In order to turn a square into a circle, there are some awkward shapes that take place in the process.
We are currently a hybrid square-circle when it comes to AI adoption for business operations.
Recently, I’ve had the opportunity to build processes that save people minutes to hours a day, and it’s interesting to see the challenges with these systems beyond the technicalities.
So I’m documenting them here to share what I’m seeing.
First, technology is getting really good and moving extremely fast. Speed of technology affects adoption of efficient systems. One day, you build a system around doing multiplication by hand with your employees. The next day, the calculator comes out. The technical capability of a calculator is tremendous. You can now do multiplication at scale. But how does the calculator fit into your system now? Is the system you built for doing multiplication completely useless now? How do you create a new system that properly leverages the calculator, knowing that a computer may be built in the near future?
I think there needs to be more emphasis on the distinction between technical capabilities and adoption within the AI hype. I feel that society overemphasizes what technology can do.
Getting technology into the hands of people such that it improves our lives, and understanding where we are in this adoption, deserves more focus. This is where systems and operations come into play. You need a system to adopt the technology, and you need to build a system using the technology such that its impact is meaningful.
Let’s look at a modern example:
AI is very capable of sending personalized emails to 100 people every hour.
How do you set up a system to send personalized emails to 100 people every hour? Do you use Mixmax? Do you use Apollo? Do you build from scratch?
Let’s say you pick Apollo, but Apollo doesn’t personalize well enough or measure the metrics you want to measure. How do you build a system to solve these last-mile issues?
I see the world entering a new era within the workforce. The old workflows of manually managing Excel sheets and slide decks are going to be beyond us soon. We are seeing these legacy business workflows being phased out with tools like ChatGPT, Canva AI, and other AI-fluffed tools out there. We are moving away from systems of heavily manual work to nearly fully automated work. But, in order to turn a square into a circle, there are some awkward shapes that take place in the process.
And here’s the tricky part.
How do you convince/lead/manage your organization of squares to convert to a circle?
Your organization knows the circle state is the ultimate ideal state, but your organization has been a square for so long that converting to a circle feels awkward. Being this hybrid circle-square is awkward. It wastes time. It seems to move us away from our day-to-day focus on things that actually move the needle.
In practical terms, the circle state is some dream-state AI-automated workflow era where AI does 99% of our work. But how do you get people to unlearn manual workflows that have been ingrained in their heads their entire career, to adopt better new systems? And how do you build a system for the transition, such that it is smooth and natural?
I have no idea.
The thing with system building is that you don’t know if it’s strong until months to years down the line. Some systems seem great when they’re built, but unexpected systemic issues can still arise. And systemic issues can get really complex. Just look at the US government. Or any government, for that matter.
Here are some things about building business operations and systems I’ve learned:
Make sure the system is fully thought out.
Proposing half-thought-out solutions can be the downfall of business operations because you spend months implementing a system that doesn’t give you the solution you want. I’m learning to “fully bake” an idea about a system before finalizing on one. Thinking about common workflows, edge cases, and tackling problems from first principles.
Effectively communicate the purpose and value add of a new operational workflow.
When you “fully bake” your system idea, articulating it to the rest of the team should be a piece of cake because you’ve put some calculated thought into how and why this process makes everyone’s lives easier. When people resonate with the intention, they are more likely to put effort into adopting the system.
There will be edge cases in your biz ops system.
And that’s a tradeoff you have to make. Small outliers are not worth readjusting a workflow system for. If the system solves 80% of your problems, it usually isn’t worth spending another 80% of your time to account for the last 20% of your outlier problems. It’s better to get things done than think about how one can do it more efficiently. Instead, try to build a system that handles edge cases gracefully. However, if you work in an environment that requires a small margin of error, this doesn’t apply. Build a perfect system instead. Can’t be that hard!
Measure the effectiveness.
One of the things that makes operations tricky is that it can be difficult to directly correlate some process to getting more revenue, customers, etc. Thoughtfully mapping out how this process increases some KPI and saves people time is important because tangible metrics are what encourage system adoption and define a system’s impact.
In conclusion, here is my take on all of this:
Building efficient operational systems is all about saving employee time so that the company can make more money.
If you cannot build efficient systems and directly correlate these operations to making more money, implementing AI to improve workflows serves no purpose but to make you feel like you’re remaining relevant.