eyes. Understand each step, and ask yourself: are these steps necessary?
Check whether information is entered in multiple places or pulled from systems that aren’t the direct source of truth.
Often, you can fix the process in person before you even automate it. That way, you can adapt to that flow before making an automation change. But not always necessary.
I would also say, make sure you’re automating all the stuff that’s boring, stale, or super repeatable, and leave the decision-making to people. At least to get started, until you can find a way to pattern that out as well.
The other side
Now, let’s talk about the other side of this. I usually try to bring some other truth into this to make it two-sided.
Sometimes you should automate a broken process, because the time you’re spending on it is killing your business. That’s totally understandable. I would say do that thing, but don’t stop there. The idea is to take it from there and make it better each iteration.
Also consider this. Not every process needs to be automated, because that process probably shouldn’t exist in the first place. You also have to take a step back and ask: what are the things that we’re doing that are just a waste of time in today’s world, because maybe we’ve adopted or adapted the way we do things.
The last thing I would say is, I’m advocating for you to fix it first. But if fixing it first means you never automate, or never become more efficient, then automate it. Don’t let fixing it first get in the way of making it better and faster. It takes discipline and encouragement here, and that’s where I’m trying to help you and say, go and do these things. Do them right, if you can, the first time. If not, make them right eventually. Don’t end up in the same broken process that you’re in now by putting band-aids on and calling it a day.
You know, automation is just a mirror. It looks at how you do things, and it just happens faster.
In two weeks, we will talk about something that interests me a lot: scammers. Scamming is getting harder to spot, and it’s not because your tools are getting weaker. Most scammers use the same tools you do. If that sounds interesting, subscribe so we can send it to you.
If you have a tech decision in front of you, like AI, automation, security, whatever it is, and you want a straight read on it, that’s the kind of thing I like doing. You can grab some time with me here.