Workflow Wednesday: AI Builds Data Products

Art by @basilonmypizza: https://lnkd.in/eF8FkWzN - https://basilhefti.ch/

Today, AI can do many things. One capability stands out: coding.

And coding has a close relative: defining, creating, testing, and maintaining data products.

So, does AI work there too?

Yes.

It inspects raw data, infers semantics, identifies relationships, generates transformations, and validates results by cross-checking the data.

But wait - I am the expert in this field. Surely AI cannot do my work?

Think of the sewing machine. It produces stitches with speed and consistency. It does not decide on patterns, sizes, or what should be produced in the first place.

The same applies here.

We also know: machines can produce scrap at impressive scale, leaving you with a pile of unusable output very quickly.

In other words: The machine brings speed, breadth, and persistence. The human brings guidance, judgment, context, and monitoring.

Applied well, AI becomes a substantial accelerator for data work.

The important question thus is: how do we organise the work so that AI does this reliably? This is what we do.

• Art: Basil Hefti, https://lnkd.in/eF8FkWzN https://basilhefti.ch/