Konvergen AI Introduces Data Capture Platform to Facilitate Photo Transcription

Public and private sectors in Indonesia have massively digitized their services. The goal is clear, to make existing business processes faster and more efficient. In fact, some digital processes are yet to fully automated. For example, with a digital application, a company still has to go manual on validating the online submission to ensure that the attached files are appropriate.

These problems encouraged Lintang Sutawika, Roan Gylberth, and Timothy Devin to initiate the Konvergen AI. It's a startup focused on developing artificial intelligence technology for data capture - refers to the process of collecting data from paper or digital documents using optical character recognition (OCR) components. OCR is functioned to produce a transcription of texts in digital format from photos, images or scans.

How does it work? For example, there is an application for shopping records, developers can use the Konvergen AI feature to scan the receipt with a mobile camera. It allows the application to recognize the products and nominal listed, users don't need to write manually.

"In brief, Konvergen AI's products and services are used to improve performance efficiency involving manual and repetitive processes, as data input. Currently,  the Konvergen AI system can process various documents ranging from identity-related such as eKTPs or Family Cards, and documents such as invoices," The Co-Founder, Lintang Sutawika said.

Konvergen AI provides data extract service for various purposes / Konvergen AI

He also mentioned, to begin with, the Konvergen AI product was aimed at financial companies, such as banking, insurance, and financing; however, it can also be adjusted to other industries such as export/import and legal processing. As SaaS, products are commercialized through a subscription license.

"Currently, we are assisting a private bank in registering new customers. In this case, our system helps to input customer identity data as stated in documents such as eKTP and NPWP. With this data capture system, the banking registration and checking process should be faster," he added.

Seed funding from GDP Venture

Konvergen AI was first founded in late 2018 and proceed more than one million documents from a number of customers. They also secured seed funding from GDP Venture, a venture capital actively involved in artificial intelligence startups. Aside from Konvergen AI, GDP also invested in Datasaur, Prosa,ai, 6Estates, Qlue, and Glair.

Konvergen AI's Founders, Timotius Devin, Lintang Sutawika, and Roan Gylberth / Konvergen AI

In terms of founders, Sutawika mentioned the first time the three of them met while studying master's degree of Computer Science in Universitas Indonesia.

"At first, the three of us ran several computer vision and image processing projects. A simple thing, such as reading or inputting data can become very overwhelming when it was done repetitively and on a large scale. We thought we can engineer the development of artificial intelligence to reduce manual work."

In actual days, there are many vendors offering data capture services, such as cloud computing providers, they have various cognitive services, one of which is an OCR-based system. Konvergen AI aware of the issue and came up with such differentiation.

"The system we developed is an end-to-end service. For example, some cloud computing service vendors have OCR API services, yet in daily practice, these OCRs still have to be processed before the outputs considered to be suitable for use. The system we've developed has been adjusted for non-ideal conditions," he said.

Closing the interview, he explained his plan in 2020, "We plan to increase the number of  ready-to-extract documents and increase the language types, therefore the Konvergen AI can serve more industries and customers, especially in Indonesia."

–Original article is in Indonesian, translated by Kristin Siagian