LPPM Universitas

Persada Indonesia Y.A.I

KedaiReka - Kemendikbud
Perspektif Industri Dunia Usaha dan Kerja

Bandung, 20 Maret 2021

Teguh Eko Budiarto

CEO / Co-founder of Prosa.ai

linkedin.com/in/teguheko/
teguheko@prosa.ai
Prosa Introduction
Company Overview
Prosa is the acronym for “Pemrosesan Bahasa” (Language Processing).

Vision Mission
Hyper local tech and data for
To be the leading company Natural Language and Speech
Develop technologies and
in the world to provide Processing Technology
produce the highest quality for Bahasa Indonesia
AI Technologies
data for AI in Indonesia Computer vision processing
for Indonesia
with local dataset
Artificial Intelligence Industry Trends 2021

AI sebagai salah
Perkembangan Trend
satu fokus
Industri 4.0 Conversational AI
Strategi Nasional

Peningkatan kebutuhan di
Peningkatan Digital Users,
berbagai bidang - BFSI, Telco,
Transactions, Data
Healthcare
 Market Overview
The World’s Largest archipelago
Population: 250 million
GDP Growth: 5.27% (Q2, 2018)
746 Languages & Dialects
(1 unifying Language - Bahasa Indonesia)

https://techcrunch.com/2018/11/18/google-report-southeast-asias-digital-economy-to-triple-to-240-billion-by-2025/
NLP and Speech Technology

Products Speech
Natural Language Processing (NLP)
Processing

Virtual Assistant
 Core Research

People Vehicle

Face Masked Face Face Matching Body Attribute Vehicle Counting License Plate
Recognition Recognition (1-to-1) Detection & Classification Recognition

Mask Usage Activity People Crowd Illegal Parking

Detection Detection Counting Detection Detection
IN-OUT
 Sentiment Analysis 2.0

Concept Level

1 post / article

1 or more concept (key phrase)

sentiment value per concept

 Sentiment Analysis 2.0

Aspect Level

1 post / article

categorized to pre-defined
categories

sentiment value per category

 Sentiment Analysis 2.0

Intent detection

a. Suggestions
b. Reasons
c. Complaints
d. Feedback
 Concept Relation Analysis

Apakah penggunaan kuota bantuan oleh siswa/mahasiswa sudah sesuai dengan tujuan?

11
 Customer Experience Time Trends of Sentiments
ke blokir

mengalami masalah
Positive Neutral Negative Daily Weekly Monthly Yearly Range+

bank raksasa tersendat

internet banking 2800
Sungguh amat teramat sangat ribet sekali bisa reset 2400
hapus posisi back office
macet mulu
error 1600
bank 1200
bni
kena biaya
terkuras transfer
800
saldo
gagal 400
transaksi
aplikasi
0
p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p p
Se Se Se Se Se Se Se Se Se Se Se Se Se Se Se Se Se Se Se Se Se Se Se Se Se Se Se Se Se Se Se
pembayaran bca minta diulang 01 01 02 03 04 05 06 07 08 09 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30
dapat di transfer ke rekening resmi bri
resmi memiliki 99,9 persen
bisa memanipulasi sistem
canggih
List of Sentiment Marker
mobile banking
eror
mobile banking dilarang melakukan pemblokiran bisa melakukan enable kartu mobile
canggih tampil
bingung2 gt atm mobile banking juga tersedia
kaga bisa mempengaruhi info / layanan melalui Tetap bisa bertransaksi melalui mobile
internet / mobile banking banking dan internet banking
dipakai untuk 2 mobile bisa melakukan enable kartu
tiba2 ke logout sendiri eror bingung2 gt
block kartu atm mobile banking juga tersedia tidak bisa block kartu
di suruh
dirindukan lama-lama kerasa bisa memanipulasi sistem dilarang
ketelen tanpa harus ke bank dipake untuk 2 mobile kesel
cuman ngadem
membantu untuk membunuh uang ku
secara tidak perlahan
benci
atm aman ga disalahgunakan internet banking
bisa reset
rusak
di blokir takdir ke blokir mengalami masalah
boros jangan eror gak ada biaya admin
tarik semua tambah followers
bisa
Chatbot for
better survey
Messenger in the most used
application on respondent mobile

Can capture free text responses

Highly customizable and adaptable

questions

Resume survey sessions for each

respondent without worry of
connection drop
Robo Call for
better survey
Auto generated sound from text
with natural voice

Transcribe the respondent voice

to determine next action

Transcribed text also

automatically being analyzed to
capture the response essence

Fully automated to save human

effort
 General Purpose
1. To provide monitoring new regulation
issued by authority automatically and
comprehensively
2. To evaluate (review & comply) new
regulation with existing regulations
3. To manage regulation document

For Compliance Specific Purpose

Prosa Regtech Compliance started by
focusing on Banking / Finance domain
Product Highlight
Aplikasi web based dengan teknologi kecerdasan
buatan (AI) yang dapat mengenali pembicara
dalam rapat dan mentranskripsi ucapan yang
disampaikannya secara tepat serta dalam waktu
yang singkat.

Kelola rapat dan kirim pemberitahuan ke peserta

rapat, merevisi, mempublikasikan dan eksport
MoM rapat.

