[BDD 2025 - Artificial Intelligence] AI for the Underdogs: Innovation for Small Businesses

BDD
AI for the Underdogs:
Innovation for Small
Businesses
Vina Sari Yosephine, Ph.D. | Dosen UPNVJ

Education
Vina Sari Yosephine
Work Experiences
● Logistics Analyst, HI
● PostDoctoral Researcher,
KAIST
● KAIST, Doctor of Industrial & Systems Eng.
● KAIST, Master of Industrial & Systems Eng.
● ITB, Bachelor of Engineering
Dosen UPN Veteran Jakarta, System Engineering, Practical AI, Project Based Learning

The Generative AI Leap
Multimodal Growth Rapid Evolution Global Impact
The AI landscape is evolving rapidly,
diverse modalities like text, image, and
video.
New advancements in AI are enabling
unimaginable creativity across various
media and formats.
.
The implications of this AI evolution
extend worldwide, transforming
industries and everyday life.

Indonesia’s Labor Market Facts
● Indonesia is part of the EAP region where
only about 10 percent of jobs involve tasks
complementary to AI
● Higher educational attainment is associated
with greater AI exposure
● Workers in non-agriculture sectors,
especially workers in the commerce sectors,
are much more exposed to AI than
agricultural workers
● In Indonesia, the relatively low level of robot
adoption is concentrated in speciﬁc,
generally lower–value added manufacturing
subsectors

In the age of
algorithms,
PURPOSE is our
greatest parameter.

The Engine of Indonesia Economy
± 65 million UMKM
≈ 99% of businesses
Contribute ~60–61% of
Indonesia’s GDP
Employ ~117 million people
≈ 97% of total jobs

The Engine of Indonesia Economy
But only ~15–16% of non-oil & gas exports
● Don’t meet international quality consistency, manual
QC, high variability
● No digital records or traceability (needed for export
compliance)
● No automated production planning
● Cannot scale production (small factories struggle with
bottlenecks)
● No logistics or supply chain integration (where AI and
IoT tools can help)
● Manual communication in data formats required by
importers
± 65 million UMKM ≈
99% of businesses
Contribute ~60–61% of
Indonesia’s GDP
Employ ~117 million people
≈ 97% of total jobs

The Reality of AI in Indonesia SME
Low operational
efficiency
Signiﬁcant barriers
to AI Adoption
Manual processes
and fragmented

Practical AI
Our economy depends on AI that SMEs can actually use.
Big AI inspires us. Practical AI transforms us
65M UMKM - 99% of all businesses - 60% of GDP - 117M workers
Small Business Needs
Not Big AI
Practical AI
When SMEs adopt AI,
Indonesia adopt AI
“
How do we build AI that SMEs can actually use?

The Need for Both AI Approaches
Driving innovation,
Advanced solutions
Inﬂuence entire industries
Create opportunities
in Indonesia's economy
Practical applications
Tailored for SMEs
Empowering
Leverage technology effectively
increase competitiveness
in a rapidly evolving market
Integration of Big AI and Underdog AI
small businesses grow & beneﬁting
from advanced technology.
Big AI
Underdog-AI

How This Looks in the Real World
● SMEs operate with uncertainty
● They have small datasets
● They need interpretable decisions
● They need fast deployment

Case Study: Textile Quality Control
Typical Traditional Textile Factory

Case Study: Textile Quality Control
A computer vision technology, with smartphone images and lightweight CNNs.
Question: what is the type of this fabric? Where is the defect located?

Case Study: Predicting Bottlenecks in Operations
A decision tree algorithm predict delays, streamline workﬂows, and enhance overall operational
efficiency within SMEs
Question: which station is the bottleneck?

Case Study: AI Lite Inventory Control
Inventory of a typical big warung or toko grosir

Case Study: AI Lite Inventory Control
A Q-learning algorithms and accessible tools like Google Sheets and ESP32, SMEs can effectively
manage their inventory without the need for complex ERP systems.
Question: Given my historical demands, how many products should I order?

Common SME AI Patterns
Small Data Clear Reasoning Low Cost
Limited Resources Integrating human in the loops No ﬁnancial burdens

Integrating SME AI in Education
Question: how is our education readiness for AI?

Real-World Project-Based Learning

Developer Playbook
Identify Pain Point Build Simple Pipeline Choose Explainable Model
Start by identifying the key
challenges that SMEs face
in their operations.
Create a clear and streamlined AI
pipeline that addresses the identiﬁed
pain points.
Select AI models that provide
transparency and understanding for
end-users and stakeholders.

The Underdog AI Movements
Impact-First Mindset Inclusion Empowerment
pain points.

Build AI that Matters
Empower SMEs Focus on Impact Simplify Processes
pain points.

Learnability
start small
representative data
Legibility
make models explain
themselves
Longevity
operate and improve
over time
3L Framework - Human in the Loop Systems
AI should lift people up — not leave them behind.

Thank You
@vinayosephine
Vina Sari Yosephine
Get in touch
vyosephine@upnvj.ac.id

[BDD 2025 - Artificial Intelligence] AI for the Underdogs: Innovation for Small Businesses

More Related Content

Similar to [BDD 2025 - Artificial Intelligence] AI for the Underdogs: Innovation for Small Businesses

More from Wintari Yasmin

Recently uploaded

[BDD 2025 - Artificial Intelligence] AI for the Underdogs: Innovation for Small Businesses