Choose the working set
The analyst selects Revenue Build, Debt Schedule, and Assumptions from Model_v12.xlsx. They add the deal notes from SharePoint and the Capital IQ screen they already ran. Nothing else enters the context.
How it works
ValuClaw starts with one deliverable: the analyst selects the Excel tabs, notes, and approved data sources. A versioned skill and approved team memory shape the output. The draft lands in Word with every number tied to a cell, and the reviewer approves before it saves or sends.
Working set
Model_v12.xlsx · Revenue Build · Debt Schedule · Assumptions
Chosen for this run
Word section
Model commentary for review
$48.2M tied to cell F87. 21% YoY tied to H87. The churn assumption that changed from 3.2% to 4.0% is flagged. The model route is logged. All of it sits beside the draft, not in a chat history.
A human decides what happens next.
Your team stays in control
The first implementation is deliberately narrow: one deliverable, named inputs, an explicit model policy, and a human approval point. That makes it easier to evaluate before anyone approves a broad rollout.
The analyst selects Revenue Build, Debt Schedule, and Assumptions from Model_v12.xlsx. They add the deal notes from SharePoint and the Capital IQ screen they already ran. Nothing else enters the context.
Auto picks from the approved pool, or the team pins GPT, Claude Opus, or a local model behind the firewall. The model used for this run appears beside the draft in the run log.
The draft lands as an editable Word section. The reviewer checks it, accepts the redline on the churn assumption, and approves. Only then does it save to SharePoint or send via Outlook.
A practical first workflow
The aim is to make the senior review faster and more defensible. The analyst spends less time on the mechanical build-up. The reviewer sees where every number came from. Nobody moves to a new portal, and nothing hides how the answer was produced.
Start with one deliverable