Irzan Raditya of Kata.ai on building conversational AI tech: Startup Stories

Kata.ai builds sophisticated chatbots.

By

Irzan Raditya of Kata.ai on building conversational AI tech: Startup Stories

Entrepreneurship is no easy feat—at least that’s what Irzan Raditya believes. Raditya is the co-founder and CEO of Kata.ai, Indonesian AI company that gives digital systems the capability of processing human speech.

Young and tough, Raditya has founded a number of startups since 2012 while he was still residing in Germany. The most significant one was YesBoss. Founded in 2015 after Raditya returned to his homeland, the platform offers virtual assistant services through text messages and a mobile app. However, he had to deal with the fact that YesBoss’ business model was not sustainable and needed to pivot the business into Indonesia’s B2B segment by rebranding it as Kata.ai.

Raditya says that the decision was a tough one to make, especially since the company had to lay off a number of employees whose functions were no longer relevant to the new business. Even so, Raditya and his co-founders, Wahyu Wrehasnaya and Reynir Fauzan, quickly recovered and focused on Kata.ai’s platform, building and offering technology for companies to develop their own chatbots. The idea turned out well this time, with Kata.ai showing consistently promising development. According to CrunchBase, the company has raised a total of USD 3.5 million in two rounds, with its Series A funding coming from the Trans-Pacific Technology Fund.

KrASIA recently spoke to Irzan Raditya about the importance of chatbots for businesses and his aspirations as an entrepreneur.

KrASIA (Kr): How important is the chatbot solution for businesses, especially small and medium enterprises (SMEs)?

Irzan Raditya (R): According to our internal research, Indonesians can have up to four chat applications on their smartphones. And with the high number of online transactions made through social media channels, we believe chatbots are the best means for businesses to reach consumers. SMEs can serve their customers easier on social media accounts like Instagram, WhatsApp, and Line without the help of [human] admins. In serving the SME segment, we partner with Halosis, a platform that supports SMEs to improve their online sales performance through social media. Our technology is used to build chatbots at Halosis that function as virtual admins to serve buyers.

Kr: What are the key features of Kata.ai’s chatbots?

R: Our key advantage lies in natural language processing (NLP) capabilities applied to the Indonesian language. NLP is a technology that allows computers to understand the meaning of human speech in the form of text. We can guarantee that our NLP is the most comprehensive in Indonesia because we utilize more than 500 million text messages from all active chatbots on our platform as training data to improve our model for language comprehension.

We also have the “Kata Supermodel” feature, where chatbots can instantly recognize several components of the conversation without configuration and training data, such as the name of the person, area, location, time statement, number, email, and the currency in a series of sentences.

In early 2019, we became an official partner to offer the WhatsApp Business API, which allows businesses to utilize official WhatsApp accounts that are directly connected to the company’s operating system. This way, they can offer a variety of services, using WhatsApp as a secure communication path. In addition to integrating a virtual chatbot assistant, the service also includes sending one-time passwords, notifications, and so forth.

Kr: What services do you offer to customers? How do you train chatbots to grow in line with the customer’s business development?

R: Besides developing chatbots, we also have a consulting service to help our clients maximize the benefits of utilizing the chatbot. We believe that the sustainability of our chatbot product is very dependent on ROI for the customer, whether it provides better customer service, cost-efficiency optimization, or increased sales.

There are two components on our platform that can be used to continue developing the chatbot function, namely Bot Studio and Natural Language Studio. Bot Studio is a feature to develop the conversation flow according to the needs of the user, while Natural Language Studio aims to further train the NLP model so it can be more accurate in detecting the meaning of a user’s speech. Renewing the flow of conversation and training data are two things that must be done continually to improve the quality of all chatbots.

Kr: How does Kata.ai impact customers’ business?

R: Based on data from one of our clients, 96% of queries that enter their customer service channel are regularly asked questions that can be handled directly by chatbots, so their customer service agents can focus on handling more serious complaints. Several clients also use chatbots to sell phone credit and data packages, which leads to a new source of revenue that continues to grow—the average sales figure can increase up to 49% every month.

Kr: What are your thoughts on AI adoption in Indonesia today? What are the challenges?

R: Beyond the field of language, the implementation of AI technology is quite varied in Indonesia. For instance, image recognition is integrated with CCTV and AI camera systems for credit score predictions, which are increasingly used by banking companies and fintech services. However, the adoption of this technology is still limited to large companies. So we want more SMEs to adopt AI in the future.

A big challenge in achieving wider AI adoption is the availability of data. AI can only develop if there is a lot of data to build predictive models. This is also the reason why AI implementation is still focused on industries that have access to large sets of consumer data, such as banking, telecommunications, and e-commerce.

Kr: As a young entrepreneur, what is your philosophy in developing a business and leading a team?

R: I have the principle of “people first, product second, then profit.” I believe that once we take care of our people and empower them with ownership and guidance, the potential is limitless and eventually a good product and profit will follow. For me, the biggest asset in a company is not the technology or business model, but the talent. I believe what Jim Collins, the author of the management book Good to Great, says that if you have the right people on the bus in the right seats, you can go anywhere you want.

In hiring, I believe in always hiring people who are smarter than me because it will help me accelerate the company’s growth and save me from headaches.

Kr: What does failure mean to you?

R: In this life, either you win or you learn. I have failed in building four startups before establishing YesBoss / Kata.ai, and each failure contained lessons that helped me improve in the next venture. As a software engineer, I refer to myself as software with “permanent beta” status. When a program is in the beta phase, the software will continue to be tested in order to get feedback for developing a better version. In order to grow together with the company, I need to educate myself and there should never be a “full version” of me.

Kr: What’s next for Kata.ai?

R: This year, we’ll explore the “vertical AI” strategy as we believe that AI can solve big problems in many industrial sectors in Indonesia, from financial services to health, to education.

This article is part of KrASIA’s “Startup Stories” series, where the writers of KrASIA speak with founders of tech companies in Southeast Asia.