Obidur Rahman is a Bangladeshi data scientist who has innovated solutions catering to both local and international clients by developing AI-integrated marketing tools. In a candid interview, the current Group Business Director of X Solutions Limited & Founding CEO of iGroott, the tech hub of the marketing automation tools, shared with us his journey as a data scientist and details regarding his two biggest AI-integrated marketing tools: Dotkom and Zerocium.
Q: How did your journey into data analysis begin?
After getting an opportunity to work at Maverick Studios, which was doing a data-driven innovation drive for Robi, I left my established job at a Grameenphone call centre and joined Robi as a data analyst. After the official Facebook page for Robi was created, I created a program called Smashboard to track social interaction with customers. It calculated metrics such as average interaction time, common points of interest and specific issues Robi could be struggling with.
With Smashboard, I had created an automatic business model. But just gathering data from Robi wasn't enough, as I needed to match the competitors as well. Thus, I formed the Robi Listening Centre, the first formal data analysis platform in Bangladesh. From there, I made a social listening tool called Social Troops, which kept track of other local telecom domain customers, monthly users, negative experiences shared with Robi and any other data that was available from the Robi Listening Centre. Social Troops used all this data as outbound and enabled me to create what I would say is our signature product: Dotkom.
Q. What is Dotkom? How does it work?
Dotkom, an AI-centric social listening tool, is our flagship product. Launched in October 2021, this tool is able to scale the digital lifestyle of a customer by profiling their interactions with a brand's social media page. Unlike other language-processing AI, Dotkom is programmed to specifically pinpoint sentiments in the Bangla language, which can range from positive, negative or neutral comments left by customers in a brand's Facebook post. Its language-processing capabilities enable it to find relevant and core problems being discussed by the customers, thus setting priority management for managing public complaints better.
Q. How does Dotkom perform market-based predictions?
Using data derived from customer comments, the AI creates profiles for each user and determines what socio-economic class they might belong to based on the regular pricing options of their availed services. Since there might be a noticeable difference in preference between, for example, 1 GB package buyers as opposed to regular 5 GB package buyers, the AI charts what needs are being addressed by the customers and how to provide better product management to specific classes of customers. Dotkom is also able to delegate and design new campaigns while measuring customer expectations. Using sentiment-driven data, the AI can offer more in-depth suggestions on how an upcoming offer can be advertised better.
Q. Regarding your other product, Zerocium, what is it and how does it function?
Launched in January 2021, Zerocium is an omnichannel catered towards making customer care service more efficient. Whenever a query pops up in Zerocium's shared network, the notification is seen by all the customer care agents of a company. A human agent from one branch can then cater to a customer who might be calling an understaffed branch. As a global joint venture, Zerocium was developed by local and global resources to serve the needs of different brands around the world.
Q. How efficient is Zerocium?
Zerocium has reduced the average customer response rate to 1.5 minutes, where the global benchmark is 18 minutes. It always maintains zero pending customer requests and because of the AI-integrated query system, nearly 100% of the customers are served within the first 5 minutes of the call. Keeping track of the customer replies, the AI sends a detailed analysis to the respective company regarding resource allocation, customer engagement index and employee work rate.
