What we learnt from building a brand and website from scratch in an hour, exclusively using AI Tools like ChatGPT 5.0, Claude, and Lovable.
AI has moved from buzzword to everyday tool. As prompting becomes an essential skill across disciplines, we wanted to test its potential in design work. At Supercharged, we ran an experiment: could we create a complete brand and website in just 2 hours using AI, starting from nothing?
Tools like ChatGPT, Claude, and Lovable have quickly entered the workflows of designers, developers, and product teams. The critical question in 2025 isn't whether to use AI, but how to integrate it without compromising design quality. We wanted to know: would AI function as a true collaborator or merely as an assistant?
This post covers our complete setup, what failed, what we learned, and the final brand we created. By running this hands-on experiment, we gained concrete answers about AI's role in real brand creation.
ChatGPT 5.0: Our primary collaborator for ideation, branding, and content wireframes.
Lovable: A no-code website generator we used alongside ChatGPT for rapid prototyping.
Claude: Tested for comparison with Lovable, but showed lower performance in structure and efficiency.
We assembled a team of 3-4 UX designers with similar backgrounds. Our ambitious goal: generate a complete homepage for a new brand within a one-hour meeting. The process would flow from concept to brand identity to functional homepage, with ChatGPT guiding ideation and Lovable building the site.
Our hypothesis was skeptical: AI would struggle to produce quality work without strong direction, functioning more as an assistant than a true collaborator.
We chose this tight timeframe deliberately. Designers regularly face pressure to produce quickly for investor decks, pitches, or client meetings. The constraint forced us to prioritize ruthlessly and test whether AI accelerates or hinders early-stage design when time is limited
The prompt that we gave to AI was very vague, but we explained our design studio and what our goal was: to run an experiment to create a viable digital business. So, initially, ChatGPT created 10 ideas based on 2025 trends; however, initially, AI gave us service ideas that were related to past search history. So we had to go back and edit the initial prompt to come up with 10 product ideas, which successful entrepreneurs could build a business with.
After we were relatively satisfied with the ideas that AI generated, we asked it to validate and shortlist these business ideas by researching relevant discussions on Reddit. Our main goal was to use a business idea that would generate more revenue. So we had two choices that our team liked, and we created pros and cons based on the ideas.
We then asked ChatGPT to compare Prompt 3 and Prompt 4, and our main goal was to choose an idea that would generate higher revenues.
We are a design studio running an experiment: using ChatGPT as our end-to-end creative and strategic partner to launch a viable digital business. The process includes ideation, customer research, branding, wireframing, and copywriting — culminating in the creation of a landing page designed to generate hype, validate demand, and build a waitlist. The final site will be built using Lovable.
To begin, generate a shortlist of 10 innovative business ideas that meet the following criteria:
Aligned with trending sectors in 2025 that show strong potential for scalable growth
Currently underserved or have low competition
These need to be phsyical or digital products that any seasoned and successful entrepreneur could have build a business with.
Present the ideas with a brief rationale for each: what trend it aligns with, why it’s scalable, and what makes the competitive landscape favorable.
-----
Do this all from scratch, do not reference any of our chat history for this exercise. Do not personalize these ideas to things I have previously spoken to you about.ChatGPT suggested that Idea 3 for Smart Home Accessories was likely to have a “larger upside potential in terms of total revenue”. There was also a greater user population for this, which would be older adults and adult children. We then started moving forward with Idea 3. The reason we went ahead with this is that the larger user pool has more revenue potential. There were existing product samples which was easier to validate. And there were recognizable real-world end pain points, which were generated later by AI.
Pit ideas 3 and 4 against each other.
What is a business that is more likely to generate more revenue?ChatGPT then convinced us to move on with Idea 3 based on its pros and cons list.
Direct Comparison: “Which is More Likely to Generate More Revenue?”
