On this page
There are two types of companies that use AI.
Company A aims to use AI to do existing things faster. Now work that took a day runs on itself. And then, they use the freed up time to watch more TV.
Company B aims AI to scale their outcome. Each time AI freed up more time, they reinvested the time in finding better strategies, and how to do things on a bigger scale: "Can we serve 10x more customers? And potentially earn 10x?", "We did 5 experiments last month, can we try 100 next?" They repeat this process over and over again, every time AI removes a bottleneck, their ambition gets bigger, and they gain experience to scale the outcome.
Company B will be successful in a short time while Company A will perish in 6 months.
If you prefer Company A, you will not like Zortex. Zortex's goal of using AI is to scale outcomes by 100x. In other words, we leverage AI to make the previously impossible possible.
Think in terms of years, not days
Some view AI like a productivity tool to work faster. They never scaled their operations, never thought of how to do more. If you use AI this way, you are thinking in terms of days.
View AI as a long term investment: first it's going to be bad, then it gets better as time goes.
Guide towards using AI for scale
If something satisfies all three conditions listed below, you should just do it, you don't need any permission.
1 Scaling this might resolve a business constraint
Scaling for the sake of scaling is stupidity; if someone can't choose what to scale, he won't do it well. Ask this question: if we scaled this process 100x, or 10,000x, would we unlock outcomes or possibilities that weren't possible before?
For example, making a monthly plan 1000 times won't do much. But 10,000x the process of collecting potential leads, researching industry and company specific pain-points and producing personalised emails for potential markets could change the game.
2 Long termism on the learning curve
See the learning curve in years, both AI's and yours. Something AI can be super terrible at today could be great in 2 years. Ask this question: if AI continues to improve at the rate it is today, is it possible for us to do 10,000x of this work 2 years later? How about 5 years?
3 Control
Every scale without control becomes a disaster. The limiting factor of scaling is control.
Before using AI, someone must understand it well enough to tell: (1) is the result unusable? (2) is quality degrading? (3) when there are multiple problems, could they have a common root cause? (4) what are the primary control loops?
More importantly, can you still do it at 10,000x scale? Scale should be built progressively, as each time you gain experience. Do not go from 0 to 10,000x all of a sudden. Aim for 10x the previous iteration is enough.
Learn to drive before AI takes over
If you haven't learned how to drive, you will not be allowed to drive on the road today just because the car is self-driving. Every country has this rule, because if it goes wrong, you need to take control. Similarly, you need to learn the work before you use AI.
In short, the handbook to using AI is:
- Understand how to do the work.
- Build a proof of concept of doing the work with AI.
- Using that knowledge, build systems that scale work 10x.
- Using that knowledge, build systems that scale work another 10x.
Meanwhile, improve the AI workflows in Zortex AI hub and accumulate re-useable, modular parts, so every system we build makes the next ones easier.
