Three layers, one system.

Layer 01

Connect

Pineloop reads every system the business already runs and normalises all of it into one model of your financial position. On-premise, cloud, or a database nobody has opened in a decade.

Nothing is migrated and nothing is replaced. The constraint is access, not architecture, which is why your IT head belongs in the first conversation rather than the ninth.

• Systems of record — ERPs of any generation, hosted or on-premise.
• Operations — warehouse and transport management.
• Money — banking, statements, remittance advice.
• Channels — marketplace settlement files.
• The rest of it — spreadsheets, and documents attached to email.

Layer 02

Run

Processes run continuously, on schedule, or on trigger. Most AI in finance suggests, drafts, and summarises, and someone on your team still has to act. Pineloop completes the process. What you feel is that the task is gone, not that it is easier.

What it learns is how your company actually operates, including the parts nobody wrote down: which customer always short-pays freight, which cost centre a site's consumables belong to, which docket stands in for a missing goods receipt note. It needs less input the longer it runs.

Layer 03

Control

Rules are proposed from how the work is done today, reviewed by your controller, and only then put into effect. Nothing posts without one, and anything posted under one can be reversed.

This is the layer your auditor cares about, and the one your IT head will test first.

• Thresholds — you set the value and conditions above which a person decides.
• Routing — exceptions go to the people you name, in finance or outside it.
• Reversibility — every action can be undone.
• Trail — ledger entry, rule applied, approver, source document.
• Your data — never used to train models.