Gowthami Kuttapalayam Mothilal: If your quality of data is poor, then whatever technology you bring in is not going to help your business.

Mongolian Economy Magazine spoke with Gowthami Kuttapalayam Mothilal, a Business Intelligence Developer and Analyst at a startup company in Stockholm. She has an extensive experience working in Data and AI Consulting. Furthermore, she aspires to help women who want to transition to the field of Data and AI, and she has authored several children’s books. In this interview, she sheds light on what is AI, how can poor data lead to biased AI systems, and how can we prepare for the future of AI.

-Could you briefly introduce yourself to our readers?

-I am a data and AI consultant working at a startup company in Stockholm, Sweden. In my role, I serve as a Business Intelligence Developer/Analyst and actively contribute to AI-related projects. Additionally, I am passionate about mentoring women who are seeking to enter the data and AI field, especially those transitioning from unrelated fields. Alongside my professional endeavors, I indulge in writing children’s books. I have authored and published a few books in the past.

-What led you to a career in data and AI?

-Honestly, it happened gradually. Throughout my school years, I was one of the young student scientists who consistently participated in STEM-related projects. From an early age, my passion for numbers was evident. When you delve into AI and data analytics, the fundamentals are closely tied to mathematical concepts and statistics. Therefore, my natural inclination towards data and AI was a logical progression for someone passionate about math.

-Many readers might not have a profound understanding of what is AI. Could you explain to us what is AI in simple terms?

-I suppose many readers might be familiar with AI, thanks to ChatGPT. AI stands for artificial intelligence. In simple terms, we aim to make machines or computer systems simulate human intelligence. For instance, when you wake up and use your smartphone, the first thing that operates is the facial recognition system – an example of AI. If you want to travel, you access Google Maps, check the distance, and choose your means of transportation, be it a train, bus, or car – all facilitated by AI. Using Alexa is another example; it recognizes your voice and performs the tasks you ask – also AI. Whether you are aware of it or not, you use AI in your everyday life. AI has already become an integral part of our lives.

-People fear that AI will take over their jobs in the future. What skills and professions we should focus on to remain relevant in the changing landscape?
-This question is becoming increasingly relevant. According to the Future of Jobs Report 2020 published by the World Economic Forum, it’s estimated that by 2025, 85 million jobs will be replaced by AI. However, at the same time, 97 million new roles will be created by AI. Therefore, a shift in our skill set is imperative.

The fundamental thing we need to understand is which kinds of jobs will be impacted first then, we can focus on developing new skills. If you believe your job can be automated, lacks creativity, or is mundane, then those jobs can be easily eliminated by AI. For instance, the role of a driver is already being replaced by driverless autonomous vehicles. Uber has entered into a partnership with Waymo to introduce autonomous cars on its app. In the future, driverless vehicles might become commonplace, so drivers should be prepared to reskill themselves. Another example is customer service agents being replaced by chatbots or virtual assistants.

Regarding the skill set you should focus on, it depends on your interests and curiosity. This is the right time to explore new fields. We can observe on LinkedIn that new roles, including Prompt Engineer, AI Engineer, Large Language Model Engineer, and Software Engineer LLM Ops, are being created. The demand for these skills is on the rise.

I recommend actively using LinkedIn to explore job opportunities in the fields of data and AI, as the demand for these skills continues to grow. Understand the job descriptions and explore them. There are many books, courses, and YouTube videos available. Additionally, you can approach mentors working in the field you are interested in. Numerous networking opportunities exist as well. For instance, I am currently a member of Women in AI and Stockholm AI where you get opportunities to connect with people in your area of interest. All you need is curiosity, the rest you can figure out easily.

-Regardless of our education and background, what can we do to become more data literate than we are now?

-This is an important question. In 1820, the global literacy rate was 12 percent. Now, it is 99 percent in developed nations. When it comes to data literacy, there are no clear statistics. However, I would say that the data literacy rate is still relatively low. I think that data literacy is for everyone in this digital world, not limited to corporations. Data literacy means numerical literacy which is the ability to understand statistics, charts, and graphs. These are the fundamentals that you should know and you have various sources to use for your learning. You can read books such as ‘Storytelling with Data: A Data Visualization Guide for Business Professionals’ written by Cole Nussbaum Knaflic. This book is designed for everyone interested in becoming more data literate regardless of their knowledge of mathematics or data. Furthermore, you can take courses to develop your skills. 

-You have a rich hands-on experience in the entire data journey, from collecting data to creating visualizations. What are the most common mistakes that people make in transforming data findings into persuasive presentations?

-Generally, people tend to include all data points in their reports or presentations, assuming that more data is better. However, this is not the case. It is important to know which data points to highlight or include in presentations. Imagine that I am the CEO of the company, I might not be interested in a leave report of employees compared to other key business metrics such as employee attrition. When people include irrelevant data, executives are likely to overlook the data points that need more attention. Therefore, minimalism is the key. You need to consider their needs, the time they have, and which data points to highlight. By doing so, you can make your report more relevant and valuable.

The second mistake people make is failing to include their recommendations. People who work with data often spend a lot of time on research, identifying patterns, and gaining insights but they fail to state their recommendations. The reason could be fear that the recommendation might be wrong. This is a common concern I hear from professionals in the early stages of their careers. You should remember that no one will take your recommendation as it is, it will merely serve as a discussion point. Since you have spent quality time, you should add recommendations by saying, “Based on my findings, we should consider stopping this line of product because sales are rapidly declining.” Even if they do not immediately scrap the product, your recommendation will spark a discussion about it.

In conclusion, being minimalistic and including your recommendations will help you transform data findings into persuasive presentations.

