Automating SECPHO's strategic communication with Artificial Intelligence
Isabel González is an alumni of the Master in Artificial Intelligence at UDIT, a programme she studied after completing her studies in Computer Engineering at the University of León and an Erasmus stay at the University of Pardubice (Czech Republic). It was during this academic stage when she began to develop a marked interest in data analysis and the application of artificial intelligence to process improvement.
His Master's thesis ("Automation of communicative content generation through web scraping and natural language processing"), developed in collaboration with Secpho, focuses precisely on this applied approach .
Faced with the exponential growth of information produced by the members of this scientific cluster, Isabel has designed a software solution that automates the collection, structuring and generation of personalised content oriented towards strategic communication. Using web scrapingtechniques, natural language processing and the ChatGPT API, her system generates texts adapted to the association's philosophy, with the aim of attracting new members and reinforcing its positioning. This project optimises time, reduces errors in manual processes and, most importantly, demonstrates how AI can become a real ally in decision-making.
Looking to the future, Isabel aspires to continue developing her career in projects where artificial intelligence has a direct impact on improving business processes, especially in those areas where technology acts as a bridge between innovation and real business needs .
How did you come up with the idea of automating the generation of content aimed at attracting partners and customers ?
Isabel González: The idea of automating the generation of content arose from a specific need detected at Secpho: to efficiently manage the large amount of information generated by its partners.
Until then, this task had been carried out manually, which meant a high consumption of time and resources, as well as a greater risk of making mistakes. The marketing area, with which we worked closely, identified the need to strengthen our presence on social networks, continuously disseminating the news and achievements of our partners, while improving internal communication.
From this context, the opportunity arose to automate part of the process, taking advantage of the data already available, reducing the operational burden and allowing the team to focus on tasks of greater strategic value.
What role have web scraping techniques and the use of ChatGPT played in the system you have designed? What has been the most complex to integrate ?
Isabel González: They have played an important role. The process starts with the automatic collection of relevant information and news from the official websites of the partners to store them in CSV files, as they are easy to manage when you have a lot of information .
From there, a database was developed by combining this external information with Secpho's internal data and then connecting it to the ChatGPT API to generate content for news dissemination and partner or customer acquisition.
The most complex part was to understand how to structure the requests so that the model would generate coherent texts, aligned with Secpho's communication style and objectives. Also, ensuring that the data was clean and organised so that the system would work without errors.
What type of content is finally generated?
Isabel González: The system generates different types of content from the information stored in the database. For example, it produces general summaries of news published by partners, it writes texts oriented to social networks such as LinkedIn posts, and it is also able to perform deeper analyses, such as detecting the main ideas of each news item or identifying common patterns in different content.
This allows for agile information sharing and a more structured view of all publications, which helps at a strategic and communication level.
What are the advantages of this system compared to a traditional, more manual communication strategy?
Isabel González: The main one is that it saves a great deal of time and resources by eliminating many repetitive tasks such as gathering information or writing content.
It also reduces the margin of error and allows for more constant updating, which allows for more frequent publishing and a more active presence in networks. Another important advantage is that the system allows for more efficient use of the available data, generating more relevant and contextually adapted content.In addition, it facilitates the analysis of trends and the topics most addressed by the partners, which is difficult to detect manually.
What technical and strategic learnings do you take away from working with real data and an entity like Secpho?
Isabel González: Working with real data and Secpho has been an experience that I recommend. On the technical side, I have learned how to do web scraping and I have strengthened my knowledge of data cleaning and data processing, as well as database construction. In addition, using the ChatGPT API to generate automated content has been a completely new experience from which I have been able to learn a lot.
On a strategic level, I have gained a better understanding of how an information-intensive organisation works and the importance of optimising processes to save time and reduce errors .
If you could apply this technology to another area, what would it be and why?
Isabel González: I think that applying artificial intelligence in the education sector can help a lot to personalise learning, adapting content to the pace and style of each student, and automating administrative tasks that take up a lot of teachers' time. It can also facilitate access to educational resources and improve decision-making in schools based on the analysis of real data.
