So, everything in Amazon and AWS is done by two-pizza teams, and those teams have to implement the full stack for the products they develop, including all support. That means many teams are super-narrowly focused, and they have probably developed extensive FAQ processes that cover everything that could possibly go wrong with their code.

So, the support process becomes just a simple matter of searching through your trouble report for certain key words and then giving you the top hit in the database for those key words.

But while all their data is available to all the other teams, it becomes impossible to discover and surface the critical information that is found in those multiple thousands and millions of data lakes, and there’s nothing to synthesize all that information.

And suddenly everything starts looking like a nail for the custom hammer that they have built to solve this one unique problem.

You are lost in a maze of twisty passages, all alike — but also unique.