If you rely on your website or social media to land clients, the way people find you online just took a massive turn
A fresh landmark study reveals that AI doesn’t look for answers the way we thought it did. What does this mean? Your current marketing strategy might be completely invisible to your next big client.
If you run your own business, work as a contractor, or freelance for a living, you already know the struggle of getting found online. However, lately, potential clients are skipping search engines altogether. Instead, they are typing their questions straight into AI tools such as ChatGPT. They ask things such as, “Who is the best freelance designer for a tech startup?” or “How do I choose a trustworthy contractor for an office build-out?”
A study by Semrush examined how ChatGPT actually decides who to mention and recommend.
The findings are a huge wake-up call for anyone selling professional services. ChatGPT acts like two completely different systems depending on how hard it has to think.
Here is what is going on under the hood, what it means for your pipeline, and how to make sure the AI platform points clients to you instead of the competition.
The discovery: “Fast mode” v. “deep thinking mode”
When someone chats with ChatGPT, it processes questions in one of two ways:
Instant mode (Fast mode): This is the default setting. It gives quick, straightforward answers to simple prompts (like “What does an independent project manager do?”).
Thinking mode (Deep reasoning mode): This kicks in automatically when a user asks a complex, high-stakes question (like “Compare three top freelance accountants for a small business and break down their pricing models”).
Most of us assumed ChatGPT pulled from the exact same bucket of information for both. It turns out it doesn’t.
The study found that when ChatGPT switches from Fast Mode to Deep Thinking Mode, nearly 75% of the websites and sources it cites completely change.
In short, where ChatGPT looks for quick answers is almost entirely different from where it looks when a client is hovering over the hire button.
Why this matters for our day-to-day business
To see why this changes the game for us as freelancers, we need to look at where the AI gets its information depending on the mode.
| Aspect | ⚡ Instant Mode (Simple, quick questions) | 🧠 Thinking Mode (Complex, hiring decisions) |
| 🌐 Source Material | Pulls from forums (e.g., Reddit) | Pulls from official websites |
| 📖 Content Type | Uses quick blog posts | Uses case studies & guides |
| 🔍 Depth & Accuracy | Relies on casual reviews | Checks precise documentation |
1. Casual forum chatter gets dropped when money is on the line
A lot of us spend hours active on Reddit, Quora, or social media trying to build a reputation.
In ChatGPT’s fast default mode, those community threads make up a huge chunk of its answers. But as soon as a serious lead asks a detailed evaluation question, ChatGPT shifts into Deep Thinking Mode. When that happens, its reliance on forum posts drops by half.
If your entire online presence relies on casual social chatter, the AI is likely going to pass you over when a high-paying client asks for recommendations.
2. The AI does “deep thinking mode” research on your setup
When ChatGPT enters Deep Thinking Mode, it doesn’t just run a single quick query. It actually performs almost five times as many hidden sub-searches behind the scenes.
If a client asks, “Who should I hire to build my SaaS app?”, the AI silently breaks that down into smaller searches:
- “What is this freelancer’s actual track record?”
- “Do they have concrete examples of past work?”
- “What specific tech stacks do they master?”
If your portfolio site only says something generic like “I build great websites,” the AI will skip right past you to a competitor who explicitly details their exact skills, past results, and workflow.
3. Being helpful early on locks in the final lead
The study also found that when ChatGPT cites a business early in a client’s research phase, it tends to keep recommending that same business all the way through to the final hiring choice.
If you publish clear, practical content on your site that solves a prospect’s initial problem, the AI is much more likely to keep steering that client back to you when they are ready to sign a proposal.
3 simple ways to make your portfolio AI-friendly
You don’t need to be a tech wizard to adapt to this. You just need to make your work easy for an AI to read and verify.
Step 1: Ditch vague pitch lines for explicit details
- Instead of: “I help brands grow through modern marketing.”
- Use: “I build email marketing campaigns for local fitness studios, focusing on customer retention strategies.”
Name your specific tools, target niche and exact deliverables. The clearer you are, the easier it is for the AI to match you to a client’s specific sub-searches.
Step 2: Put real case studies on your own site
Don’t rely solely on third-party job platforms or social feeds. Put detailed breakdowns of your best projects right on your domain. Include:
- What problem was the client facing?
- The exact steps and tools you used to fix it
- The measurable result (e.g., boosted client repeat bookings by 20% in 90 days)
This gives the AI the concrete evidence it looks for when in Deep Thinking Mode.
Step 3: Write how-to guides for common client headaches
Think about the very first question a client asks before they even realise they need to hire an expert. Write a straightforward guide on your site answering that exact issue.
By helping the client right at the start of their search, you teach the AI that your site is a go-to source on the topic.
Where do we go from here?
AI search is raising the bar for what counts as reliable information. Move away from broad summaries and put clear, detailed and proven examples on your own site and social media accounts. This should pay off because when your ideal client asks an AI who to hire, your name is the one that shows up towards the top of the list.
