Agentic AI for Disaster Risk Management
The Indonesian city of Samarinda faces recurring risks of landslides and floods. Local authorities like BAPEDA (Regional Development Planning Agency) and NGOs like Kota Kita needed tools to predict, assess, and manage disaster risk. Challenges included: • Fragmented data sources: climate models, topography, community surveys were not integrated. • Reactive workflows: local authorities only responded after disasters occurred. • Low decision support: existing tools gave raw data but no actionable foresight.
Why did it matter to us?
For Kota Kita and BAPEDA: without a predictive, explainable tool, communities remained exposed to sudden floods and landslides, putting lives and infrastructure at risk.
For IBM: this project was a chance to demonstrate that agentic AI could drive social good, positioning IBM as a trusted partner in humanitarian innovation.
For the city of Samarinda: disaster preparedness was tied directly to public safety, governance credibility, and urban resilience.
How my model addressed it
I designed an agentic AI solution platform that didn’t just predict risks, but also recommended actions across multiple stakeholders.
Risk Mapping & Scenarios
AI analyzed geospatial + hydrological data to surface possible risk zones.
Interactive maps highlighted varying degrees of flood/landslide probability.
Persona-Based Design
Conducted TIGER team interviews with BAPEDA officials, city planners, and community representatives.
Built tailored views:
Planners saw “long-term urban planning impact.”
Community leaders saw “near-term evacuation guidance.”
Agentic AI Workflows
Platform didn’t just give probabilities; it suggested mitigation actions (e.g., reinforce drainage, reroute traffic, pre-position resources).
AI agents simulated consequences of interventions (e.g., what happens if evacuation is delayed by 2 hours?).
Collaborative Research & Iteration
Rapid TIGER sprint cycles with IBM researchers and developers.
Iterated prototypes in collaboration with Kota Kita NGO stakeholders.
How it will impact business
For Samarinda/BAPEDA:
Improved disaster readiness by moving from reactive to proactive.
Better allocation of limited city resources before floods hit.
For IBM:
Showcased IBM’s agentic AI capabilities in high-stakes environments.
Strengthened IBM’s positioning in public sector AI and resilience solutions.
For Kota Kita:
Equipped the NGO with a digital advocacy tool to push for policy changes and infrastructure investment.
What are the consequences of taking actions through this model?
Positive Consequences:
Citizens gain safer evacuation and earlier warnings.
Local government earns trust by showing proactive disaster management.
IBM demonstrates AI for good, expanding credibility in new geographies.
Strategic Consequences:
Adoption could scale to other Indonesian cities or even broader ASEAN disaster-risk projects.
Builds a replicable AI framework for humanitarian + civic use cases.
Missed Consequence if ignored:
Without this system, Samarinda remains vulnerable to reactive crisis management, with lives, homes, and city budgets repeatedly at risk.
