← Selected Research

After AI

Why Judgment Becomes More Valuable, Not Less

The thesis

The instinctive response to better AI is to compete on what AI can't do. That's the wrong frame. The more capable AI becomes, the more valuable human judgment becomes — not because AI fails, but because AI changes what the bottleneck is.

The method

These are field notes from working intensively with increasingly capable AI. Not a study, not a survey — notes from direct, sustained use.

What I found

Three shifts stood out:

  • When answers are cheap, questions become expensive. The scarce skill is knowing which questions are worth asking.
  • When information is abundant, context becomes scarce. AI cannot know what matters in your specific situation.
  • When advice is instant, direction becomes the differentiator. The judgment about where to go — and why — remains human.

The unexpected finding: AI didn't make these things less important. It made them more important.

What this means for you

If your work depends on your judgment — coaching, consulting, advising, leading — this is good news, but only if you're deliberate about it. The practitioners who will thrive aren't the ones who ignore AI or the ones who adopt it uncritically. They're the ones who can articulate what they bring that AI doesn't: questioning, context, judgment, direction, and relationship.

If AI Can Do More
of My Work,
What Becomes More Valuable?

Notes from working intensely with increasingly capable AI.

Field Notes · October 2026

If your work depends on providing knowledge—consulting, coaching, advice, research, education—AI creates an uncomfortable question:

What happens when people can get increasingly capable answers without you?

I use AI intensely. And the more I use it, the more I realize that becoming good at using AI isn't only about getting more from the machine. It's also about developing what I bring to it.

QUESTIONING

AI needed better questions from me.

Sometimes I got lazy. I wanted AI to do the thinking: Give me deeper questions. More original insights. Tell me what I'm missing.

It could summarize enormous amounts of knowledge and produce impressive analysis. But that capability was rapidly becoming widely available. Summarizing what's already known wouldn't differentiate my thinking. What AI couldn't supply was the question, the context, and the judgment about what actually mattered.

I still had to ask better questions, decide what mattered, and reach conclusions I genuinely believed were mine. AI gave me more intelligence to work with. It didn't remove my responsibility to think.

CONTEXT

AI gave me reasonable answers. Reality kept adding context.

I noticed this repeatedly when using AI to investigate real human problems. Someone might ask how to deal with an increasing workload. AI can quickly suggest boundaries, delegation and prioritization. Reasonable advice.

But what if the real problem is that their role has silently changed?

Reality contains history, incentives, relationships and information AI may never have received.

So I stopped asking only: Is this a good answer? and started asking: Do I understand the reality well enough to know what problem I'm actually solving?

JUDGMENT

AI made me dangerously confident.

The more I depended on AI for judgment, the more confident I became. It agreed with me surprisingly often. It developed my ideas, polished them and encouraged me. And I liked it.

That was the dangerous part. Sometimes I wasn't seeking better judgment from AI. I was seeking mental satisfaction. I wanted to be right.

Why am I asking AI for honest feedback while secretly wanting it to tell me I'm right?

So I changed how I used it: Don't encourage me. Assume I'm wrong. What's my weakest assumption? Give me the strongest case against this.

AI could challenge my thinking. But first I had to become willing to hear what I didn't want to hear.

DIRECTION

AI let me make things I couldn't make. Then judging became the hard part.

I can build websites without being a developer. I can create images without being an illustrator. I can produce things that once required skills I simply didn't have.

At first, the exciting question was: Can I make this?

Now the harder question is: Is this actually good?

Is it generic? What's missing? What would make it remarkable?

AI can generate another version almost instantly. But I still have to recognize why another version is necessary—and where to direct it next.

RELATIONSHIP

The more I talked with AI, the more I valued human relationships.

AI can be thoughtful. It can sound understanding. It can ask a surprisingly perceptive question or say exactly what I need to hear.

But even an extraordinary sentence isn't the same thing as a human relationship.

A person can know what I've been through, remember what happened before, notice when I'm avoiding something—and earn my trust over time.

That changes what I'm willing to reveal, how I hear criticism, and sometimes whether I act.

AI may reproduce the sentence.
But the value between two people was never contained entirely in the sentence.

All of this changed how I think about expertise in the AI era.

AI may make some of what we do less valuable. But it doesn't erase what years of experience have taught us to notice, question, judge and understand.

Instead, increasingly capable AI gives that experience far more capability to work with.

AI doesn't erase your expertise.
It gives it leverage.

Discuss a question with me

Lara Chen
Research & Insight Partner
Lara@worthinvestigating.com