Suggestions
Hiten Shah
Co-Founder and CEO at Nira
Hiten Shah is a serial software entrepreneur and co-founder of Nira, a company that aims to keep sensitive documentation safe from unauthorized access by integrating smart identity management tools. Shah began his career in the early 2000s and has since launched multiple successful software companies, among them Crazy Egg and KISSmetrics. Prior to launching Nira, Shah co-founded Quick Sprout. Shah's drive and expertise in software development and entrepreneurship have propelled him to be at the forefront of fast-growing software companies. He completed his undergraduate degree at the University of California, Berkeley, and attended high school at Brea Olinda and La Mirada.
Highlights
When you work with AI, the output is never random. It’s a mirror held up to how clearly you think.
Every response is a reflection of your own edges and blind spots. The model doesn’t create ideas. It digs through your words to find the one you meant. What you see on screen is your own reasoning, rendered in pixels.
The first time I realized this, I was testing a design prompt. The layout it produced looked competent but lifeless. I rewrote the prompt a few times, added adjectives, and even name-dropped styles I liked. Each result came back different, but none of them felt right. They were all accurate reflections of my words and incomplete reflections of my intent. That’s when it hit me. The model wasn’t missing imagination. I was missing articulation.
Working with AI shows how much of our thinking hides between the lines. In conversation, other people fill in those gaps for us. They infer tone, context, intention. Machines don’t. They give you exactly what you say, stripped of every human assumption. It’s unnerving at first, but it’s also clarifying. The gaps in your output are really gaps in your direction.
Once you notice the pattern, prompting feels less like control and more like composition. You don’t tell it what to do. You trace the perimeter where meaning can form. When your intent is vague, AI compensates with templates. When your intent is sharp, it composes. The difference feels like creativity, but it’s just comprehension.
I started thinking of prompts as compressed creative briefs. Each one defines the edges of a world the machine can build inside. The more complete the brief, the more coherent the world. That changes what it means to design.
The real challenge is in learning how to think clearly enough to describe what you mean.
AI takes over the mechanical parts of design such as the grids, the layouts, and the repetition, so your judgment has more room to work.
What’s left is the part that’s hardest to automate: Taste.
We ran 180 AI web design prompts to see what actually works. Everyone says these tools can design full websites in minutes. I wanted to see if that was true.
Across Base 44, Bolt, Figma Make, and Lovable, we tested every variation we could think of. We changed how much detail we gave, how structured the prompt was, and how much branding we included. What we learned was simple. AI can generate clean layouts, but it can’t design in the way a person can. It needs direction. It needs a system to follow.
The best results came from prompts that had order and structure. When we asked for layout first, then styling, then interactions, the designs started to make sense. When we asked for everything at once, they fell apart.
AI-generated prompts often beat the ones we wrote ourselves. They were more complete. They included the small details that people forget to mention, like spacing rules, hierarchy, and interaction logic. We started letting AI write the first draft of each prompt, then used judgment to refine it. The output quality jumped immediately.
Brand identity made the biggest difference. When we gave AI our colors, fonts, tone, and a few reference screenshots, it produced results that felt specific. Without that context, every tool drifted into the same generic templates.
What mattered most wasn’t the platform but the process. The same rule held everywhere. Vague prompts produced average results. Structured prompts produced usable prototypes.
AI design only gets good when your thinking does. The clearer the intent, the stronger the result.
Tools change. The skill that lasts is articulation.
AI takes over the mechanical parts of design such as the grids, the layouts and the repetition, so your judgment has more room to work.
If you treat it like a shortcut, you’ll get templates. If you treat it like a collaborator, it will make you faster.
The full results, along with examples and screenshots, are here: https://t.co/lki9G3tOwX…