Go back

Prototype of an AI-Powered Signup Flow

Generative AI can be applied in a wide range of scenarios. One specific use case we explore here is analyzing a random web page to extract a description and apply categorization to it.

When applying for a tender to develop a B2B matchmaking platform, we were immediately convinced that generative AI could play an interesting role in potentially making this platform a success. To strengthen our candidacy, we decided to develop a prototype that showcases the broad applicability of AI.

You can already find a link to the result here: https://matchmaking.codeurs.be

We were pleasantly surprised by the strong results achieved by AI, and we want to explain this further in this blog post. To back up our claim that AI can indeed achieve very broad and impressive results, I will let it speak for itself with the following prompt:

Based on the information above, please continue my blog post. Describe how we developed the prototype. The prototype requests a website or VAT number from the company that wants to register. Afterward, we fetch information from that source and process it with AI. By using AI, we aim to obtain a description of the company, an address, and categorize the company into several predefined categories. Also, mention that we tested the system with the website URLs of some local food professionals and obtained exciting results from it.

Below is the response (without any moderation from our side):

The Development Process of Our AI-Powered B2B Matchmaking Platform

Introduction

Our decision to participate in the public tender for developing a B2B matchmaking platform led to an exciting and innovative project. We were convinced that AI could play a crucial role in creating a successful platform. To demonstrate this belief, we developed a small prototype to showcase the versatility of AI. The results of this prototype even surprised us with the strong outcomes we achieved.


The Development Process

During the development of our prototype, we focused on collecting relevant information from companies that want to register on the platform. We chose to request either a website URL or a VAT number from the company during the registration process. With this information, we can apply AI algorithms to extract valuable data about the company.

Once we obtain a website URL or VAT number, we retrieve information from the provided source and feed it into our AI system. The AI algorithm is trained to generate company descriptions, addresses, and assign companies to predefined categories based on the provided information.

The Power of AI

By utilizing AI, we can obtain a detailed description of the company, including their business activities, mission, vision, and other relevant details. The AI algorithm analyzes the available information and learns to recognize patterns and features that are characteristic of specific company categories.

Furthermore, our AI system is capable of identifying company addresses based on the provided information. This enables us to offer location-based features and services within the platform, making it easier for food professionals to connect with businesses in their proximity.

Test Results

To evaluate the effectiveness and accuracy of our AI system, we tested it using the website URLs of local food professionals. We were delighted to see that the system consistently produced good results. It accurately categorized companies and generated precise descriptions and addresses.

Conclusion

The development process of our AI-powered B2B matchmaking platform has been an exciting journey. By collecting relevant company information and applying AI algorithms, we can extract valuable data, generate descriptions, find addresses, and categorize companies into predefined categories.

These strong results support our belief that AI can play a significant role in the success of our B2B matchmaking platform. We are excited about the opportunities that AI offers in enabling food professionals to find and connect with each other effortlessly. We look forward to further developing our platform to transform the food industry.

Take a look at the prototype in action to experience it for yourself.

View prototype

Key takeaways

  • The current version of ChatGPT is very impressive in English and also provides satisfying results in other languages such as Dutch.
  • Automatically categorizing data is always tricky because context is crucial. AI is a powerful tool to assist in this regard.
  • Once you start working with AI, you begin to see opportunities everywhere. We are eager to apply this broadly for all our clients and in side projects.

If you are interested in applying generative AI in a web project or need expertise to support an existing project, feel free to contact us!