REAL CASE 101

From About $150/Month to $8.6K MRR: How Visualizee Found Growth Again

Piotr Obidowski built Visualizee, an AI architectural visualization product for architects and interior designers. The market need was real and the AI could generate architectural images, yet after nearly two years the product was still producing only about $100–$150 per month, mainly from one-time payments. The core problem was not that the AI failed to work. The first node-based version was too complex for its actual target users. Piotr later rebuilt the core experience around a simpler chat-based interaction, moved more prompt and technical complexity behind the product, shifted the business model toward subscriptions, and invested systematically in high-intent SEO. Piotr later publicly reported that Visualizee reached $8.6K MRR. The case shows that technical capability creates business value only when customers can reach the desired result easily.

PolandLiving / AI Architectural Design & Visualization SoftwareSuccessful TurnaroundThe technology worked, but customers struggled to use it: how Visualizee turned around through product simplification, subscriptions, and SEO

BEFORE

Approx. $100–$150/monthNumber

AFTER

$8.6K MRRNumber

OUTCOME

Successful TurnaroundCase Result

MODEL

AI architectural visualization SaaS for architects and interior designers; shifted from primarily one-time payments toward subscription revenue while retaining some one-time paymentsBusiness Model

Read the Case

## 1. Nearly Two Years In, Revenue Was Still About $150 a Month

An AI product had been under development for nearly two years. The technology could generate the results customers needed, and the target market was real, yet monthly revenue was still only around $100–$150. Where was the problem?

That was the situation Piotr Obidowski and Visualizee once faced. Piotr is a software developer who built Visualizee while maintaining a full-time job. His goal was to use generative AI to help architects and interior designers create architectural visualizations faster.

Piotr later clarified publicly that the early $100–$150 per month was not MRR. It came mainly from one-time payments. In other words, the product had shown that some customers were willing to pay, but it had not yet developed stable, recurring SaaS revenue.

## 2. He Found a Real Problem in the Architecture Industry

Early in the design process, clients often have only sketches, screenshots from SketchUp or Revit models, or design concepts that are not yet fully developed. Architects still need to turn those ideas into visual results that clients can understand, discuss, and revise.

Traditional architectural rendering requires specialized software, professional skills, and time. Generative AI created a clear opportunity: if AI could rapidly turn sketches and design concepts into architectural images, it could reduce the time and cost of early-stage visualization.

So Visualizee's original market direction was not fundamentally wrong. The need existed, and AI was already capable of performing an important part of the task.

## 3. The First Version Wasn't Useless — It Was Too Difficult

Piotr spent roughly five months of his spare time developing the first version of Visualizee. It used a node-based workflow similar to ComfyUI. Technically, the approach was flexible and allowed users to control different parts of the generation process.

But Visualizee was not primarily serving AI engineers. Its target customers were architects and interior designers.

Those customers wanted design results. They did not want to learn nodes, prompting, and AI workflows before they could obtain those results. User feedback increasingly pointed to the same problem: the learning curve was too steep and the product was difficult to understand.

Visualizee therefore faced a classic product contradiction:

The technology could perform the task, but the customer's path to value was too complicated.

## 4. Customer Feedback Revealed the Real Bottleneck

Faced with persistently low revenue, Piotr could have reached several conclusions. Perhaps the AI needed more features. Perhaps the product needed more traffic. Perhaps marketing was insufficient. Or perhaps the market did not exist.

But customers were not saying, "We don't need architectural visualization."

What they were effectively saying was:

This product is too difficult to use.

Those are very different diagnoses. The first suggests a demand problem. The second suggests a problem in how the product delivers value.

Piotr later acknowledged that he had heard some of these signals earlier but did not act on them quickly enough. That distinction matters: recognizing a problem is one capability; responding to evidence quickly is another.

## 5. The Real Turning Point: Make the Product Carry the Complexity

Visualizee's real turning point was not the discovery of a completely new or more powerful AI model. It came when Piotr redesigned the product around how his target customers actually wanted to work.

He eventually abandoned the node-based core experience into which he had already invested substantial development time and rebuilt Visualizee as a more direct, chat-based product.

The new workflow no longer required architects to become prompt-engineering experts first. Users could upload a sketch or model and describe the desired design result in ordinary language. The AI assistant could help interpret the request, ask necessary follow-up questions, and handle more of the prompt construction and generation logic.

The complexity did not disappear. It moved away from the customer and into the product.

