Skip to content
looma

AI products · Internal AI

AI apps people use every day. The same patterns, built for companies.

Looma is an early-stage startup in Indonesia. We run three public AI apps (Snap Money, Flex, and Kala), then bring the patterns proven there into companies' internal systems.

For companies: two kinds of work, both finished.

Written scope: example

Built

  • Document assistant over 240 SOPs, every answer carrying its source
  • Search in both Indonesian and English
  • A log of questions that went unanswered, reviewed monthly
  • Two hours of training, documentation, and full access to the code

Not built

  • HRIS integration: deferred, not dropped
  • A native mobile app
  • Migrating archives from before 2019
  • Support outside working hours
The right column is written at the start. Projects sprawl because that column never existed, not because the left one was too short.

Four stages, and the second is the most decisive.

We start from the problem, not the technology: which rules are scattered, which procedures are unread, which repeated work is still manual. This mapping is free.

  1. 01

    Map the problem

    We start from the problem, not the technology: which rules are scattered, which procedures are unread, which repeated work is still manual. This mapping is free.

  2. 02

    Written scope

    The agreement is written before the code is: what will be built and, just as important, what will deliberately not be built.

  3. 03

    Build with regular review

    We build directly and show progress regularly, so the project does not drift off course.

  4. 04

    Full handover

    The system is handed over with documentation, brief training, and full access to the code. A maintenance retainer is offered, never required.

Limits we state up front.

Discuss your needs

Tell us the problem. If it is not our field, we will say so.

The first reply contains questions, not a sales pitch. If your problem does not fit the work we do, we will tell you.

Discuss your needs