Chatbot Stories in Indonesia

The last five years chatbot technology has stolen a lot of attention. No wonder so many companies end up a bunch of implementing chatbots they themselves are complete with cool names and seem friendly. However, many chatbots still seem rigid, and need a lot of "practice". Moreover, the chatbot placed in the position of customer service.

Chatbots work by analyzing words sent by users through message channels. Then the words are sent and matched with the existing system to then provide a predetermined response.

In some cases, many chatbots cannot recognize the sentence in question, this happens because the sentence does not use the language/word registered from the system or the vocabulary in the system is incomplete.

Chatbots can be an alternative solution to UI/UX. Instead of doing tap or memorize the steps to find an info in the application or website, users just write what they are looking for on the conversation platform. Chatbot with "friendly" will find the information in question. In this case the chatbot can guide the user if he does not find the meaning of the sentence given.

However, for the case of customer service, chatbots have to work "harder and friendlier". The reason is that not all users are satisfied with the answer templates or general information. It could be that the contacting users are those who are waiting for why there is an interruption in their account or a transaction that has not been completed.

Often complaints are not written in complete and coherent sentences. Even more often it is written with passion and emotion so that sometimes sentences can be uncoordinated. Besides having to understand, chatbots must also be trained to be more polite. So, the biggest obstacle lies in understanding the language and the database of solutions provided.

If the chatbot is equipped with a learning engine that allows chatbots to be trained, many chatbots in Indonesia still need a lot of training to be able to understand more word choices, sentence structure, and most importantly the solution database. That's why many chatbots are in position customer service still often accompanied by user service agents in order to continue to provide a human touch.

Various kinds of chatbot innovations in Indonesia

In Indonesia, there are not many technology companies engaged in chatbots and AI. Two names that are consistent in business development and innovation are Kata.ai and botika. Both follow each other in terms of innovation.

Kata.ai, led by Irzan Raditya, started introducing the Kata Conversational Platform. A platform designed to provide complete solutions for Kata.ai users. In it there are various kinds of features, including Omnichat Words, Assist Words, Voice Words, and several other features.

Kata.ai began to knit the hope of becoming a "super app" in terms of chatbots and AI by starting to offer a variety of solutions. In 2020, synergies and partnerships are expected to strengthen Kata.ai and the solutions it builds.

“We believe collaboration with other service providers and platforms is key to our growth strategy, as we have done in the last two years with tech startups others, such as Qiscus and Halosis, or also partners system integrator such as Accenture, Medlinx, Sprint, Telkom Infomedia, and others. We are very open to welcoming more synergies and partnerships that can be generated in 2020," light Irzan some time ago.

Botika also does the same thing, innovation. There are two new services from Botika, namely Voicebot and Omnibotika. Voicebot is a virtual assistant that can be commanded via voice messages, while Ominbotika is a dashboard that is able to control various communication channels ranging from WhatsApp, Line, Telegram, WeChat, email, to telephone.

In 2019, approximately 20 million message traffic passed through the Botika system. More than 1500 downloads for the Chatbotika application for online shops and handling more than 2000 companies as clients. CEO Ditto Anindita Botika said that in 2020 his party will develop Voicebot capabilities and increase the number of communication channels to be integrated with the Botika system.

"Botika formed a consortium with several companies and made a Smart Speaker, namely Widyawicara, which is planned to be launched in 2020. Through our subsidiary, ARSA technology, we innovate AI Chip, which enables computing machine learning can be done on the hardware side of the IoT independently [edge computing]. This will increase the speed of the computing process and reduce the server load," explained Ditto.

Heavy duty chatbot satisfies Indonesian customers

Although it has been widely implemented in Indonesia, in fact chatbots still have a lot of homework to do, especially regarding the understanding of messages that are constrained by the various languages ​​in Indonesia and the AI ​​capabilities in it provide suitable solutions--not only providing links to FAQ articles.

Despite being one of the most advanced technologies, chatbots have the potential to fail. Especially if the data used as "practice" is lacking or inaccurate, cannot understand the conversation well, and last but not least, protect personal data.

Although not something new, chatbots still have the potential to grow. The main task of a chatbot is to satisfy customers by making themselves seem like real humans.