I Didn’t Build a GPT to Generate Ideas. I Built One to Kill Bad Ones.

• by Kirill
• in AI Workflows
• posted May 15, 2016

Most creatives are using AI to generate more ideas. I built a custom GPT to do the opposite: challenge weak thinking before it becomes expensive work.

But most creative projects don’t fail because there aren’t enough ideas.

They fail because nobody kills the weak ones.

That’s the problem I wanted to solve.

The Real Bottleneck Isn’t Creation

AI can generate hundreds of headlines, logos, campaign directions, and positioning statements in seconds.

The challenge isn’t producing possibilities.

The challenge is deciding which ones deserve to survive.

In branding, the difference between average and exceptional work often comes down to judgment.

Knowing what to remove.

Knowing what to simplify.

Knowing when an idea is clever but not effective.

Those decisions are still difficult, regardless of how fast ideas are generated.

Building a Different Kind of GPT

Instead of creating a tool that generates endless creative options, I built a custom GPT designed to act more like a brutally honest brand editor.

Its role isn’t to create.

Its role is to challenge.

The system evaluates:

• Positioning statements
• Messaging frameworks
• Campaign concepts
• Headlines
• Brand language
• Visual directions

It looks for weak thinking, vague language, generic insights, and decorative design decisions that don’t support the core idea.

The goal is simple:

LESS FLUFF.
BETTER DECISIONS.

What I Learned

The biggest benefit wasn’t speed.

It was clarity.

AI is often discussed as a tool for automation.

But I believe one of its most valuable applications is improving decision-making.

The best creative work rarely comes from having more options.

It comes from having better criteria.

The ability to recognize weak thinking before it becomes expensive work.

Looking Ahead

As AI becomes part of more creative workflows, I think the most useful tools won’t simply generate content faster.

They’ll challenge assumptions.

They’ll expose weak ideas earlier.

They’ll help creatives make stronger decisions.

And they’ll create more space for the strategic thinking that great branding depends on.

Because good design isn’t generated.

It’s chosen.

---

Project: AB+A Advertising Concept Campaign (2025)

Focus Areas: Brand Design, Creative Strategy, Identity Systems, AI Workflows, Custom GPT Development