The problem
Most finance teams produce budget versus actual commentary every month, yet the process is rarely as clean as it should be. Actuals come from the general ledger, budgets sit in a separate planning tool or spreadsheet, and forecasts are often maintained in yet another model. Analysts spend the first few days of every close pulling exports, aligning cost centre codes, rebuilding variance tables and chasing budget holders for explanations.
The commentary itself is often inconsistent. Different analysts describe similar variances in different ways, materiality thresholds are applied loosely, and the same recurring drivers appear month after month with slightly different wording. By the time the board pack is ready, the numbers are accurate but the narrative has absorbed a disproportionate amount of senior finance time.
Why it matters
Variance commentary is the bridge between the numbers and the decisions. If it is late, inconsistent or unclear, leadership loses confidence in the management information and starts asking questions that should already be answered in the pack. That delays decisions on cost control, hiring, pricing and investment.
There is also a control angle. When commentary is produced manually, there is little audit trail of who said what, which thresholds were applied, or whether the same explanations are being reused without challenge. For listed groups, regulated entities and PE-backed businesses, this matters for both internal governance and external reporting.
The opportunity
The preparation of variance commentary is highly repeatable. The data sources are known, the materiality rules can be codified, and the structure of the commentary itself follows a predictable pattern. This makes it an ideal candidate for no-code automation combined with embedded AI.
Automation can handle the data joins, the variance calculations and the threshold checks. Embedded AI can then draft first-cut commentary in a consistent house style, flag where explanations are missing, and summarise recurring drivers across cost centres. Finance retains full review and sign-off, but starts from a much stronger position.
Example workflow
1. Connect the source data
Pull actuals from the general ledger, budgets from the planning tool or budget workbook, and the latest forecast from the relevant model. Include cost centre, account, entity and period dimensions so the data can be sliced correctly.
2. Standardise and prepare the data
Align chart of accounts mappings, cost centre hierarchies and entity structures. Resolve known naming differences between systems. Convert currencies where required using a controlled rates table.
3. Apply business logic
Calculate variances against budget and forecast at the required levels. Apply materiality thresholds, for example absolute value and percentage tests, to flag the variances that genuinely need commentary.
4. Run checks and controls
Check that actuals tie back to the trial balance, that budgets reconcile to the approved plan, and that no cost centres or accounts are missing. Highlight late postings, journals after cut-off, and any unmapped items.
5. Produce outputs
Generate a structured variance pack with tables by cost centre, function and entity. Use embedded AI to draft first-cut commentary against each material variance, drawing on prior month explanations and known drivers.
6. Review exceptions
Finance business partners review the drafted commentary, amend where needed, and confirm explanations with budget holders only where the AI draft is insufficient or the driver has changed. All edits are tracked.
7. Move to governed operation
Lock the workflow down with version control, access controls and a clear owner. Schedule the run to align with the close calendar, and retain the full audit trail of inputs, thresholds, drafts and final commentary.
What good looks like
- A single, repeatable process that runs to the close timetable rather than to analyst availability.
- Materiality thresholds defined once, applied consistently, and reviewed annually.
- A clear separation between data preparation, AI-drafted commentary and human review.
- Prior period commentary available as context, so recurring drivers are described consistently.
- A full audit trail of source data, calculations, drafts and final sign-off.
- Commentary that focuses on what changed and why, not on restating the numbers.
Benefits
For the finance team
Less time on data wrangling and first-draft writing, more time on genuine analysis and business partnering. Junior analysts get a structured starting point rather than a blank page, and senior reviewers spend their time on judgement rather than formatting.
For leadership
Faster, more consistent commentary in the board and exec packs. Variances are explained in the same language month to month, which makes trends easier to spot and decisions easier to take.
For the wider business
Budget holders are only asked for input where it is genuinely needed, rather than chased every month for the same recurring items. This improves the relationship between finance and the business.
Where to start
A good first version focuses on one entity, one P&L view and a clear set of materiality thresholds. Prove the data joins, the variance calculations and the AI drafting on a single month, then run it in parallel with the existing process for one or two cycles. Once the team trusts the output, extend it across entities, add balance sheet and cash variances, and bring forecast comparisons into the same workflow.
How 4th Revolution can help
4th Revolution is finance-led. We combine data engineering, no-code automation and embedded AI to build governed workflows that finance teams actually trust. We understand close timetables, materiality, controls and the realities of working with imperfect source data.
Our goal is not just to build a workflow that drafts variance commentary. It is to leave you with a governed, repeatable process, owned by finance, with clear controls, version history and a sensible role for AI inside it.
Example outcome
Before: a team of analysts spends the first week of every close pulling exports, rebuilding variance tables and writing commentary from scratch. The board pack lands late, and recurring variances are re-explained every month in slightly different language.
After: the variance pack is generated automatically once the ledger is closed. AI drafts first-cut commentary against material variances, drawing on prior month context. Finance business partners review, refine and sign off. The pack is consistent, the audit trail is complete, and senior finance time shifts from preparation to analysis.