AI-Based System for Monitoring Organizational Reputation From Online News Articles
In June, Tokopedia hosted the annual START Women in Tech conference, a unique hybrid event that brought together female leaders from the industry. With the theme “A Way Forward and Beyond for Women in Technology,” the conference aimed to empower women in tech through inspiring conversations, keynote sessions, and networking opportunities.
The event featured offline keynote talks and online breakout tracks, where Tokopedia’s female tech professionals presented and shared their innovations. Fitra Malida, Senior Media Intelligence Associate for GoTo Group, presented on AI-based Systems for Monitoring Organizational Reputation from Online News Articles.
Let’s learn more about her participation in START Women in Tech!
Can you explain what the Media Intelligence team does and what is your role in the team?
The Media Intelligence (MI) team analyzes big data from both conventional and online media. Our primary tasks involve data mining and data analytics to gain insights for the Corporate Affairs (CA) team. One of our main responsibilities is media monitoring, which includes the collection and enrichment of data from various sources. As for my role, I help the team adopt and implement technology and automation, and I contribute to expanding the products that the MI team offers to improve our internal processes.
Can you share your team’s journey from serving only Tokopedia to expanding your scope to GoTo Group?
The MI team was initially established at Tokopedia to understand public opinion about the company, mitigate crises, and analyze and assess PR efforts and strategies. As the volume of data grew, we also started stumbling upon various critical challenges. This led us to develop an in-house machine learning system since manually analyzing such a large amount of data became impractical.
Now, also serving GoTo Group, the core objective of the team remains the same. However, we had to adapt our perspective and adjust our focus depending on the operating company (op-co) we are working with. When the MI team transitioned to also serving GoTo, we had to make substantial back-end engineering changes to ensure the tool could cater to nearly 10 industries. The MI tool has also developed and changed a lot — we not only added features but also removed some to make it more effective.
What were the main challenges your team faced? How did the AI-based media monitoring tool solve them?
One of our major challenges was relying on manual processes, which were repetitive and prone to errors. It also became difficult to monitor media content in real time, especially with the unpredictable publication hours of online news. Dealing with the enormous volume of data manually was time-consuming and posed a huge risk of us missing potential crises, not managing misinformation, measuring PR performance inaccurately, and ineffectively allocating our teams’ human resources.
To address the above problems, our AI-based media monitoring tool automates our day-to-day operations, allowing us to gather data 24/7 instead of being restricted to specific working hours. The tool also keeps us alert to ongoing conversations by detecting negative articles and notifying us via Slack or email. This automation truly improves our ability to monitor media, quickly address crises, and make more informed decisions.
How was the collaboration throughout the development of the AI tool?
The development of the AI-based media monitoring tool took several years and involved many different teams. We started working directly with the Data Science (DS) team, after which they suggested we involve the Software Engineering and the IT Enterprise teams.
Partnering with the DS team, we developed a sentiment analysis model and Named Entity Recognition (NER) model. Throughout the development period, which required a substantial amount of data, the MI team focused on providing clear definitions and criteria of our metrics — such as what kind of articles is positive, neutral, or negative, and ensuring data quality. Then, the logic and data will be processed by the DS team to develop the model. Afterward, both teams focused on checking the models’ predictions. Simultaneously, the IT Enterprise team, consisting of Software Engineers and a Product Manager, was responsible for building a tool with which to integrate the models. The tool includes various features like automatic data collection and alerting systems based on our business requirements, identifying the problems and solutions, and how the team could support those goals.
As a result, we can slash manual working hours by half, allowing team members to focus more on cognitive and elaborative tasks that cannot be executed by a machine — such as conducting various in-depth research to answer specific needs and ad hoc reports.
As a female in the tech space, what have been your biggest lessons learned? How has having more women in the MI team brought value to your work?
It is quite interesting to point out that the MI team is predominantly women and that the DS team who currently work with us is majority women. I believe that having a balanced number of women and men working together has been valuable to our work as a collective and fostered a stronger bond among us as we are urged to be more understanding and nurturing of one another, which creates a supportive environment for personal and professional growth.
However, as I didn’t come from a technology background, what was equally important for me was having people from diverse educational backgrounds on the team. As I had an International Relations (IR) background, one of the most important lessons I’ve learned is not to let fear stop me from trying new things and discovering my passion in the tech field, as it’s always possible to learn and grow along the way.
Do you want to say any words of encouragement to those not working in tech but wanting to jump into the field?
Absolutely! It’s never too late to start a career in tech. You can shift your career path (strategically) and excel in the field. Don’t be discouraged by any perceived time constraints or previous experiences outside of tech. Take that first step, embrace the learning process, and see where it takes you. With determination and a growth mindset, you can make significant progress and unlock many opportunities.
Fitra Malida, Senior Media Intelligence Associate for GoTo Group, has shared valuable insights into the work of the Media Intelligence team and their journey leveraging AI technology to automate tasks and conduct real-time media monitoring. As a female professional in the tech space, she also highlights the importance of taking that first step into the world of technology, supporting men and women alike, and embracing every opportunity.
Together, let’s continue to create a more inclusive technology industry for everyone!
Read more about START Women in Tech 2023 through this link.
