§ Demo · Finance · Narrated by Lyra

Month-end reporting — cycle from days to hours.

A retrieval-assisted drafting surface for the finance team — variance commentary generated with citations to the underlying ledger, a drafting UI the CFO can edit, and a consistent narrative structure month over month.

Month-end reporting — cycle from days to hours.

Narrated by Lyra · 3:45

The demo runs on a synthetic mid-market P&L, but the pattern is exactly what we shipped for a UK insurance broker last year. The most valuable outcome wasn’t the time saved — it was the CFO getting hours of her month back and a narrative shape she could actually defend to the board.

The eval spine matters here. Numerical claims are checked deterministically against the ledger before drafting; the model produces the language, not the facts.

The pattern

What the walkthrough shows.

  1. Draw the numbers cleanly.

    Structured pulls from the ledger, KPI store, and prior-month comparators — into a canonical dataset the model reasons over. Nothing invented; every number traceable.

  2. Generate the variance commentary.

    For each variance above the materiality threshold, a first-pass commentary drafted with an inline citation to the driving lines. The model doesn't invent explanations — it surfaces them for the finance team to confirm.

  3. Assemble the narrative.

    Executive summary, KPI headlines, top-of-mind risks — assembled to a consistent shape month over month, so the CFO can compare trends without decoding a new deck each time.

  4. Edit surface for the humans.

    Every drafted section is editable, with the source data one click away. The CFO keeps her voice; the machine keeps the grunt work.

Order-of-magnitude impact

What clients typically see.

4–6 hrs

Cycle time (from ~3 days)

~£18k

Typical annual saving

189%

Year-2 ROI

Want this shape, for real, in your business?

A 30-minute discovery call. We'll say honestly whether the pattern fits.