### Summary Good bet, if: - Vertical AI + WhatsApp-native UI is the right wedge to digitize fragmented SEA B2B markets, and relevant in expansion. - The founding team can scale SaaS and PaaS on top of an initially operations heavy D2B wedge. - We can stomach some early working capital strain for long-term infra/SaaS upside. This is not a pure SaaS profile with recurring high-margin revenue at this stage yet. ### The Core Proposition Farmio is building an **AI-powered food supply chain platform** across three models: 1. **SaaS** – WhatsApp-native agentic software for food suppliers. 2. **PaaS** – Marketplace charging take-rate on sales. 3. **D2B** – Importer/distributor of eggs and oil under its own brand. The real bet is on **AI-native automation of messy B2B workflows** in Southeast Asia’s food supply chains. ### Why It’s Interesting #### 1. **Clear Product-Market Fit in D2B** - ~$500k/month gross revenue (as of Apr ‘25) with just 5 FTEs is unusually lean and productive. - EBITDA positive ($60k in Apr), rare for pre-Series A in this category. - Core wedge in eggs and oil imports shows they can win trust, deliver volume, and handle ops-heavy fulfillment. #### 2. **Team Is Strong** - CEO built and exited Pickupp (last-mile delivery, $53M raised, 143% IRR for seed investors). - CTO and COO come from credible tech/logistics backgrounds (Xendit, Meta, Uber). - Founders retain ~70% of cap table – high ownership, high skin in the game. #### 3. **AI-Native from Day 1** - They’re not tacking on LLMs. They’re building modular, task-specific agents (e.g. invoice automation, credit limits, reconciliation). - These are hard to replicate once the data loops kick in. The long-term moat could be real. #### 4. **Valuation Is Reasonable** - Raising $2M at $16M post (with 20% discount baked in). - 1.6x FY25 run-rate revenue multiple is attractive for a vertical SaaS/infra bet in this space. - MOIC modeled at 10x, IRR 27.5% – in line with good early-stage targets. #### 5. **Market Timing** - SEA is moving toward digitized procurement, but most SMBs still run ops via WhatsApp or pen/paper. - Farmio fits right into this gap with a lightweight, chat-native interface and “infra + ops” automation. ### What Needs Watching #### 1. **Overweight on D2B** - Most traction and margin (~30%) comes from buying/selling eggs and oil. This is capital and ops-intensive. - SaaS and PaaS are still early in monetization – unclear if these will scale without heavy onboarding/support effort. #### 2. **Working Capital Risk** - D2B involves holding inventory and depending on third-party logistics. - Reported deterioration in DSO/DPO/DIO in 2024-25 – they’ll need tight cash ops as volume grows. #### 3. **AI Delivery Risk** - Need more clarity on which agents are _live_ vs roadmap. - “AI-first” pitch is strong, but value will depend on adoption, integration depth, and actual performance of the agents. #### 4. **Regulatory + Ops Complexity** - Operating in food import means dealing with high-reg oversight (especially across 4 countries by 2027). - Expansion to UAE adds complexity even if the opportunity is real. ### Questions 1. What is the [[AI Agents Stack]] and models behind the "[[AI agents]]"? Ref: [[AI Agents Landscape - Snapshot of progress and potential]] | [[Autonomous AI Agents - The rise, potential and challenges]] 2. What is the level of customisation / extent of prompting that someone would need to go through to build similar of these agents? i.e. will a feature upgrade make it so easy for customers/users to build this with their go-to LLM or not? 3. How do they compare to existing companies that can be competitive in the "point solutions" or similar agents in the target market? E.g.: 1. Sales (KYC) Agent Automation: [Advance Group](https://www.advancegroup.com/) (Singapore) and [Wise AI](https://wiseai.tech/) (Kuala Lumpur), [Jumio](https://go.jumio.com/brand-ppc-24?gad_campaignid=836682152&gbraid=0AAAAADtrUaNKXz_cwCHIhm4hvvj87LiKp) (USA) 2. Sales Operation Agents (Order & Invoice Processing): [Staple](https://www.staple.ai/solutions/purchase-order-processing#:~:text=Classify%20documents)(Singapore), [Cinnamon AI](https://cinnamon.ai/en/)(Tokyo), 3. Finance Agents (Payment Reminders & Debt Collection): [Flow](https://flow-tech.ai/#:~:text=)(Singapore), [Lunash](https://www.lunash.id/id#:~:text=With%20minimal%20disruption%20to%20your,while%20we%20recover%20your%20payments)(Indonesia), [AI Rudder](https://airudder.com/about-us/#:~:text=Our%20team%20has%2010%2B%20years,ai%20rudder%20website) (Singapore) 4. Logistics Agents (Shipment Tracking & Coordination): [DiMuto](https://dimuto.io/#:~:text=The%20agriculture%20and%20food%20industries,future%20for%20food%20supply%20chains) (Singapore) [GoComet](https://www.gocomet.com/) (Singapore) 5. Customer Service Agents (AI Chatbots & Talkbots): [WizAI](https://www.wiz.ai/wiz-ai-devises-ai-talkbots-for-southeast-asia-startup-stories/#:~:text=), [AI chat](https://www.aichat.com/about-us#:~:text=Innovating%20Conversational%20AI%20for%20Business,and%20drive%20results%20since%202016)and [YellowAI](https://yellow.ai/) Ref: [[Farm to Table Value Chain]] | [[Future of Food MOC]]