.png)
Harnessing Large Language Models for Social Impact: The Journey of Decisions Lab
- Rituparna Konar
- 2 days ago
- 3 min read
Technology should make society better, not just faster. This belief has guided me since 2016, when I joined the first Young Founders School (YFS) bootcamp at Credit Suisse in Hong Kong. That experience didn’t just teach me how to build a business; it planted the idea that entrepreneurship can be a powerful tool to solve real, meaningful problems.
Eight years later, in June 2024, I co-founded Decisions Lab with a mission: to harness AI to better understand human behavior and enable smarter, more impactful decisions.
Starting with SimCity for Policymakers
Our first concept was simple, yet ambitious: create a “SimCity”-style simulation platform for policymakers. Imagine being able to test recycling policies or housing initiatives in a virtual environment before implementing them in the real world. Our engine, powered by Large Language Models (LLMs), generated digital citizens—complete with routines, emotions, and biases.
The idea resonated. Policymakers loved the demos. But the reality of working with governments soon became clear: procurement cycles were slow, budgets limited, and tangible impact felt too far off. We realized that building great tech wasn’t enough.
Pivoting to the Social Impact Sector
Next, we turned to NGOs and family foundations. These organizations appreciated our vision, and our simulated personas offered a new lens for evaluating program effectiveness and anticipating beneficiary reactions. But we hit familiar obstacles: limited budgets, lengthy consensus-building processes, and scalability issues. This chapter taught us an important lesson in patience and the value of listening deeply to our customers.
Exploring Digital Transformation
An angel investor encouraged us to explore organizational change. In large companies, change managers often face unseen resistance from employees—an area where our simulation engine could shine.
We dove in, studied organizational behaviors, mapped pain points, and built prototypes. Feedback was polite but lukewarm. While people liked the idea, they didn’t feel an urgent need for it. The takeaway: even a great solution needs to solve a clear, pressing problem to gain traction.
Following the Data—and Finding Product-Market Fit
With every pivot, we collected powerful insights:
People prioritize solutions that help generate revenue over those that just cut costs.
High-quality, granular data dramatically boosts the value of AI simulations.
Immediate, measurable ROI makes it easier for clients to say yes.
These learnings eventually led us to direct-to-consumer (D2C) marketing. These businesses thrive on customer retention and already collect rich first-party data, perfect for our simulations. Instead of selling a product, we started solving real, high-priority problems.
Why Large Language Models Matter?
The power of LLMs lies in their ability to understand human nuance—cultural subtleties, emotional triggers, and cognitive biases. By fine-tuning models for regional audiences like Hong Kong or Singapore, our simulations began to feel more authentic, more human. This isn’t just about better marketing outcomes. It’s about empowering better, more thoughtful decision-making across sectors.
Lessons in Entrepreneurship—and Personal Growth
Starting and evolving the Decisions Lab has been like running a marathon. There were days we didn’t fully understand our users, couldn’t market effectively, or struggled to close deals. We got advice from mentors who didn’t always fit our situation. But through it all, we kept learning about humility, resilience, and how to navigate ambiguity.
Building a startup is a deeply personal journey. It’s full of difficult decisions and quiet victories. But the growth it brings is unmatched.
Staying True to the Vision
Our long-term vision hasn’t changed: use AI to help society make smarter decisions. While marketing is our current focus, it’s a proving ground. Our broader goal—to influence policy, empower nonprofits, and drive systemic change—remains at the heart of what we do. We’ve learned that practical impact often requires starting small and iterating with purpose. That’s how we’ll earn the trust—and traction—to make a bigger difference down the line.
Advice to Aspiring Entrepreneurs
To anyone considering Young Founders School or building something of your own:
Pivoting isn’t failing—it’s learning.
Be patient.
Talk to users.
Understand problems deeply.
Stay flexible.
Each step—even the tough ones—builds your entrepreneurial muscle. With persistence and curiosity, you’ll create something that truly matters.
Entrepreneurship isn’t about instant wins. It’s about thoughtful adaptation, sustained effort, and staying true to your mission. The story of Decisions Lab is proof of that. Our aim remains the same: leverage AI to create real, lasting social impact.
Written By Louis Cheung, Co-Founder of Decisions Lab & YFS Alumnus (Class of 2016)
