This article details the entrepreneurial journey of an AI startup, highlighting pivots, the importance of customer understanding, and the challenges of finding product-market fit, especially in the B2B space.
Early Struggles and the "Mom Test"
Initially focused on a ToC product, the startup struggled to gain traction despite positive feedback. An early learning experience involved the book The Mom Test, recommended by an advisor. This book emphasizes asking the right questions to get honest feedback from potential customers, avoiding the pitfall of simply pitching a product's features and receiving superficial praise. Asking questions about the problems a user faces instead of pitching your solution.
Pivoting Away from a Failing ToC Product
The startup realized that positive feedback on hypothetical scenarios didn't translate to actual product usage. This highlighted a fundamental problem: positive feedback isn't always indicative of product-market fit.
Embracing B2B and Finding a Niche
Recognizing the challenges of the ToC market, the team decided to pivot to B2B, specifically targeting a niche vertical. The reasoning was that ToC markets are crowded and competitive. B2B, particularly in specialized areas, offered the potential to gain traction in a smaller market and build from there.
The Challenge of Domain Knowledge
Moving to a vertical market required a significant shift. The team, with a strong technical background, needed to acquire domain expertise. They believed that focusing on necessary domain knowledge and building relationships with customers was key. By immersing themselves in the client's environment, the team can learn their processes and workflow.
How to Find the Right Pain Point
Discovering a specific problem to solve involved a combination of networking and serendipity. Reconnecting with a former colleague who worked in the fashion industry, the team learned about the pain points in photo operations, specifically the inefficiencies and costs associated with product image creation and virtual try-on. They learned that the best way to solve the problem is to be able to fully understand the problem by doing something such as spending a week in their client's studio.
Building a Solution and the Importance of Data
The initial focus was on solving the product imagery challenges in the fashion industry. Using online apps to try on new clothes has been difficult. The team learned their client spends millions of dollars on taking photos for these clothes every year.
Overcoming Data Limitations
Existing open-source models proved inadequate due to insufficient and biased training data. Open source datasets did not provide the data quality or specific angles for the images. This meant the team had to train its own models, a challenging and uncertain process.
Model Training and Uncertainty
Training a model required significant time, resources, and expertise. The model needed to be able to work with a wide range of styles and needs. They had to beat the state of the art in the area.
Customer Collaboration and Data Support
The startup got support and data from their partner. Supplementing with data from outside sources.
Customer Obsession and Iterative Development
The startup's approach shifted from pursuing cutting-edge technologies like video to focusing on solving specific customer problems. This involved delivering the customer obsession. Understanding the needs of the customer is the key to developing the product.
A Focus on Solutions
The team now focuses on providing solutions within a vertical market. Instead of creating new stories, they’re improving on existing ones, leveraging technology to fill in gaps. Customer feedback drives the product roadmap, focusing on solving pain points and building tightly coupled features.
Fine-tuning Models and Vertical Expertise
Vertical applications are where AI should be implemented. They feel it is not worth building another GPT.
Go-to-Market Challenges and Product-Market Fit
Many technical teams do not understand the difficulties of go-to-market strategies. They found that their path to success was through solving a point in an enterprise environment.
The Search for the Right Breakthrough
Finding the point to create a breakthrough is a challenge. Creating value for a space with little demand leads to low ROI.
Iterating Towards Enterprise Success
The team still faced problems scaling when taking on enterprise clients. Enterprise requires the team to reiterate.
Human-AI Collaboration
At first the team was only able to process a dozen Stock Keeping Units (SKUs). AI faces problems such as inconsistency in details. This results in low results for quality. The team combines human effort with AI solutions to ensure everything is perfect and consistent.
A System for Scalability
To improve efficiency, the team developed systems that combined AI and human workflows. Their goal was a scalable system. Enterprise clients also measure the performance and metrics.
Design Partners
Having a design partner is an advantage. In late 2024 the team went to their design partner and identified some problems.
Enterprise System Sales
The team is selling their enterprise system and has had good progress.
{# 
{# 
{# 
{# 
{#