There's much excitement about how AI can explain anything for us. And that includes how the business works and how to manage it better. 

I have said before that any enterprise is a system that we designed. We know all the pieces that make it up, and all the processes by which those pieces work together. OK, some elements are not totally in our hands - like why customers and staff behave as they do. But we even understand those well enough to work with.

We already have an explanation for how the business works

And when we put together a digital-twin, dynamic business model (DBM), we are capturing exactly that system - all of its parts and how they work together.

Most causal relationships in a DBM are simply stated in everyday language, easily restated in trivial arithmetic.

Here's the model from last time (explained here), about how under-hiring hits service quality, annoying customers, hurting our reputation and later hitting our revenue. 

Here's the first few lines of that model ... 

"customers(this week)" = "customers(last week)"  + "customers won/week" - "customers lost/week"
"customers won/week" = "customers won/week if OK reputation" * "reputation" 
"customers lost/week" = " customers" * "% customers lost/week"/100

What's that in total? - just 23 items explaining some complex reality. OK, it's not quite so simple. The expressions need a little more to work right, e.g. to make customer losses escalate faster as reputation falls. Some relationships may be "looked up" rather than calculated, e.g. how customer losses rise as their annoyance level increases. And we may need some simple research, e.g. how exactly does customer annoyance build up as they suffer poor service experiences? ... but we would have to do those things for a spreadsheet model in any case.

... and the explanation is in near-natural language

Notice how those 3 lines in the dynamic model are very nearly the natural human language we would write down.  We could even build a very simple parser that would, for example, replace ' * ' with 'multiplied by' or replace ' MIN (X, Y)' with 'whichever is the lower of X and Y'. Then we would have written out a completely accurate - and quantified - story of how the situation actually works. 

Here's the relationship between a DBM, a spreadsheet alternative and a written-out story of how that system works.

Notice those limitations below both the spreadsheet model and the written story? The dynamic model breaks all of those problems, and more.

• It beats the spreadsheet because it is in near-natural language, clearly shows what-causes-what, is less error-prone, and its visualisation makes even big models easy to understand.
• ... and it beats the written story because - although that written story is also quantified - the model actually computes, so it can be verified, and it can play out how the whole system actually works and performs, under any range of scenarios and strategies.

Implications for AI ...

So we already have a model that matches the real world - in content, structure and behaviour. It is already expressed in natural language, and that language is easily be parsed into every-day speech. So what more can AI offer? 

Let me know at contact@strategydynamics.com

___________

The working model above can be saved and adapted to your own case.
It is included in a set of several marketing models that you can get here.
Use coupon blog33 at checkout to get these models for just £26.60.