I was reading a twitter post which I think misses the mark. It focuses just on velocity, but with little concern on value (as in customer satisfaction) or revenue.

Many points are valid if you just want to move faster on a classic software dev cycle, but what for?

I no longer think that number of features == value, with cheap AI adding features is relatively easy. Each added feature increases the friction for existing users.

What if each customer got their version of the product?

One size does not fit all

This is not as weird as it sounds. Enterprise software today is stupidly complex because it needs to be customized for each customer’s workflow. But that flexibility is inflicted on all customers. All of them need to deal with all the features.

What if it was built just for you? One release per customer.

Granted it would take a radically different software architecture and development cycle, but maybe, that’s where agentic development will shine.

A different architecture

You would need to think rather than having a product, having a product archetype. That’s what you produce, and then you tweak it for each customer using agentic development built on a library of components and services.

Each customer gets it’s own release, it’s own release cycle and if something is built for one that could be sold again, you can promote it as a shared component/service and be up-sold to existing customers.

Focus on customer satisfaction

It requires a completely different relationship. Each customer rep is a product manager. The focus should be on making customers happy. You own the expertise of the vertical and how to maintain a fleet of products, related, but all different. The goal is to delight customers (and make money as you do it).

I think it could work, specially for enterprise software with a end-user facing component.