Cara yang efisien menangkap informasi rapat

dan menyediakannya untuk para peserta rapat,
serta anda yakin tidak akan melewatkan detail
apapun.
 Recording transcription to aid on Focus Group
Discussion

Transcribe the respondent voices

Capture keywords, topics,

sentiment

Enabled advanced search

Enable advanced analytics

Some marketing researches are being held on FGD forms
SPEECH ANALYTICS
Voice Bot or speech analytics for
Turn customer service calls into meaningful insights
customer service can help Identify
customer’s key concerns, classify
topics, analyze sentiments, and
categorize whether customer calls are
queries, complaints, or suggestions.
 Computer Vision
Detect Vehicle - Traffic Quantity
Computer Vision
Crowd Detection
Computer Vision
Visitor Counting
 Computer Vision
Visitor Profile
Face detection - age, gender, emotion, fashion style
 Computer Vision
Heatmap of Visitors

● MARKETING EFFECTIVENESS ● OPTIMIZE POPULAR AREAS

● MEASURE SUCCESS ● STAFF OPTIMIZATION
● REAL-TIME CUSTOMER TRACKING ● DELIST UNATTRACTIVE PRODUCTS

Source: retailflux, identifeye

 Prosa.ai
Clients and
Partners
Prosa has partnered and served
governments, universities, public
and private companies

24
Core Development Principles

0 Extensive AI Research

Human-in-the-loop automation

Scalability

Extensibility
 Product Development Iteration
Started on 2018 and has been deploying multiple products on clients production environment

Get Real Data and improve accuracyFast iteration

Extensive experiments
Products
Training Servers Model Optimisation and Product Bundling
On Prosa.ai Premise Deployment - Client
Research and Premises ● Software Dev
Innovation Data storage, ● API based
preparation, ● Model Versioning ● SaaS, On Premise Virtual Assistant
annotation Model ● Inference ● License
Partnership : Training hardware
- University Management
optimization ● Edge Devices
- Government AI Training
- Other ● Packaging
Model Servers
Industry ● Deployment
Development ● Parallel
Processing Business Development Speech
Analytics

Meeting
NLP Technology Transcription
● Price Packaging
● Marketing
Speech Processing Technology ● Partnerships
● Funding
Computer Vision Technology
 Our Team

CEO Board of Director 30+ AI Engineers 20+ SW Engineers 60+ Data

Developer
TEGUH EKO B, ST. MT. CISA ON LEE
2 previous digital startups GDP Labs CEO & CTO,
GDP Venture CTO.
Worked 7 years in Japan software industry
30+ years experience in IT
industry

Computer Vision
Text Scientist Speech Scientist Engineering Advisor Data Mining Advisor Linguistic Advisor
Scientist

Dr. AYU DESSI PUJI Dr. ACHMAD NUGRAHA UTAMA, IR. WINDY GAMBETTA, Dr. TOTOK
PURWARIANTI LESTARI, Ph.D MBA, CISA, CRISC, CGEIT SUHARDIJANTO
IMAM Ph.D Pusat AI ITB
Pusat AI ITB Pusat AI ITB Linguistik UI
Pusat AI ITB Pusat AI ITB
 Problem and Challenges

● Training Data Availability

● Talents

● Research

● Investment

● Market Adoption
University Collaboration for AI Talent

AI Industry Academic Partnership Recruitment

Key Assets
On the Job Training
Curriculum Internship
Specific Interest Coding Standard
People Foundations NLP - Speech - Vision
Architecture Standard
Specific Interest Small Projects
Final project But effective
Corporate Cultures
Data

Talent Scout Culture Fit Test Growth - Kaizen - Excellence

Know Project Management

How Experience
Sponsorship
Industry
Performance Appraisal
Research and Innovation Collaboration

Research Productization Grant

Improve Tingkat Kesiapan Government:

Technology Exploration
Teknologi to Level 7-9 LPDP, BRIN, KedaiReka

Small Scale Data Gathering Large Scale Data Gathering University:

Inovasi ITB

Industry Level Product

Prototype
Maturity Startup Accelaration dengan
Global Tech companies

Lab Scale Test Real Environment Testing Foreign Aid :

USAID, etc.
KedaiReka Launched
Apa itu KedaiReka
Platform Kolaborasi IDUKA - Dikti
Kemitraan disupport pendanaan
 Rasio Pendanaan

Maks Rp. 5M per kegiatan

Masalah yang diselesaikan
Prioritisasi Dampak
Alur Seleksi
Benefit untuk Industri
Peluang Cipta
Definisi Peluang Cipta
Peluang Cipta - lanjutan
Tampilan Layar Profil Industri
Tampilan Layar Peluang Cipta
Kreasi Reka
Eksplorasi Kreasi Reka
Penjelasan Eksplorasi Kreasi Reka
KedaiReka Follow Up Email
Aktivitas Kolaborasi
 Pembuatan Proposal

● Setelah menemukan mitra yang tepat di kedua belah

pihak, Insan Dikti + Industri

● Kesiapan pendanaan / investasi secara cash ataupun

in-kind dari Industri

● Mengikuti Template yang diberikan oleh KedaiReka

● Kelengkapan informasi
○ Cashflow kegiatan - rencana penggunaan dana
○ Proyeksi manfaat, dampak, maupun revenue
 Kegiatan Selanjutnya

● Review oleh Perguruan Tinggi

● Revisi Proposal apabila dibutuhkan
● Pengajuan proposal ke Kemendikbud via
Perguruan Tinggi
● Pitching / presentasi proposal
● Seleksi

Note: Prosa saat ini masih dalam proses Review Proposal untuk 1 kegiatan
Terima kasih
Diskusi - Tanya Jawab

Teguh Eko Budiarto

CEO / Co-founder

linkedin.com/in/teguheko/
teguheko@prosa.ai