Let’s say you ask an AI model to write a story about 3 pigs and a big mean wolf.
It will likely generate a generic story very close to the old classic we all know about.
Now let’s say you ask a multi-agent AI system with a CEO, a VP of Marketing, Story Engineer and Senior Editor agent working in a continuous loop with access to advanced image generation tools.
You’ll probably get a more verbose version of the story (and of course some odd looking AI-slop imagery).
But actually, what you wanted was a story where the big mean wolf teaches the 3 pigs about AI safety by walking them through the jungle and encountering a large mecha-AI T-Rex (to make it more interesting for education purposes of course 😉)
You might say that the reason the AI missed the target is because the context, requirements and key details to build the story weren’t clear.
You may also question whether a complex agent organizational structure including a VP of Marketing is actually relevant for the task at hand.
... And that's the point!
↳ Simply installing a solution doesn't create value. Value comes from implementation and adoption.
The story example is a bit silly, but imagine you are upgrading a key system, implementing a new solution or launching a new product.
We are getting bombarded by claims that AI will automate everything and that agentic systems can solve all our problems.
The reality is that:
If no thought has gone into specification, AI will just bring us in the direction that it's training told it to, and that could be light-years from where we want to be.
As AI solutions gain in capacity, the deviation risk actually increases.
➡️ The further AI gets us, the more disastrous a 1-degree deviation from target will be.
So where do we go from here?
The good news is 👇
We already know how to do a lot of important things that AI can use as input.
Project management, collecting requirements, writing specifications, monitoring progress and correcting course, etc.
However we must come to terms that just 'using' AI will not magically fix our problems.
We need to invest into thinking and planning how to effectively leverage these new solutions to enhance and accelerate work... in the right direction.
If anything, the place to start is simple:
Be more intentional
Be more deliberate
Be more structured in how AI is used
Have you started bringing more rigor and structure into how your teams use AI?
Interested in your thoughts 💬
Feel free to DM to discuss further!