If we’re talking revenue ceiling and growth potential, not just ease of getting off the ground:
Idea 3 (Ageing-in-place brand)
Higher potential AOV
Strong emotional driver → higher willingness to pay
Multiple channels (DTC, B2B, partnerships, insurers)
Potential for both product and subscription revenue
Can credibly grow into a multi-product, multi-country brand in a massive demographic trend
Idea 4 (Thrift engine)
Beautiful niche, mission-aligned, easier to build as a lean SaaS
Great for stable, “quietly profitable” B2B software
But revenue per customer is limited and customer base is more niche and price-sensitive
Likely tops out as a solid small-to-mid SaaS business unless you aggressively expand into broader retail categories
👉 Winner for revenue potential: Idea 3 – Ageing-in-place smart home accessories.
Because the underlying demographic trend is huge, the spending power is higher, and you can stack revenue across products, bundles, and services.We then asked AI to create user personas which has multiple stakeholders, which were not just older adults, but also their children, caregivers, and professionals. We asked AI to generate the needs and motivations for the users.
Let’s move forward with Idea 3. For this idea, identify all relevant stakeholders involved (e.g. end users, decision-makers, influencers, etc.). Then, create detailed user personas for the primary and secondary user types.
For each persona, include the following:
Name, age, location
Role or background relevant to the product
Goals and motivations
Pain points and current solutions they use (if any)
Behavior patterns or buying habits
Tech savviness or familiarity with similar tools
A short quote that captures their mindset or need
Present 3–5 personas that represent different angles of the customer journey or adoption path.These was just one of the persona's that ChatGPT came up with for us.
🧭 Core Personas
1. Margaret Clarke – Independent Senior User
Age: 78
Location: Bath, UK
Background: Retired teacher; widowed; lives alone in her family home; moderately active, loves gardening and reading.
Goals & Motivations
Maintain independence and stay in her own home as long as possible.
Reduce her children’s worry without feeling “monitored”.
Prefers technology that feels natural, not medical.
Pain Points / Current Solutions
Uses a basic panic button pendant but dislikes wearing it (“makes me feel old”).
Struggles with poorly lit stairs and worries about falls at night.
Has Alexa for reminders, but doesn’t want cameras everywhere.
Behaviour & Buying Habits
Cautious buyer; trusts recommendations from her GP or her children more than ads.
Prefers products that blend into her home decor.
Willing to pay more for peace of mind and simplicity of use.
Tech Savviness
Moderate; uses an iPad daily for calls, emails, and news; not confident with setup or maintenance.
Quote:
“I want to live safely without turning my house into a hospital.”We then clarified to AI that our primary market would be the US and Canada, which is why it then gave us an updated summary list of all core persona's.
We then started to move on to defining the website content, and wanted to get 3 products that would path the principles of the product launch, which were LumaPath, SafeStep, and AuraSense.
Come up with a initial line of 3 products that we can offer for our launch.These were the three products that ChatGpt came up with:
Then, at this point, we needed to define business-level strategy, and we used existing knowledge from SuperCharged content, and told AI to think like a senior product manager and define core strategic foundations for the business based on the potential process using the frameworks that were taken from existing knowledge. Shout to our CEO, Nivit, who provided this knowledge.
Think like a senior product manager. Based on Idea 3, define the core strategic foundations of the business and its potential products using the following frameworks:
For the Business (high-level):
Provide the Who, What, and Why:
Who is the target customer segment?
What problem are we solving, and what is the product or service we're offering?
Why does this matter now — what trends, pain points, or market gaps make this the right time?
For each Product or Feature within the business:
Use the Who, What, Why again, but tailored to product-level decisions:
Who is the specific user or use-case this feature serves?
What does the product/feature do for them?
Why will this feature matter or deliver value in the user's context?
Organize the output clearly:
Start with a concise business-level Who/What/Why
Then list 2–3 potential products or key features, each with their own Who/What/Why analysis
Then, based on all these selected business sides, we asked it to generate brand strategy components with a focus on clarity with the target audience. Then we asked for the 5 unique name options, the taglines, and the mission statement.
At this point, there were a lot of choices that our team collaborated to choose the one that we wanted our brand to pick, which could generate the most revenue.
Based on the selected business idea, generate brand strategy components with a focus on clarity, creativity, and alignment with the target audience. Please provide the following:
Business Name:
5 unique name options that are relevant, memorable, and ideally available as .com domains
Include a short explanation for each name choice (e.g., meaning, tone, brand fit)
Tagline:
5 tagline options that communicate the core value proposition clearly and compellingly
Vary the tone across options (e.g., bold, friendly, aspirational, etc.)
