Build or buy a finance context layer

Building a finance context layer in-house means more than a metric catalog. You need mapping across your source systems, an approval workflow finance will use, version history, lineage, access controls and connections to AI tools, plus a team to maintain them. Building makes sense when you have that team and unusual needs. Otherwise, buying can get finance to governed definitions sooner, at the cost of a subscription, a vendor review and some dependence on the vendor.

In-house projects often start as a JSON file or spreadsheet catalog. The hard part is keeping it current as definitions and systems change.

What an in-house build requires

1. Connectors to the core, GL, ERP, warehouse and spreadsheets.
2. A way to map each source's version of a figure.
3. An approval workflow non-engineers can use.
4. Version history and change records.
5. Lineage from report to source.
6. Access controls that carry through to AI tools.
7. Engineers to maintain all of it as systems change.

When building makes sense

1. You have a dedicated data platform team with capacity.
2. Your definitions mostly live in one warehouse.
3. Finance is comfortable working in engineering tools.

When buying makes sense

1. Finance needs to own definitions directly.
2. Data spans core banking, GL and spreadsheets.
3. You need results before a platform build could deliver them.

What to weigh either way

1. Total cost over three years, including people.
2. Time to the first governed workflow.
3. Vendor review and security requirements.
4. How easily you could export your definitions if you change course.

How Preql helps

Preql's agents propose mappings across your systems, business owners approve them, and every change is versioned and auditable. Lineage traces numbers back to source systems, and Preql inherits your existing access controls.

FAQ

Can we keep our existing catalog?

Yes. An existing catalog is a useful starting point for mapping, whichever route you take.

What does maintenance look like?

Owners approve changes. With Preql, agents monitor source systems and flag changes that need attention.