AI;DR (AI, didn’t read) is a new term calling out AI slop. Frank Landymore’s piece in Futurism puts it bluntly: why should audiences put effort into consuming what someone else didn’t put effort into creating?
I see a parallel with AI transformation in supply chains.
Two problems:
We have the data quality problem. Everyone knows it sucks, but nobody wants to own it. Bad data is up there with alcohol when it comes to influencing terrible decisions. With data governance nowhere to be found.
And we have the process problem. Supply chain operators live in the world of workarounds because of rigid tech. Franken-workflows of haphazardly attached, mismatched parts that somehow get the job done. Even though the datasets talk to each other less than Sam Altman and Dario Amodei.
These are symptoms of a broken operating system that AI won’t automatically fix.
Feeding AI agents with bad data is like feeding Mogwai after midnight. You get gremlins tearing up your operations: self-multiplying and gleefully destructive, with no impulse control and sharp teeth.
Left unfixed, the risk isn’t zero adoption. It’s disruption and distrust that is painful and expensive to eliminate.
AI does however provide an opportunity to reimagine something better. To rebuild from the ground up.
But it needs a purpose.
Which is why you start with business outcomes.
Well-defined outcomes inform metrics and drive use cases, which dictate requirements. This highlights where current workflows deliver value, and where they fall short. Use these as building blocks to create better workflows. Start small and iterate.
Anything that truly delivers value requires time and focused effort. Building a foundation of trusted data and well-defined processes is no different. It isn’t the exciting part. But it’s the key to making everything else work.