Ventry

Ventry gives every MSME an AI-powered CFO, COO, CMO & CEO through WhatsApp automating finances, inventory & strategy so micro-entrepreneurs can grow, not just survive.

Comment

Inspiration

My mother is a baker. For years, she has woken up before sunrise, baked through the day, and ended every night writing numbers in a notebook alone. No advisor. No tools. Just her best guess.

I watched her business stay small, not because she lacked skill or dedication, but because she never had access to the right intelligence. I watched her get tired the kind that builds over years of working hard without the tools to work smart. She keeps going because the family depends on it.

One day it hit me: she's not failing. She's just never had access to what large companies take for granted.

And she is not alone.

Globally, more than 400 million MSMEs form the structural backbone of every major economy providing 60-70% of employment and 50% of GDP worldwide. Yet these businesses are trapped in a low-productivity equilibrium: they lack access to the financial advisory, operational consulting, and marketing intelligence that larger firms take for granted.

In Indonesia alone our beachhead market 66 million MSMEs contribute 61% of GDP and employ 97% of the national workforce. Yet:

  • Only 38% of MSME owners have basic financial literacy
  • 59% are not yet digitally connected
  • 74% lack formal credit access
  • Over 50% of owners have only primary education
  • MSMEs contribute only 15.8% of Indonesia's exports most remain trapped in local markets

The MSME finance gap in emerging markets alone stands at US$5.7 trillion but the advisory gap is even more consequential. MSMEs cannot access credit because they cannot produce financial documentation, because they lack the tools and knowledge to create it in the first place.

A critical but overlooked consequence is resource waste. Without demand forecasting, micro F&B businesses chronically over-order perishables, overproduce, and generate dead stock contributing to the FAO's estimate of 1.3 billion tonnes of food lost or wasted annually worldwide.

MSMEs do not fail because their owners lack ambition. They fail because they lack access to the executive intelligence that drives growth.

That insight became Ventry.

What does our solution do?

Ventry is an AI-powered operating system for MSMEs that deploys a Virtual C-Suite four AI agents (CFO, COO, CMO, CEO) delivered natively through WhatsApp and a customizable web dashboard.

🔗 Explore our prototype: ventry-ai.sarana.ai

The platform provides:

  • Automated bookkeeping record transactions via WhatsApp chat, zero manual entry
  • Real-time financial intelligence margins, COGS, break-even, cash flow analysis
  • Demand forecasting & Smart Scheduling AI predicts demand spikes, auto-generates purchase orders, sends to suppliers via WhatsApp, tracks delivery
  • Proactive strategic recommendations Virtual CEO surfaces Pareto (80/20) insights, MoM growth, expansion readiness
  • Customer intelligence RFM analysis, CLV scoring, churn detection, automated comeback campaigns

Ventry's Dual Engine architecture automatically configures itself for two universal MSME types:

  • Goods Engine: COGS, inventory turnover, reorder point, spoilage rate, waste reduction
  • Services Engine: COSR, capacity utilization, idle time, booking density, consumable reorder
Agent Function Example Output
Virtual CFO Cash flow, COGS, margins, break-even, dynamic pricing "COGS for nasi goreng +12%. Raise price Rp 2K or switch to Supplier B (8% cheaper)."
Virtual COO Inventory, reorder alerts, supplier scoring, Smart Scheduling "Product X runs out in 3 days. Auto-ordered from Supplier A. Arrival: Thursday."
Virtual CMO RFM analysis, CLV, churn rate, comeback campaigns "15 regulars absent 30+ days. Comeback promo scheduled WhatsApp Friday 9AM."
Virtual CEO Pareto (80/20), MoM growth, ROI, expansion readiness "Top 3 products = 78% profit. Revenue +8% MoM. Spoilage down 18% this month."

Social Impact: By solving financial invisibility, Ventry directly addresses six UN SDGs (1, 4, 5, 8, 10, 12). Every structured financial record created builds a digital credit profile unlocking formal lending. Every demand forecast made reduces food waste. Every AI interaction teaches financial literacy to an owner who never had access to it.

How we built it

Ventry is built on three core pillars:

Pillar 1: Operational Backbone

  • Integration Mode: API-based sync with Shopee, Tokopedia, GoFood/GrabFood, and bank transaction feeds for MSMEs with existing systems
  • Greenfield Mode: WhatsApp or web input for MSMEs without any existing system Ventry becomes their primary data system

Pillar 2: AI Business Intelligence Engine

  • RAG architecture: user's business data + industry knowledge base + LLM = contextual, grounded recommendations
  • ML models: XGBoost, Prophet, LSTM for demand forecasting, anomaly detection, and trend analysis
  • Smart Scheduling: AI-automated procurement based on sales velocity, supplier lead times, and demand forecasts

Pillar 3: Virtual C-Suite Advisors

Four specialized AI agents each grounded in the user's actual transaction data surface contextual recommendations across finance, operations, marketing, and strategy.

