Why the Best AI Market Research Needs Expert Knowledge and Judgment

Building MyTelescope taught us that great AI market research needs more than data and workflows. It needs the judgment of experienced researchers and strategists.

2026-11-03

When we first started experimenting with machine learning and neural networks at MyTelescope, some of what we imagined felt like Star Trek. Something Data might do, but probably nothing we would get to build ourselves.

Now I keep catching myself taking those same things for granted.

Reading this piece on expert knowledge in AI brought me back to something we learned while building MyTelescope: getting AI to produce an analysis is one thing. Getting it to approach a problem like a good researcher or strategist takes much more.

That’s why we built around their knowledge, alongside our data. How they frame a question, challenge an assumption, interpret a signal and decide what to investigate next.

Workflows capture part of that. But two people can follow the same steps and reach very different conclusions.

The difference often sits between the lines. Knowing when the brief is asking the wrong question. When a number needs more context. When something that looks unimportant changes the whole interpretation.

That judgment is difficult to put into instructions. Often, the expert doesn’t even realise how much they’re bringing to the problem until you start asking why.

For me, that’s one of the most interesting parts of building AI today. Learning how to make that experience and judgment available to others, in the moment they need it.

A few years ago, that felt like science fiction too.

Back to blog