Before, customers had to learn how to adapt to AI.

After the redesign, the product increasingly helped AI adapt to the customer.

## 6. The Business Model Had to Change Too

Even after improving the product experience, another question remained: how could one purchase become sustainable recurring revenue?

Visualizee initially relied largely on one-time payments. Even if new customers continued buying, that structure did not naturally create stable MRR.

Piotr later shifted the business more toward subscriptions. Customer retention and continued value delivery could then translate more directly into recurring revenue.

Visualizee's later growth therefore should not be reduced to "the interface became simpler." Both the product experience and the revenue model changed in important ways.

## 7. From Traffic Volume to Traffic Quality

Early on, Visualizee received roughly 30–50 visitors per day, including traffic from AI tool directories and recommendation channels. In his public retrospective, Piotr said conversion was below 1%.

He later invested systematically in SEO by building blog content, researching keywords, and targeting questions architects and designers were actively searching for. After the shift, high-intent traffic increased to roughly 180–200 visitors per day.

Piotr also shared a revealing counterexample: Pinterest generated around 100,000 views but produced almost no meaningful trial conversion.

The lesson was clear. Exposure and purchase intent are not the same thing. The important question is not simply how many people see a product, but whether those people are actively looking for the problem it solves.

## 8. The Result

A Public Milestone of $8.6K MRR

In his public founder retrospective, Piotr reported that Visualizee had reached $8.6K in monthly recurring revenue.

In a later Starter Story interview, Visualizee was at approximately $8.2K MRR while still generating some one-time payments. Including those one-time payments, total monthly revenue at that point was close to $10K.

This case therefore should not be described as "$150 MRR to $10K MRR." The early approximately $100–$150 per month came mainly from one-time payments, while $8.6K was an MRR figure. They are different revenue measures.

A more accurate trajectory is:

Early stage

approximately $100–$150 per month, mainly one-time payments.

Public growth milestone: $8.6K MRR.

Later interview

approximately $8.2K MRR, with total monthly revenue close to $10K after including one-time payments.

The real lesson is not a dramatic revenue multiple. Three changes formed a business loop:

The product became easier to use.

The revenue model shifted toward subscriptions.

Customer acquisition shifted toward higher-intent SEO traffic.

Technology working is only the starting point. Real product value comes from making it easier for customers to get the result they want.

YouTube · English

Image / Ad

GoGoUpGoGoUp House Ad

DECISION

What would you do?

Visualizee had been under development for nearly two years. The need for architectural visualization was real, and the AI could generate useful images, but the product was still producing only about $100–$150 per month and users were clearly saying that the first version was too difficult to understand. If you were Piotr, what should you prioritize next?

AAdd more AI models, nodes, and advanced parameters so Visualizee has even stronger technical capabilitiesBKeep the existing product unchanged and dramatically increase advertising and marketing to generate more trafficCRebuild the core product experience around real feedback from architects and designers, reduce user complexity, and let the system handle more of the prompting and AI technical workDStop the project because nearly two years of low revenue proves that there is no real market for AI architectural visualization

Full Analysis

Visualizee's real turning point was not the discovery of a more powerful AI model. It came when Piotr began redesigning the product around how his target customers actually wanted to work. The first node-based system was technically flexible, but architects and interior designers did not want to learn an AI workflow before they could produce architectural images. After repeated feedback that the product was too complicated, Piotr decided to abandon the node-based core experience into which he had already invested substantial development time. He rebuilt Visualizee as a more direct, chat-based product and introduced an AI Prompt Assistant. Users could upload a sketch or model and describe the desired result in natural language. The system could help interpret the request, ask necessary follow-up questions, and handle more of the prompt construction and generation logic behind the scenes. At the same time, Piotr changed both the business model and customer-acquisition strategy. Visualizee moved from relying mainly on one-time payments toward subscriptions, while Piotr invested systematically in SEO focused on searches with genuine intent from architects and designers. The turnaround ultimately combined three changes: an easier product experience, a subscription-oriented revenue model, and higher-intent customer acquisition.

Turning Point

Before vs After

BEFORE

Product: Visualizee's first version used a node-based workflow similar to ComfyUI. Users needed to understand nodes, generation flows, and some AI operating logic. Customer experience: Architects and interior designers had to learn how to operate an AI tool before they could obtain the architectural visualization they actually wanted. Users clearly reported that the learning curve was too steep. Business model: Revenue came mainly from one-time payments, without a stable recurring subscription base. Customer acquisition: Early traffic was roughly 30–50 visitors per day, including visitors from AI tool directories and recommendation channels. Conversion was below 1%. Core problem: The technology worked, but the customer's path to value was too complicated.

