AD
Copy 2
Moderate emotion★3.7/ 5.0
Behavioral AI for Southeast Asia
Map human behavior.
We combine behavioral science and AI to simulate how people decide at population scale.
Backed by 917VenturesGlobe’s venture builder · Velocity Program
Considered choices
Open to the new
Pragmatic middle
Spontaneous choices
Research arXiv preprint, May 2025
69.8MModeled Filipino psychographic profiles
92%Survey-response alignment (QWA)
86%Human-response alignment
Commercial AdSpark blind backtest
3.4×Higher click-through rate
2.6×Lower cost per click
Methodology and sources
Research
Tia, Lanuzo et al., Sentiment Simulation using Generative AI Agents, arXiv preprint 2505.22125, May 2025. Not peer reviewed.
The 69.8M profiles model Filipino consumers using behavioral and psychographic evidence across 17 regions. Survey-response alignment was 92% (QWA); sentiment alignment was 81–86% across tested scenarios.
Commercial
Blind backtest with AdSpark: approximately 19,900 ad-pair comparisons and 800M impressions. Predikta ranked the better-performing ad in 59–61% of pairs overall, against about 50% by chance, and in 64–70% of pairs on the conversion subset.
Results are retrospective and specific to the tested data. See the full case study.
The human question
Understand why
people choose.
Data shows what happened. Behavioral science helps explain why.
-
1Observe
Data shows what people did.
-
2Understand
Behavioral science explains why responses differ.
-
Illustrative
3Simulate
AI models how populations may respond.
Decision contexts
One model.
Many industries.
The Behavioral Model provides use cases across key industries.
Marketing & advertisingWhich message will land with your audience?
Test messages and ideas with simulated Filipino audiences.
Explore PrediktaBusiness & enterpriseHow might people respond to a new offer?
Explore consumer response, product decisions, adoption and loyalty.
Healthcare & wellnessWhat shapes response to a health message?
Understand patient behavior to improve health communications, services and interventions.
Government & public sectorHow might a community respond to a policy?
Explore public response to policies and programs.
Defense & intelligenceHow does context shape population response?
Understand population dynamics, information environments and human response.
Each use case needs its own data, calibration, validation and safeguards.
Predikta by Netopia AI
The Filipino Market
on Call.
Know your audience.
Find the message that connects.
Audience Builder
See who they are.
Know what drives them.
Set filters like region and age, and Predikta builds the audience. See the persona mix inside it, so you can shape a message around how these people choose.
Try your campaignPrediktaSimplified preview
1Define the audience with filters
Audience name
NCR Gen Z
Filters:National Capital Region (NCR)Young Adults (18–27 years old)
2See the personas inside it
SegmentationPersona mix
- Traditional Consumers56%
- Cautious Consumers18%
- Modest Explorers15%
- Prudent Explorers11%
Traditional ConsumersLargest persona
“I prefer familiar brands that I can trust to deliver good value.”
Familiarity, reliability, and value guide their choices.
Audience Builder · NCR Gen ZSample data
Campaign Simulation
Upload the creative.
Understand the response.
Compare overall ratings and emotional responses before launch.
Try your campaignPrediktaSimplified preview
Testing and refinement
Campaign comparison
Three ads. One modeled audience.
Emotion profile
Emotion score · 0–8Copy 2Copy 5Copy 6
Campaign comparison · 3 adsSample data
Predikta Listen
Read the response.
Understand the conversation.
After launch, Listen groups live comments into themes and pulls out representative comments, so you can see what people are saying and what to change next.
Try your campaignPrediktaSimplified preview
Predikta Listen
Sample campaign: new iced coffee flavor
Live comments, grouped into themes
Sentiment breakdown
52% Positive31% Neutral17% Negative
Top themes · representative comments
41% of comments · mostly positive
“Not too sweet, just right. Sana permanent na ’to!”
27% · mixed
“A bit pricey for the size, but I’d buy it on promo.”
19% · mostly questions
“Available na ba sa Cebu? Wala pa sa branch namin.”
Summary
People like the taste. Price is the main hesitation, and many ask where to buy it. Next post: list the stores and the promo.