-I saw that you shared a post saying “IBM found that poor data quality strips USD3.1 trillion from the US economy alone annually.” How can we improve our data quality? What challenges can companies face in the process of improving the quality of their data?

-It is a really big problem. The main issue that I am noticing is a lack of budgeting. First, we need to allocate a budget to employ people who are specialized in data quality, or data governance. With a proper data owner and structure in place, everything will run smoothly and better.

You might have heard the common term, ‘Garbage In and Garbage Out’. It means the quality of the output is determined by the quality of the input. Tech leaders need to keep in mind that if your quality of data is poor, then whatever technology you bring in is not going to help your business.

However, in many board meetings, they agree on adopting new technologies and allocating budget to recruit AI and ML engineers but have less interest in recruiting data quality and data governance specialists. We have to understand that data scientists cannot create magic unless they have very good data.

-Can poor data quality lead to bias in the algorithms and AI systems?

-Absolutely. You might have heard about the story of a very renowned and large global organization that had previously released a resume-screening tool. Usually, recruiters receive hundreds of job applications, but it is challenging to review them all. Therefore, the organization created a resume screening tool using past historical data from their organization. Unfortunately, the tool was biased and started preferring male candidates when compared to female candidates. It is not because women were not qualified, but the training data was based on male candidates’ resumes. That meant that the AI system automatically thought that only male candidates were preferred. Later on, the organization scrapped its resume screening tool. These kinds of biased tools still exist in today’s world. For that reason, good data quality can play a very important role in AI systems. Data quality is a broad concept, but you can say one of the critical elements of good data is being inclusive and one that represents a diverse group of people.

-What should we keep in mind when developing inclusive products? 

-When it comes to developing products and services, there are three key factors. First, have an inclusive team.

Only a team of diverse people can identify each other’s blind spots. It is very essential when you are developing products for a mass audience. Second, avoid participant-washing (having diverse teams for the sake of diversity).

While many claim to have a diverse team, the crucial question is: do they actively listen to them, comprehend what they are conveying, and genuinely value their input? Finally, having proper regulation in place by authorities and driving compliance.

-Some say that with the help of AI people will work a few hours a day. What is your opinion on that? 

-It might become a reality 20 years later. Until then, we have to contemplate how we should change our economy because most of us depend on regular income. If we work less, then we will be paid less. There needs to be some kind of change in that regard. For example, Finland ran Europe’s first national universal basic income experiment. It could become a normal thing in the future. There is a counter-argument that if you pay everyone, then people might not be motivated to work. I think we can slightly change that system. For instance, people who contribute more can get paid more and so on. Eventually, there should be some further thought and modifications to the concept of universal basic income. We must use this time in the best way to facilitate the future.

-You authored a kindergarten and first-grade decodable story book TEAMWORK WINS! How did you get interested in writing children’s books?

-When I took a planned career break to nurture my son during his formative years, I wanted to help him become a voracious reader. I was looking for books that would enable kids to read, but I could not find any decodable books. I discussed it with other parents, and they seemed to have experienced the same issue. The fundamental problem I figured out was that the English language has 26 letters but 44 sounds, and children do not know how to differentiate between long vowels and short vowels. Therefore, I tried to address the problem. I took a course from Queensland University, did a lot of research, and finally created a book with graphical representations.

‘TEAMWORK WINS!’ is a book meant to help children read faster and understand the logic behind how words are formed. After I published the book, I spoke about this with my son’s teacher, and she got extremely interested in buying the book for all the children in the class. Later on, I published it on Amazon, and now it’s available in more than 8 countries. I received very good reviews from parents saying that it was profoundly helpful for their children who were starting to read.

Since I started working again in data and AI, I have not been able to focus on writing books. However, whenever I have time, I try to do some sketching for my books. The process from drawings to publishing the book was a wonderful journey to learn and explore. My son is good at chess, and he is writing a book to help other kids understand how each chess piece moves in the form of a story. Therefore, I am planning to help him create this book and release it soon. I hope other children will enjoy reading it.

-What does the day in the life of a Business Intelligence Developer/Analyst look like?

-Currently, I am working in a startup company which means every day looks different. Generally, I spend 10 to 15 percent of the time understanding what is happening in the field of Data Analytics and AI, including the new tools that are being released. It’s essential to have a proper understanding of why we are choosing a specific tool over others. This means I need to read a lot and try to understand different services and tools.

The most important aspect of my job is understanding the stakeholders’ needs. Then, my actual process begins where I do data collection and ETL (extract, transform, and load). After cleaning, transforming, and loading the data, I create dashboards and visualizations using visualization tools like Power BI. There are always opportunities to learn new tools or new processes, making it an interesting field. You never feel bored. I highly recommend entering the field of AI and data because you will never do the same things. There will be new challenges and opportunities every day. It is quite an interesting job.

-You are a woman who aspires to help women break into the data and AI industry. Are you mentoring women who are interested in the field? Do you have any advice?

-Yes, I am mentoring women who are interested in the field of data and AI. In general, the type of mentorship differs depending on the candidate’s experience and exposure. Among the many women I have mentored, some of them are trying to help others make the same transition. It is like a ripple effect. If you help one woman, that woman could help hundreds of other women. That is how we can bring more women into this field. In AI and data, only 26 percent are women. So it’s important to encourage and support women.

Lastly, I want to emphasize the importance of believing in your ability to secure your dream job. Undoubtedly, you will face failures along the way, as is common when learning new things. However, it’s crucial to persist, continually move forward, and experiment with different approaches. Through this process, you will definitely encounter successes and, over time, build the confidence needed for your journey. Believe in yourself.

M.Ariunzaya, Mongolian Economy magazine journalist in Sweden