Mission Statement:
1–2 sentences that describe the company's purpose and the problem it aims to solve
Vision Statement:
1–2 sentences that describe the long-term impact or change the company wants to create in the world
Brand Values:
List 4–6 core values that define the company’s principles, personality, and way of operating
Each value should have a 1-sentence explanation to clarify what it means in practiceThe AI pumped out the following options for names:
Our team liked Options 1 and 4, so we went with Option 1. The decision was unanimously with our team liking both option 1 and option 2 however after a brief discussion, tagline 1 felt like a better option for a product-based service.
Then, since we have a title and tagline, we asked AI to help us define a cohesive brand identity for this business. At this point, we also gave AI clear requirements to generate 2-3 options for colour palettes, typography, logo concepts, brand elements, brand vocabulary, and provide keywords.
Then the colour palettes that AI suggested were a warm and “forest” colour, calm technology that reminded us of Apple, and a dark navy midnight theme. Since we wanted to differentiate our design from Apple, we chose to go wth option 3, which was the navy midnight theme.
Then, after giving our options that we liked to AI, it helped us create a locked brand identity with the colours, font, and logos that we liked.
Next, we used AI to generate a detailed written description of the homepage wireframe for our business including Color Palettes, Typography, Logo Concepts, Brand Elements and Image guidelines.
Help us define a clear and cohesive brand identity for our business. Provide the following components with thoughtful reasoning and alignment to our target audience and brand values:
Color Palette
Suggest 2–3 distinct color palettes (primary + secondary colors)
Each palette should include 4–6 colors with hex codes
Briefly explain the mood or personality each palette conveys
Typography Guide
Recommend 1–2 Google Font pairings (for headers and body text)
Specify weights, usage (e.g., H1, H2, paragraph), and why the pairing fits the brand
Prioritize readability, flexibility, and brand alignment
Logo Concept
Describe 2–3 logo directions (wordmark, icon-based, monogram, etc.)
Explain the style, symbolism, and how it scales across formats
Optionally, suggest inspiration sources or design references
Brand Elements
Identify key visual elements or motifs (e.g., shapes, patterns, icons, illustrations) that can support the brand look
Mention how they can be used in web, social, or print
Image Guidelines
Recommend a visual style for imagery (e.g., photography vs. illustration, color grading, subject focus)
Include direction on tone (e.g., aspirational, raw, friendly), use of stock vs. custom visuals, and composition tips
Brand Personality & Tone of Voice
Define 3–5 personality traits that describe how the brand should feel (e.g., bold, friendly, innovative)
Describe how this translates into tone of voice in copy and communication (e.g., conversational, authoritative, playful)
Brand Vocabulary
Provide a list of 10–15 words and phrases that align with the brand’s tone and values
Optionally, include a “words we avoid” list to maintain consistency and avoid misalignment
Then we asked Chatgpt to put all the information into a comprehensive brand and website identity document that consolidated everything that we’ve confirmed so far. Then we uploaded this document Lovable and asked it to create a homepage based on our chosen information. This was the moment of truth!
Our first website design had very strong problems at first glance. The image background for the hero was the same as the image chosen beside the text. So then we asked Lovable to create a website that did not have the same image.
But alas, *ding ding ding*. The thing we had anticipated from the beginning but were desperately hoping would not stop us appeared. The dreaded daily limit. So in the end, this was our stopping point.
.jpg)
Our initial output landed at a 4/10 from the team.
It was not great, but not terrible, and it definitely not even close to the strongest website we’ve ever seen.
But given that it was produced in under an hour, it already underperformed what a non-designer founder could reasonably build in a day using something like Wix. In that sense, a 4/10 is surprisingly high for a rapid AI-driven sprint.
And it raised a real question for us: If one hour can get you to a 4, what could a full workday achieve?
With Lovable as the baseline generator and using ChatGPT as a guiding structure, getting that same concept to a 7 or even an 8 is very realistic.
To test that theory, we repeated the process with a real client, “The Urban Astrologer”, applying that same workflow but wth clearer brand inputs. Within an hour, the output jumped dramatically. This version landed at a 7/10, and is still not polished, but absolutely strong enough to support real business goals like trust-building and service clarity. This experiment was an absolute win for us.
