Getting things right requires iteration. And we all know that requires running screens  --  they engage users far more than documents or wireframes.

Framework complexity leads to this unpleasant situation: months of effort to get running screens, only then to uncover a misunderstanding. Ouch. Low code can help, but it still takes time to design databases and paint screens.

The marriage of GenAI technology with Logic Automation means you can now describe a system with a prompt:

And get a complete running system in one minute  --  a database with test data, working screens, and an API. Even a fun little landing page:

We are moving veryrapidly.

Now, our team can review the screens and identify changes. These might be structural ("customers have multiple addresses") or logic ("it needs to check the credit limit").

So, we define "check credit" logic using declarative rules expressed in Natural Language and iterate:

And now that's running, also in about a minute. Suddenly, iteration cycles that used to take months are happening in minutes. The time wasted in misunderstandings  --  they still happen, of course  --  is inconsequential.

We are moving very rapidly, iterating in the right direction.

The logic above, while stated in Natural Language, is immensely powerful. It's automatically ordered (simplifying iterations) and is automatically reused over all the relevant use cases. So, perhaps conceived for placing an order, our logic also understands how to process selecting a different product:

The five lines of logic above apply to about a dozen use cases. Without logic automation, this would be about 200 lines of code.

That's a 40X reduction, for the backend half of your system

So, we've applied automation to all the elements of our architecture: the database, the screens, the API, and the logic.

The simplicity of GenAI  --  it's just Natural Language  --   means we're no longer limited to rocket scientists. Anyone can help get the requirements right.

Organizational velocity: more people, moving faster, in the right direction.

We are not claiming, by any means, that GenAI can automate complete enterprise-class systems. You will want custom front ends, integration with business partners and internal systems, etc. This means developers absolutely need to get involved.

With... what? All we've provided is a few prompts...

As you might have expected, a project has been created. It's a standard Python project you can download and extend in your favorite IDE, completely configured and ready to run.

Developers are cautious about generated "Franken-code", for good reason. It can be monstrously difficult to understand, debug and extend.

So, GenAI-Logic creates models, not code.

The initial project creation included an API with OpenAPI doc:

Because it's a model, the API is four lines:

And here is the logic - five Python rules. It's still at the same high level of abstraction, but you can now use code editors, debuggers, and source control:

Note the automation process did not just generate "raw" Python code. It would be around 200 lines. And we tried it: basic GenAI creates code that is poorly optimized, and wrong (it missed the corner case above). FrankenCode, at its worst.