AFTER

Product: Piotr removed the node-based core experience and rebuilt Visualizee as a simpler chat-based product. Customer experience: Users could upload a sketch or model and directly describe the desired design result. The AI assistant handled more of the prompting and generation complexity. Business model: Visualizee shifted from relying mainly on one-time payments toward subscriptions and began building genuine MRR. Customer acquisition: Piotr systematically invested in SEO around problems architects and designers were actively searching for. High-intent traffic later increased to roughly 180–200 visitors per day. Core change: Customers no longer needed to understand how complicated the AI was internally. The product increasingly helped them reach the design result faster.

GoGoUp Five Powers

Discover

0/5

☆☆☆☆☆

Analyze

0/5

☆☆☆☆☆

Execute

0/5

☆☆☆☆☆

Decide

0/5

☆☆☆☆☆

Create

0/5

☆☆☆☆☆

Discover0/5

Piotr identified a real professional need: architects and interior designers needed a faster way to turn sketches, models, and early design concepts into visual results that could be presented, discussed, and revised. The underlying pain point existed before generative AI. AI simply created a new way to solve it. Piotr's important discovery was to apply AI to an existing professional workflow rather than starting with AI technology and forcing it into an artificial use case. The score is 4 rather than 5 because the first version did not understand early enough how customers wanted to access that value. Later feedback that the node-based system had a steep learning curve showed that Piotr understood the need for architectural visualization better than he initially understood the customer's preferred way of obtaining it.

Analyze0/5

Piotr eventually stopped treating low revenue as a single problem such as "the market is too small" or "we need more traffic." He identified several different issues: substantial product friction, a one-time-payment model that did not naturally build recurring revenue, and major differences in purchase intent between traffic sources. Pinterest provided a useful example. A large number of views produced almost no meaningful trial conversion. That helped distinguish visibility from commercial intent. The score is 4 rather than 5 because Piotr had received signals that the product was too complicated earlier, but did not act on them quickly enough. The eventual diagnosis was strong; the speed of response was not perfect.

Execute0/5

Piotr did not stop at the conclusion that the product "should be simpler." He actually removed the node-based core experience into which he had invested substantial development time and rebuilt the workflow customers used to complete the task. The new chat-based experience and AI Prompt Assistant allowed the system to absorb more of the prompting and technical complexity. At the same time, Piotr moved the business model further toward subscriptions and systematically built an SEO acquisition channel. These were not cosmetic interface changes. They were concrete changes to the product, revenue model, and acquisition system. The later recurring-revenue growth shows that the execution produced measurable commercial results.

Decide0/5

Piotr's difficult decision was accepting that the first core product structure, despite months of development work, was not appropriate for the actual target customer. He could have continued adding features, bought more traffic, or abandoned the project after a long period of weak revenue. Instead, he accepted part of the previous development effort as a sunk cost, rebuilt the core experience, and reallocated his limited time and resources. For an independent founder, this was a significant resource-allocation decision. The project's later development indicates that the decision became an important foundation of Visualizee's turnaround.

Create0/5

Visualizee's creativity was primarily at the product and professional-workflow level rather than the foundation-model level. Piotr did not invent a new generative AI foundation model. Instead, he used existing AI image-generation capabilities and redesigned the relationship between professional users and generative AI. The first version required customers to understand more about the AI. After the redesign, the product increasingly helped the AI understand the customer. That is meaningful product and workflow innovation, but because the underlying AI capability was not invented by Visualizee, the Create score is 4 rather than 5.

Key Takeaway
Good technology does not automatically make a good product. Reduce the barrier to use and help customers reach the result faster.

Evidence / Sources

Main sources: Piotr Obidowski's Visualizee founder retrospective on Indie Hackers; a later Starter Story interview with Piotr / Visualizee; and official Visualizee.ai product information. The public sources support the key figures used in this case, including roughly $100–$150 per month in the early period, the public $8.6K MRR milestone, and approximately $8.2K MRR in the later interview, with total monthly revenue close to $10K after one-time payments were included. GoGoUp Five Powers scores and commentary are GoGoUp educational analysis based on these public sources, not official ratings provided by the founder or source publications.