After completing the team experiment, we applied our process to create a full website for The Urban Astrologer, a long-standing client. We were given their Instagram brand and used similar techniques learned from the team experiment.
We already had a logo, brand vibe, and Instagram references. The goal was to create a polished website using Lovable and ChatGPT, with an estimated timeline of 3-4 hours.
We started by feeding The Urban Astrologer's brand information—client references, color palette, and logo—into ChatGPT. It produced the mission and values, brand vocabulary, and system of motifs with less prompting compared to our initial Havenry project.
We also asked AI for a homepage wireframe, which gave better results than before because our inputs were more specific and the brand identity already existed.
We used the same prompt as Havenry's website to create a fully functional website using Lovable, with slight edits to fit the context better. We then created prompts for typography, CTAs, testimonials, and product panels.
After applying this process to The Urban Astrologer with more time, we were pleased with the output.
.jpg)
.jpg)
The Experiment: Required significant collaboration and debate to produce refinement. AI generated irrelevant or service-based ideas. The group debated ideas before prompting again, making human input essential. There were multiple rounds of refining the problem statement.
The Pilot Project: The Urban Astrologer project was more aligned because there were clear brand inputs and faster iterations. The brand tone was consistent and more efficient. I controlled the entire flow through prompting and brand direction. This was built for an existing business with an established identity, not a hypothetical brand. Because the context was stronger, AI generated clearer output on the first try.
For Designers: AI can multiply your ideas, but cannot replace initial judgment and design expertise. Prompt specificity matters more than tool choice. Learning frameworks matter more than the tools themselves.
For Businesses: AI can produce near-complete prototypes in hours, but design quality is low. Experienced designers are needed to complete the work and give it proper polish. Businesses must decide whether AI speeds up their work or, in the long run, damages it.
The tools tested in this project were:
We wanted to test the limits of AI tools to see if they could meet the standards of a modern design team—maintaining brand consistency and understanding hierarchy.
We compared multiple tools like Lovable and Claude. Claude was used briefly to check if different models created different quality, but we quickly learned that Claude was not capable of handling a large project and kept giving errors. Meanwhile, Lovable became a crucial part of the pipeline because it allowed us to turn ChatGPT's output into visual prototypes instantly.
A key insight that emerged was how AI performance depends on the direction running the experiment. In the team experiment, we sent brand and expert knowledge to AI based on our CEO's and designers' expertise to explain the brand process. This showed that AI's quality depends on the designer's leadership and knowledge, and can change drastically based on the user.
By running this experiment in a real meeting, we gained a realistic understanding of how designers can use AI in 2025. This project helped us understand how AI behaves when a designer is working independently but needs faster, more effective prompting to guide it.
Yes, there are tools like ChatGPT 5.0 and Lobable that can create full homepage wireframes and functional website layouts. However, it still requires human direction and refinement for brand consistency.
Lovable is one of the great tools for AI-generated documents to turn into instant prototypes. It works extremely well when it is used alongside ChatGPT 5.0.
An effective strategy is to feed AI structured brand inputs, providing it screenshots of the brand logo and design colors, and examples of what the intended end result is, and avoid vague instruction prompts.
AI still struggles with ambiguous prompts, visual hierarchy, and complex multiple-page flows. Wiht newer models being released every week, this gap will be closed in the future.
No, AI can only accelerate ideation and early prototypes from designers, but cannot replace the initial brand interpretation or creativity, or at least yet. Designers will still be needed to guide the vision and process and refine the work.
However the rate at which AI is growing, it's definitely going to change the way designers do their job. Designers will soon start moving towards a more strategic role in directing AI to produce designs.
Supercharged Studio is a creative technology agency that has created dozens of websites, brands, and apps for emerging innovators, industry leaders, startups, and small businesses.
Work with us. Your ideas will like it here.
We'll supercharged 'em with AI.
Supercharged Studio is a creative technology agency that crafts websites, apps, logos, and brands. We help emerging innovators, industry leaders, hustlers, and dreamers create a competitive edge through design.
Your ideas will like it here.
