Build a finance ontology one workflow at a time. Pick a workflow that matters, list the 20 or so concepts its figures depend on, define each in plain language, write down how they relate, map them to your systems and name an owner for each. Then extend it with the next workflow instead of modeling the whole enterprise first.
Ontology projects stall when they try to describe everything before anything is useful.
1. Modeling the enterprise first. Months of design before anyone can use it.
2. Starting with the tool. Picking a graph database before agreeing what a customer is.
3. No owners. Definitions drift as soon as nobody is accountable for them.
4. Technical names. Concepts named after tables instead of how the business talks.
Preql's agents scan your systems to find how each one represents the same business concept and propose shared definitions. Business owners approve them, and every change is versioned with an audit trail and lineage, so each workflow builds on the last.
How long does the first workflow take?
It depends mostly on agreement, not modeling. Naming owners up front keeps it short.

