B3bitbybit · venture board ← All ventures

Venture Research

The estate floor still runs on paper.

We assessed whether B3 should build offline-first field-operations software for Sri Lanka's tea, rubber, and coconut estates. This is the market, the paperwork, and the incumbents.

Verdict
Conditional.
Validate before building.
Market
Sri Lanka
Prepared for
Methika, Thanuka
Date
2026.07.25
Evidence
25 sources, verified

The call

The problem is real. The field is not empty. Sri Lanka's plantation sector is large, structurally loss-making, and dominated by a rising labour bill. That makes payroll accuracy, labour-cost visibility, and audit-ready compliance a bottom-line lever, not a convenience. The idea holds.

The greenfield assumption does not. At least one Sri Lankan company already ships mobile-first, offline-capable estate software and is winning government deployments. Do not build yet. Spend two weeks on competitor due-diligence and one real estate conversation. Build only against a wedge the incumbents miss.

The ground

A large sector on thin margins with a rising wage bill.

The pain gets a budget because labour is the cost, and the cost is climbing.

405RPC tea estatesPlus 27 state estates. A small, concentrated B2B buyer set.
~½of RPCs loss-makingMedian net margin near 3–4% across CY2015–20.
Rs1,700mandated daily wageUp from Rs500 in 2019. Adds over Rs35bn a year in cost.
2.5Mworkforce supportedTea alone. Sri Lanka's largest employer.
45%have no internet96% have mobile. Offline-first is mandatory.

What we tested

Four hypotheses, scored against the evidence.

H1

Estates still run on paper

Field records are handwritten and often discarded once wages are paid. The data exists, then evaporates.

Supported
H2

The value is payroll and labour cost, not less paper

In a sector this margin-stressed, muster-to-wages accuracy and labour visibility move the P&L. Compliance is a second mandated workload.

Strongly supported
H3

Incumbents are heavy and head-office only

Wrong. A local player already ships mobile-first, offline field software and is deploying at government estates. The gap is smaller and contested.

Partly wrong
H4

Offline-first is our technical wedge

Mandatory, yes. But the government and the incumbents already do it. It is table stakes, not a moat.

Table stakes

The finding that changes the plan

The assumption that broke.

Read this first

Competitor coverage was the thinnest part of the research, and it still surfaced this. Sri Lankan companies are already in the on-estate field layer. Treat these as leads to verify with demos, not settled fact.

PlayerWhat it doesField, offlineTraction signal
Agrithmics
Agrigen, rebranded Cultive8

Integrated estate-management ERP. IoT and agri-fintech. Procurement and payment disbursement.

 Mobile-first, offline

Deployed at Goomera, the first government tea estate. SLSPC rollout to all estates reportedly planned.
iTechro
iHarvest, with SLT-Mobitel

Cloud platform that digitises field activity and operations for enterprise plantations.

 Field layer

Telco-backed plantation-ecosystem partnership.
Blue Lotus 360

General ERP with a tea-plantation module.

 Head office

Marketed SL tea-plantation ERP. Field depth unclear.
Cropin, Farmforce, SAP Agri
named globals

Enterprise agri-ops platforms.

 Unknown in SL

No evidence found active in SL estates. Assume any could enter.

Read the terrain

Where a wedge might still be, and what could kill it.

Edges B3 could contest

  • Payroll-accuracy guarantee. Own muster-to-wages leakage, with a hard ROI number per estate.
  • Compliance in a box. Rainforest Alliance turns tacit field practice into mandated records with real deadlines.
  • Price for the strugglers. Enterprise ERP will not cheaply serve small, loss-making RPCs. A lean per-estate price can.
  • Socio-technical fit. Low-literacy Tamil and Sinhala field UX is where incumbents are likely weakest.

Risks that could kill it

  • Entrenched local incumbent already mobile-first, offline, and winning government estates.
  • Tight budgets. Half the buyers lose money. Spend must be cheap and the ROI obvious.
  • Adoption is not signup. Evidence shows high sign-up masks short-lived use. Fit decides survival.
  • Long, relationship-driven sales with RPC head offices. Slow and semi-political, and unquantified.
  • Union sensitivity to digital worker surveillance, such as attendance and biometrics.

Before a single line of code

The two-week validation gate.

Both build-gating unknowns resolve with about five conversations and two product demos. Cheap to settle before committing engineering.

Run competitor due-diligence on Agrithmics and iTechro

Get demos and pricing. Which estates are live. What field staff dislike. Whether they nail muster-to-wages payroll. Whether they are beatable.

Gate · Thanuka

Run one discovery interview with a real estate contact

Who holds the budget and how long is the sales cycle. How muster-to-wages runs today, and what inaccuracy costs per estate. Listen, do not pitch. Use Methika's network.

Gate · Methika

Define the sharp wedge

Pick one. Payroll-accuracy guarantee, compliance in a box, or radically cheaper per estate. Less paper is a claim an incumbent already makes.

Then · Both

Decide go or no-go

Build only if due-diligence finds a defensible gap. Otherwise park it. The research folder stays warm.

Decision