Case studies · Real estate: land lot projects

From Excel to a custom platform: managing 13 land lot projects, built with AI

DataLab used AI to build the custom platform a confidential land lot developer in Tolima, Colombia, uses to replace the Excel templates behind its 13 projects. Lot inventory, deals, receipts, cash, receivables with automatic late-payment alerts and a portal for each buyer now live in one place, in the cloud, with role-based access.

Subdivided land at sunset with wooden stakes and a tablet on a field table: land lot management platform that replaces Excel case study — **Confidential client** · Colombia

Client: Confidential client
Industry: Real estate: land lot projects
Country: Colombia
Period: 2026

AI MVP development

Context

Our client is a land lot developer in Tolima, Colombia, whose name we keep confidential. It runs 13 projects at once, each with its own lot inventory, buyers, cash and receivables. That operation lived in Excel templates the team filled in by hand.

The challenge

With 13 projects in Excel, the same data gets typed in several times, and one broken formula can throw off cash or receivables without anyone noticing. Knowing how much a buyer has paid, which installments are coming due or how much is in each account meant rebuilding the information by hand. On top of that, management needed to see the whole business, while each project assistant should only see their own project and each buyer, only their own lot.

The hypothesis

If every payment, contract, receipt and commission is tied to the lot, the team enters data once and the platform builds cash, receivables and KPIs on its own. And if it's built with AI and grows in packages, the first version arrives in weeks and new features are added when the operation needs them.

What we did

A clickable prototype before building, with the lot inventory, a lot record, cash and the KPI dashboard, to ground the vision of the platform.
A full process review with the client's team and a signed scope agreement before development.
Lot inventory by project (block, number, area and price), importable from the existing Excel files, plus cash or financed deals.
Each customer's data entered only once, and cash receipts with sequential numbering for every payment.
Cash in and cash out with categories, and sub-accounts by source: Bancolombia, Nequi and cash.
Receivables and payment plans with automatic late-payment email alerts, and a portal where each buyer sees their lot, payment plan and receipts.
Proof of payment uploaded by the buyer and approved by the team before it hits the cash account.
Roles (management sees everything, each assistant their project, each buyer their lot), a dashboard with monthly sales, cash and commissions, and Excel export.

The readout: results

We don't publish figures for this case; the readout is qualitative. The platform replaces the Excel templates behind all 13 projects: every lot has its full history of payments, receipts, sales agent, commission and status, receivables send their own alerts when an installment is coming due or overdue, and each person sees only what they should. The first version was planned with a maximum of 50 business days, and from there the platform grows in packages that are quoted and approved before work starts.

Stack and tools

Custom platform built with AI
Clickable prototype
Lot inventory
Cash and sub-accounts by source
Receivables with late-payment alerts
Buyer portal
Roles and permissions
Excel export

What we learned

Enter data once. When every payment, receipt and commission is tied to the lot, cash, receivables and KPIs stop being reports someone has to put together: they come straight out of the operation. That's why a specific problem, like a shared spreadsheet the business has outgrown, is better solved with a small, precise platform than with a big, generic system.

Case studies