Predikta Listen · comment themesFictional sample
MAP ME
Your data.
A deeper understanding.
Turn CRM, loyalty and transaction data into behavioral profiles. Explore what drives your customers, then test ideas with them in Predikta.
Talk about MAP MEYour data today
Transactions · CRM · Loyalty| Customer | Top category | Orders |
|---|---|---|
| C-10482 | Snacks | 12 |
| C-10483 | Beverages | 4 |
| C-10484 | Personal care | 9 |
Enriched: C-10482
Ready in PrediktaIllustrative example of MAP ME psychographic scores
AdSpark blind backtest
Better signals.
Stronger campaign results.
Top 20%
3.4×
Bottom 80%
1×
Top 20%
0.38×
Bottom 80%
1×
Top 20% of ads by Predikta score compared with the bottom 80%. Each metric is indexed to its bottom-80% baseline (1×). Retrospective results; future performance may vary.
Partner perspectives
From our partners.
“With Predikta’s ability to understand the audience, we can prescribe what our clients and brand partners should do even before they implement…”
“Solutions like Predikta can then help businesses move from reactive strategies to proactive, data-led success.”
“At Brave Connective, our mission is to bridge the gap between technology and human connection, and Predikta’s predictive capabilities align perfectly with that vision.”
How we map
The science behind human response.
Behavioral science explains why responses differ. AI lets us explore those differences at population scale.
- 01
Measure
Start with human evidence.
Use surveys and observed behavior to describe personality, values, beliefs and the context of a decision.
Behavioral surveys · Decision journeys · Quantitative pulses
- 02
Simulate
Model different possible responses.
Represent people as agents with different traits and motivations. Explore how their choices form population patterns.
Psychographic profiles · Agent-based models
- 03
Calibrate
Learn from real outcomes.
Compare simulated and observed responses to identify gaps and refine the model’s assumptions.
Observed responses · Model refinement
A learning loop: evidence, simulation and calibration.
Illustrative
Inside the Netopia Behavioral AI Engine
Human evidence
National-scale surveys and observed behavior.
Behavioral profiles
Personality, values, beliefs and context.
Population simulation
Explore how different people may respond.
Scientific foundations
Beyond demographics.
Who people are is only part of the story. Our models also consider what they value and the situations they face.
Traits & motivations
Personality, values, beliefs and lifestyle help explain differences in response.
Context & influence
Historical context and social factors help explain why the same person may choose differently.
Individual to collective
Agent-based models explore how individual choices and interactions become population patterns.
Simulations explore possible responses. They complement field research and depend on the evidence and assumptions behind them.
Read the researcharXiv preprint · May 2025 · Not peer reviewedIn the news
Recent coverage.
AdSpark partners with Netopia AI to bring Filipino behavioral intelligence into marketing work
Signal Lab
Ideas on human behavior.
Signal Lab is our open behavioral observatory: research, experiments and lessons from building behavioral AI.
FeaturedPerspective
Prediction tells you what. Simulation tells you why.
Jason Albia · 8 min read
The people behind the map
Meet the minds
behind Netopia.
Behavioral scientists, AI engineers, researchers and business builders.
Founders
Axel Kornerup MPA
Co-founder & CEO
Built and sold Netopia, the Philippines’ largest internet café network.
Jason Albia MS
Co-founder & CTO
Computational scientist with 40+ papers. Chief architect of Predikta.
Advisors
Jane Walker
Advisor
25+ years across PLDT, Singtel and San Miguel.
Aldo Carrascoso
Advisor
InterVenn Biosciences.
The wider team
Research, engineering, data and product.
Portraits and names of 19 members of the wider Netopia AI team, including advisor Jojo Flores.
Combined experience
Across the team’s careers and previous ventures.
Business building
90+years
Building, operating, investing in and advising businesses.
Behavioral research
80+years
Marketing and research, including university collaborators.
AI engineering
40+years
Machine learning, data science and computational research.
Value created
$1.5B+
By the founders’ previous ventures.
Years are combined across team members’ careers. The $1.5B+ figure is value created by the founders’ previous ventures. It is not Netopia AI’s own funding or revenue.