Introduction

This article shows how to do user research with AI when you have no users and what 2025 AI adoption data really means.

What's the problem it solves?

Founders often need user insights long before they have real users. They also need clear signals about how companies are actually adopting AI, not hype or fear. The article solves both problems by showing how to use AI to simulate users and by breaking down real adoption data.

Quick Summary

The article explains how early founders can use AI to create synthetic users. These AI personas help teams understand goals, pain points, habits, and buying triggers without running long research cycles. By running role-play interviews with ChatGPT, founders can test ideas, messaging, pricing, and features quickly and cheaply. These personas do not replace real users, but they help teams surface weak spots fast and prepare for real interviews.

The article also looks at data from 1,000 ChatGPT Enterprise customers. Adoption is slower than people think. Only a small part of the Fortune 500 uses it and many big tech companies still avoid it. Startups adopt faster because they have fewer blockers. Companies that avoid ChatGPT Enterprise tend to hire more AI talent so they can build their own tools. ChatGPT Enterprise customers often use OpenAI instead of hiring engineers for similar work.

The last section covers the return of unicorns. October 2025 saw a sharp rise in new unicorns. The difference this time is that investors reward strong AI-native products with real traction, not hype. If AI is the core of the product, the odds improve. If AI is just a feature, it is not enough. The market is back for the best teams.

Key Takeaways

• AI personas help founders understand users before they have real users.
• Role-play interviews with AI surface pain points and blind spots quickly.
• Use synthetic personas for early ideas, then validate everything with real people.
• Only 5 percent of the Fortune 500 uses ChatGPT Enterprise because of privacy, risk, and slow processes.
• Startups adopt ChatGPT Enterprise faster because they can move without long reviews.
• Companies that avoid ChatGPT Enterprise often hire more AI staff to build tools in-house.
• Unicorn creation is rising again, but only for AI-native products with strong traction.
• Investors want real proof, not hype or momentum.

What to do

• Use AI to create personas shaped by goals, tasks, and frustrations, not age or hobbies.
• Run simulated interviews to test ideas before talking to real users.
• Ask AI personas to react to your landing pages, features, and pricing.
• Use AI to summarise real interviews once you have them.
• Treat AI insights as early guesses, not truth. Confirm with real users.
• If building AI products, aim for AI-native value that cannot be added by competitors later.
• Prove traction early so investors see a durable business, not a trend-driven one.

‍