Recently, OpenAI launched Deep Research, a new product with advanced search and inference capabilities, which led me to reflect on the impact of internet-connected AI models on businesses. Both OpenAI and Google have developed models that integrate real-time information through internet searches, but interestingly, these capabilities are only available within their respective platforms and not through an API. This raises questions about the usage policies and constraints set by these companies.
However, there are alternatives that enable you to access models with this capability via an API. One example is Perplexity, which has evolved by incorporating various language models and refining its information sources to improve the context of its answers. In this article, I want to share a practical case where we applied this technology in a business environment, demonstrating its real-world usefulness.
Use Case: Customer Database Enrichment
Imagine a company that has been operating in its sector for years and has a customer database. It knows who its best customers are based on their transactions, but it lacks deep insight into what those customers actually do. This scenario is common in sectors where clients can operate across multiple industries and where there has never been a dedicated data team to map out customer profiles.
To tackle this challenge, we developed a system that, starting with the client’s name, searches the internet to identify what each company does and then enriches the database with that information.
The process follows these steps:
Automatic Search: We use Perplexity to retrieve business descriptions based on verified internet sources.
AI Categorization: We apply language models (LLMs) to classify each company into a specific category and subcategory.
Data Visualization: We create an interactive interface that lets users explore the distribution of customers by category and subcategory.
Identifying Similar Customers: We add a feature that, upon selecting a segment, searches the internet for similar companies and displays them along with their website links.
Impact and Applications
This approach enables companies to gain a deeper understanding of their customer base and make informed business decisions. By categorizing each client clearly, they can spot market opportunities and optimize their customer acquisition strategy.
Additionally, the feature for finding similar companies allows sales teams to quickly and efficiently discover new potential clients. Instead of doing manual research, the AI automates this process and delivers results within minutes, improving sales strategy and outcomes.
Conclusion
This is just one example of how AI models with internet-search capabilities can be applied in real business processes. Combining LLMs with up-to-date information makes it possible to enrich data, refine customer segmentation, and open up new business opportunities.
If you have ideas about other ways to apply this technology, I’d love to hear them. You can share them in the comments or send me a direct message. The potential of these models is enormous, and their application in the business world is just beginning to be explored.
Comments