Dec 20, 2025 8 min read

Building an AI Personal CFO: The Sejalanin Story

AI Startups Fintech

Most Indonesians make major financial decisions with zero data. Should I buy this house? Can I afford to resign? Am I saving enough for retirement?

The answers usually come from gut feeling, not calculation. Sejalanin exists to change that.

The Insight

I noticed a pattern with friends and colleagues. Smart, educated professionals making terrible financial decisions. Not because they were irresponsible — because they had no framework for evaluating tradeoffs.

Buying a house isn’t just about the monthly installment. It affects your emergency fund runway, your retirement timeline, your ability to handle unexpected expenses. Most people only look at one variable.

The idea: what if you had a personal CFO — an AI that understood your complete financial picture and could simulate the downstream effects of any major decision?

Building Lino

Lino is our AI Personal CFO. Unlike generic financial calculators, Lino understands your specific situation:

  • Income and expenses
  • Assets and liabilities
  • Investment portfolio
  • Life goals (house, marriage, children, retirement)

Lino’s recommendations are personalized, not generic. “Save 3-6 months of expenses” is generic. “Based on your spending of Rp12M/month and your goal to buy a house in Jakarta in 3 years, you need to save Rp3.2M/month and maintain a 7-month emergency fund” is personal.

The Tech Stack

Sejalanin runs on a modern, lean stack:

Next.js (Frontend) → Node.js (API) → LLM (Reasoning) → PostgreSQL (Data)

Why Next.js? SEO is critical for a consumer fintech product. We need every article, every simulation page, every FAQ to rank. Next.js’s SSR and static generation are perfect for this.

Why a custom LLM pipeline? We use GPT-4 and Claude for the reasoning layer, but with heavy customization. Our prompt engineering includes Indonesian financial context (inflation rates, tax brackets, KPR structures, BPJS calculations). The LLM doesn’t just crunch numbers — it provides contextual, empathetic advice in Bahasa Indonesia.

Why PostgreSQL? Financial projections are complex SQL queries. PostgreSQL’s window functions, CTEs, and aggregation features are unmatched for this use case.

The Business Model

We went with freemium:

  • Free: Life Score, basic simulator, 10 Lino AI credits/week
  • Hobby (Rp29K/mo): Full simulators, priority support
  • Pro (Rp49K/mo): Unlimited Lino, advanced scenarios, export

This isn’t a get-rich-quick pricing strategy. We’re playing the long game — build trust with a free tier, convert power users to paid, and grow through word of mouth.

Lessons Learned

Start with one killer metric. For us, it was the Sejalan Score — a single number that tells you if your finances are on track. Everything else (simulators, retirement planning, investment optimization) supports that one metric.

AI isn’t the product — the outcome is. Users don’t care that Lino uses LLMs. They care that Lino tells them exactly when they can afford to buy a house and what they need to change to make it happen.

Indonesian context matters. US-centric financial advice doesn’t work here. Our inflation rate is different. Our mortgage structure is different. Our retirement system is different. Every recommendation is built on Indonesian data.

Trust takes time. Financial data is sensitive. We invested heavily in security (end-to-end encryption, no data sharing, transparent policies) but building trust still takes time. Our most effective growth channel is users showing their Sejalan Score to friends.

What’s Next

We’re expanding the simulation engine to cover more life decisions — education planning, business startup costs, migration planning. And we’re exploring integration with Indonesian banks for real-time financial data import.

If you’re building in the Indonesian fintech space, reach out. I’d love to exchange notes.