The deep-tech company teaching machines to read, write and speak Bahasa Indonesia - a language most global AI models still fumble.
Most artificial intelligence was trained to think in English. That leaves a strange gap: Bahasa Indonesia, spoken by more than 270 million people across the world's fourth most populous nation, has long been a second-class citizen inside the models that increasingly run banks, call centers and government services. Prosa.ai exists to close that gap.
Founded in 2018 in Bandung, Prosa.ai is a deep-tech developer of natural language processing (NLP) and speech technology built specifically for Indonesian. Its name is an acronym - "Pemrosesan Bahasa," or language processing - and its origin is academic. Two researchers from the Bandung Institute of Technology (ITB), Ayu Purwarianti in NLP and Dessi Puji Lestari in speech, wanted their work to leave the journals and reach industry. They brought in Teguh Eko Budiarto, a telecom veteran, as chief executive.
The pitch to customers is blunt. "Businesses and institutions have a lot of data, but they do not know how to extract insights out of the data," Budiarto has said. Indonesian enterprises sit on oceans of recorded calls, chat logs, documents and forms - almost none of it readable by off-the-shelf global AI. Prosa.ai turns that raw material into structured insight.
It does so through a stack of building blocks: an NLP toolkit exposed as APIs, speech-to-text transcription, text-to-speech synthesis, voice biometrics, chatbots and analytics over text and audio. The company benchmarks its NLP engine as the highest-accuracy option for Bahasa Indonesia - a claim rooted in local data and local linguistics rather than raw scale.
That focus is the strategy. Where global cloud providers offer shallow Indonesian support inside broad multilingual products, Prosa.ai goes deep into one language and its regional accents. It is a bet that in language, local beats large.
Text analytics, entity extraction, language understanding and sentiment analysis that businesses integrate directly into their systems. The primary revenue engine.
Automated transcription of Indonesian audio, including real-time meeting transcription for enterprises.
Cloud-based speech synthesis producing natural Indonesian voice-overs, expanded to around 40 distinct voices.
Voiceprint authentication and real-time fraud detection that identifies speakers by the characteristics of their voice.
NLP-driven virtual assistants for customer service and information - including the AntiHoaks fact-checking bot.
Text analytics that detects and flags misinformation and hoaxes circulating in Indonesian.
Insight, sentiment and compliance signals mined from call-center and meeting audio.
Electronic know-your-customer verification that combines language and biometric technology for onboarding.
Prosa.ai competes on the axis global giants tend to ignore: how well a system truly understands Indonesian, its slang and its regional accents. The illustrative comparison below reflects the company's positioning - specialized local depth versus the general multilingual coverage of large cloud platforms.
Figures are illustrative of positioning, not audited benchmarks.
Prosa.ai serves banks, insurers, call centers, corporations, SMEs and government agencies - organizations drowning in Indonesian text and voice data. Individuals can also buy its transcription and text-to-speech tools directly, from around IDR 50,000 (about USD 3.50) a month, using the same core technology enterprises license via API.
It authenticates customers by voice and flags fraud in real time. It transcribes meetings and calls that would otherwise go unread. It fields customer questions through chatbots, gauges sentiment across conversations, and - through the AntiHoaks chatbot built with Indonesia's Ministry of Communication and Informatics - helps the public check misinformation as it spreads.
Researcher at ITB's AI Center and lecturer in electrical & informatics engineering; the NLP mind behind the company.
ITB lecturer whose research on Indonesian voice recognition seeded Prosa.ai's speech technology.
Telecom veteran - formerly managing director at Tokyo's Brastel Telecom - who leads the business.
ITB researchers Ayu Purwarianti and Dessi Puji Lestari join Teguh Eko Budiarto to build an NLP and speech company for Bahasa Indonesia.
Raises a Series A from GDP Venture and partners with Kominfo to launch the AntiHoaks fact-checking chatbot.
Introduces voiceprint-based authentication and fraud detection for enterprises and banks.
Expands into cloud text-to-speech, speech analytics and eKYC solutions.
Launches real-time meeting transcription, grows TTS to ~40 voices, and begins a Series B round after roughly 50% year-over-year growth.
It builds natural language processing and speech technology - NLP APIs, speech-to-text, text-to-speech, voice biometrics, chatbots and analytics - specialized for Bahasa Indonesia.
It was founded in 2018 by ITB academics Ayu Purwarianti and Dessi Puji Lestari together with CEO Teguh Eko Budiarto.
In Bandung, West Java, Indonesia.
Banks, insurers, call centers, corporations, SMEs and government agencies that process Indonesian text and voice data, plus individuals using its transcription and TTS tools.
It raised a Series A from GDP Venture in 2019 and began a Series B round in 2022; specific amounts have not been publicly disclosed.