Tech Stack

Layer Technology
Frontend Next.js + React
Backend API FastAPI (Python)
WhatsApp Node.js + Twilio API
LLM Core OpenAI GPT + open-source fallback
Vector DB Pinecone (RAG embeddings)
ML Models XGBoost, Prophet, LSTM
Transaction DB PostgreSQL
Analytics DWH BigQuery
Payments Xendit / Midtrans (QRIS)

Ventry is built as a cloud-native, horizontally scalable platform using containerized microservices (Docker + Kubernetes). Region-based deployment (GCP Jakarta) ensures data sovereignty compliance with Indonesia's PDP Law (2022).

Validation

To validate Ventry's core concept, we shared the solution with an active MSME owner in our network. The response was immediately positive they identified the AI C-Suite advisory and Smart Scheduling as features that would directly solve pain points they face daily, particularly around demand forecasting and knowing when to reorder ingredients. This early validation confirmed that the WhatsApp-first approach and conversational AI output format resonated with real MSME owners, and reinforced our decision to prioritize zero-learning-curve UX over feature complexity.

Challenges we faced

1. Coordinating a Distributed Team Under Time Pressure One of our biggest real-world challenges was finding time where all team members could meet and align simultaneously. Decisions about product direction, feature prioritization, and business logic had to be made asynchronously requiring us to build strong documentation habits and clear ownership across the team from day one. Status: Managed we adapted by working asynchronously and aligning on key decisions over WhatsApp.

2. Translating Business Complexity into Buildable Detail Designing a platform that serves as CFO, COO, CMO, and CEO simultaneously sounds compelling on paper but the development detail required is immense. Defining exactly how the AI should respond to every input type, edge case, and business scenario across the Dual Engine (Goods vs. Services) forced us to go much deeper into product specification than we initially anticipated. Status: Partially resolved core flows are defined, but full edge case mapping remains ongoing.

Accomplishments we're proud of & What we learned

Accomplishments

  • Built a Dual Engine architecture that serves both Goods and Services MSMEs under one unified platform
  • Designed a WhatsApp-native AI interface that requires zero learning curve, meeting MSME owners where they already operate
  • Mapped a clear path from data input → AI processing → actionable output across all four C-Suite agents with real example outputs
  • Demonstrated financial viability: NPV Rp 58.7B, ROI 339%, payback within 26 months, LTV:CAC 12.1x
  • Addressed six UN SDGs within a single platform
  • Projected 7,500 tonnes of food waste avoided annually by Year 3 — turning a financial problem into an environmental solution
  • Designed a globally scalable architecture with language-agnostic NLP and WhatsApp's pre-built distribution across 180+ countries

What we learned

Building Ventry taught us that the interface is the barrier, not the technology. The reason 83% of MSMEs fail to adopt digital tools isn't resistance it's that existing software was never designed for them.

We learned that financial invisibility and environmental waste are the same root problem. Solving data visibility solves both simultaneously Ventry's SDG 12 impact is not an add-on, it is a direct consequence of solving SDG 8.

We also learned that one engine can't fit all MSMEs which is why the Dual Engine became central to the architecture, not an afterthought.

Finally, we learned that the advisory gap is more consequential than the finance gap. Ventry doesn't just automate bookkeeping; it teaches financial literacy through every AI interaction.

What's next for Ventry

Short-Term Improvements (Next 1–2 Weeks)

If we had more time, two immediate improvements would significantly strengthen the product:

1. AI Training on Real MSME Data The current AI engine relies on general LLM knowledge combined with RAG. With more time, we would feed the model with real Indonesian MSME transaction patterns making recommendations significantly more accurate and contextual to local business behavior.

2. Streamlined User Flow for Non-Tech Users Currently, users need to interact with the AI through multiple prompts to access certain features. We would redesign the user flow so that non-tech-savvy MSME owners can access all features with minimal prompting through guided onboarding, quick-action buttons, and proactive AI-initiated check-ins rather than user-initiated queries.

Long-Term Scaling Roadmap

Phase 1 2026–2027: Product-Market Fit (Jabodetabek) MVP launch targeting F&B and Services MSMEs in Jabodetabek. 30-day trial funnel → 6,000 paid users. Bahasa Indonesia + English.

Phase 2 2027–2028: Geographic Expansion Tier 2 cities, Retail Engine, Smart Scheduling v2, e-commerce integration, credit scoring pilot with BRI/BCA → 37,500 paid users.

Phase 3 2028–2029: National Coverage + Credit National coverage, credit scoring live, regional language support, data flywheel maturing → 138,000 paid users.

Phase 4 2029–2030: ASEAN Expansion Vietnam, Philippines, Thailand. Local fintech partnerships per geography → 280,000 paid users.

Phase 5 2030–2031: Ecosystem Marketplace, lending, insurance layer. Data monetization. Series B/C readiness → 450,000 paid users.

400 million MSMEs. Zero executive support. Until now.

Built With

Share this project:

Updates