Scope Watch is for the partner and matter manager running a budgeted transaction. Their main financial risk is usually not the size of the fee estimate. It is the work that drifts outside it unnoticed until the invoice is drafted, when the only options are a difficult conversation with the client or a write-off. The demo matter is a fictional lender-side construction and term financing of a West Texas wind farm. On deals like that, the lenders routinely ask for work the engagement letter excludes, such as hedging documents, tax credit transfers or a mezzanine intercreditor, and associates bill it to whichever workstream is open.
It works from three things a firm already holds: the engagement terms (agreed workstreams, exclusions, budgets, rates and percentage complete), the client's incoming requests, and the team's time entries. Everything in the demo is invented, including the matter, the people, the 16 requests and the 44 time entries, and I labelled each item by hand so that Jev's answers could be checked against them.
Jev answers only the narrow questions that need judgment. It decides whether each request is in scope, excluded or unclear, and which workstream it belongs to. For each time entry, it decides which workstream the work falls under, whether the narrative is too vague to bill, and whether several tasks have been block-billed together. The code does everything arithmetical: fees from hours and rates, budget burned against work completed, projected cost at completion, and a red, amber or green status for each workstream. Keeping the two apart means the numbers are never the model's to get wrong. In the demo, that split surfaces the collateral package as 96% spent with 45% of the work done, along with five out-of-scope requests that need a fee agreed before anyone starts.