Variance & Flux Analysis

Variance & Flux Analysis ("Upload GL data, flag material variances against a threshold, and generate an AI-assisted flux memo.") takes your general-ledger data, flags the movements that matter, explains them with AI, and writes a memo you can hand to a reviewer. Open it at /dashboard/analytics/variance.

Create an analysis

From the list page, click New analysis and fill in the dialog:

• Name — for example, "Q3 vs Q4 OpEx Flux."
• Client — choose a client, or leave it as No client.
• Data layout — choose how your periods are arranged:
  • Dual files (recommended) — a Base Period file and a Comparison Period file uploaded separately.
  • Single dataset — one file that already contains both periods.

The list shows each analysis with its client, type, a Flagged / Total count, status, and last-updated date. Use the All clients filter to narrow the list, or open Reports for a cross-analysis rollup.

The editor: five steps

Opening an analysis walks you through five steps, shown across the top: Upload GL → Map columns → Thresholds → Review → Results. You can revisit any completed step.

Step 1 — Upload GL

Drag in your file(s) or click Browse Files (CSV or Excel). In dual mode you provide two files and assign each a role (Base Period / Comparison Period). No data handy? Click Use Demo Data for a realistic sample, or Download Template for a starter file.

Step 2 — Map columns

The app auto-maps your columns to the fields it needs and shows you the result so you can correct anything. The key fields are Account Name/Number, Amount, Description/Memo, and Period/Date (single-dataset mode). You can also map an optional Class/Department and add your own dimensions for grouping. Unmatched columns are marked (unmapped) — set them from the dropdown.

Step 3 — Thresholds

Set the materiality thresholds that decide what gets flagged:

• Dollar threshold ($) — the minimum absolute change to flag.
• Percent threshold (%) — the minimum relative change to flag.

Not sure where to set them? Click Suggest and the AI will sample your GL and propose thresholds with a rationale; click to apply the suggestion. You also choose the Account type and any Analysis anchors (the dimensions, such as account or department, that variances are grouped by).

Step 4 — Review

Review your configuration before running. The step previews how many rows would be flagged at the current thresholds, then click Run analysis. Running moves a Draft analysis to In Review.

Step 5 — Results

Results are organized into three tabs:

Tab	What you see
Table	Every flagged row with its amounts, the variance, and its explanation/status.
Charts	A visual breakdown of where the variance is concentrated.
Memo	The AI-generated flux memo.

Click a flagged row to open its detail panel, where you can read the AI's explanation and suggested follow-up, accept or reject it, or write or refine the explanation yourself. Use Analyze to have the AI explain the flagged rows in bulk. Export the rows with Export rows (CSV/Excel).

On the Memo tab, the AI drafts a formal variance/flux memo from your flagged items. Use Export to download it as Word or PDF for your workpapers.

Status and reports

An analysis moves through Draft → In Review → Approved → Finalized. The Reports view (from the list page) rolls flagged variances and review status across analyses, with its own exports (Export rollup, Export all flagged).

Tip: Stuck on a number? Open the AI Assistant from any variance screen and ask about a specific account — it can see the data you're looking at.