Company

Harrimont builds the systems that operations run on.

An AI, data and systems engineering company. We build Ancilor, our platform for enterprise data and operations, with the people who use it.

Based in Kosovo. Working internationally.

Film: licensed stock footage.

What we do

Applications, data and AI, built by one team.

We design the screens people work in, the connections behind them and the AI that reads them, so every part answers to the same workflow.

Systems engineering

Applications for the people who do the work: shift reports, deliveries, inspections, approvals and pay inputs. Each person sees their own site and their own work. Pay and personal details are visible only to the roles that manage them, and actions are logged with who made them and when.

Data engineering

Connections that read from the systems a business already runs, such as accounting, pump controllers, bank files and spreadsheets, without changing them. Each connection is agreed with the people who run that system and is designed to show where a figure came from and how current it is.

Applied AI

Models that help people review and summarise inside the application, designed to show the evidence behind each suggestion. AI does not change records, and a person makes the decision.

How we use AI

How we deliver

One workflow first, then the next.

We start small, work beside the people who will use the system and stay with it once it is in use.

  1. Start with one workflow

    The record your team chases most: a shift close, a delivery, an inspection. We map how it runs today with the people who run it.

  2. Build with the people who use it

    The people who will work in the screens try them on their own work, while there is still time to change them.

  3. Test against existing records

    Before your team relies on the workflow, we run it against past records whose outcome is already known and explain any differences.

  4. Operate and support

    Releases go out in agreed windows. We stay with the people who use the system for questions, fixes and the next workflow.

Operational understanding

Built from real operations.

Our first systems were built alongside a working operation: its sites, shifts, deliveries and payroll month. The habits below come from that work.

The record starts where the work happens.

Shift reports, inspections and delivery updates are entered by the people closest to the work: the site manager, the inspector and the forwarder. Inspections work on a phone without signal and sync later.

Differences stay visible.

A mismatch keeps its owner and its reason until someone resolves it. It is not overwritten to make a total agree.

Two people on the money.

Shift reports are accepted by a reviewer who did not write them, and deposits are confirmed against the bank by finance.

Closing a month is a deliberate step.

A month closes only when every shift report in it has been accepted. Reopening an accepted report takes an administrator and a written reason, and the action is logged with who took it and when.

Where we work

On site when the work needs it.

We come to the site to see a workflow run and to try it with the people who use it. Between visits, we work remotely.

Company
Harrimont L.L.C.
Office
Fushë-Kosovë 12000, Kosovo
Email
office@harrimont.com
Phone
+383 38 606 766

Research

Open reports on how AI should work.

Technical reports by Argjend Bajraktari, published on Zenodo under CC BY 4.0. They propose architectures; they are not peer reviewed and report no product performance.

Contact

Tell us about the workflow.

Start with one process, its records and the decision your team needs to make. Write to office@harrimont.com or use the contact form.