The Impact of Generative AI in Business

In recent months, generative AI has emerged as one of the trendiest technologies, thanks to systems like ChatGPT and the development of new open-source frameworks like Langchain. These advancements enable the creation of more complex systems using large language models (LLMs) as the core. These systems have the impressive capability to generate text, images, 3D objects, and even videos.

How can this technology impact my business?

There are numerous ways in which this technology can significantly and positively impact your company. However, it’s important to note that the true potential lies in integrating these models into larger systems, maximizing their power. With that in mind, let’s explore some examples of how this technology can benefit your company:

Data Management

Most medium to large companies struggle with managing vast amounts of documents and data. Implementing a Generative AI system can revolutionize your document management process, allowing for quick retrieval of relevant information. Say goodbye to endless PowerPoint presentations, PDFs, Word documents, and Excel sheets. These systems can extract valuable insights within seconds, freeing up your employees to focus on more valuable tasks for the company.

The following is an example of a system developed by our team that can extract insights and relevant information from structured and not structured data:

The above diagram illustrates a system designed to process user queries and retrieve relevant information from both structured sources such as Excel and CSV files, as well as unstructured sources like PowerPoint presentations, PDF documents, and Word files. This information is utilized to generate a comprehensive response to the user’s query. The response is generated by a Large Language Model (LLM) and aims to provide the most accurate and helpful answer by considering the extracted information from various sources as contextual input.

Content Creation

Generative AI technology can assist in content creation tasks within your company. It can work alongside your employees in various areas, such as copywriting and script creation for marketing campaigns, enhancing productivity and creativity.
The following is an example of the high level architecture of a solution that our team developed for Repsol’s HackIA 2023 Event where the goal was to create an innovative application using Generative AI. This solution was focused on insights retrieval from Repsol’s data and marketing campaign creation:

The diagram above depicts an end-to-end system consisting of two components. The first component involves answering user questions by utilizing the company’s data sources as inputs and leveraging a Large Language Model (LLM). The second component revolves around the creation and refinement of a marketing campaign based on the insights extracted from the previous step.

Customer Review Understanding

Companies often receive an overwhelming number of customer reviews about their products or services. It becomes nearly impossible to keep track of all of them. Generative AI systems can extract valuable insights from these reviews, allowing you to gain a deeper understanding of customer sentiment and preferences. Furthermore, the system can personalize responses to each customer, enhancing customer satisfaction and informing iterative improvements to your offerings.

In the following picture a simple architecture for information retrieval from documents can be seen:

Although the diagram above shows a generic architecture for the extraction of relevant information, this architecture could be easily used to extract insights from amazon customer reviews of different products. The flow would be as follows:

1. All reviews would be ingested individually into a vector database.
2. Given a user question, a similarity search of the question in the database would be performed and a fixed number of related reviews would be returned.
3. Once these related reviews are obtained, they are used by the Large Language Model (LLM) to provide an answer that best responds to the question given the reviews as context.

This system could easily save companies thousands of hours in extracting insights from their different products based on customer reviews and could help them improve both the quality and the user experience of their products.

Product Design

Imagine a small furniture company that lacks the resources to maintain a large team of designers. With the right system built around a large image generation model, even a small company could validate hundreds of designs in a single day. Additionally, it could generate 3D models of each design, revolutionizing the product design process and accelerating innovation.

The following would be an example of a system that creates different product design images given a text input:

Virtual Assistants and Chatbots

While virtual assistants and chatbots have been around for some time, their user experience has often fallen short of expectations. However, the integration of new large language models has paved the way for more personalized and intelligent experiences. These advanced systems can greatly enhance customer service and satisfaction.

Our team has created a virtual assistant/chatbot internally that helps employees search intelligently in natural language about the company’s documentation. This system is called Keeply and people is currently using it in their daily work:

Conclusion

Generative AI technologies have the potential to completely transform businesses, regardless of size. While these technologies are still in the early stages, and some applications can be costly, costs are rapidly decreasing, and the systems are continuously improving. As these technologies advance, companies can anticipate significant benefits, increased efficiency, and enhanced customer experiences.

Image: Unsplash | Robynne Hu