Finance Intelligence
Illustrative interfaceTurn Governed Financial Data into Executive Intelligence
Finance Intelligence combines approved budgets, actuals, governed variance calculations, materiality rules, and AI-ready business context to help finance teams explain performance without allowing AI to redefine the numbers.
Illustrative interface
The margin bridge on this page is a synthetic worked example. It answers a different question from the executed variance proof.
Open the Finance proof explorer. It shows the published synthetic results. It is not a live warehouse query.
The decision
Which variances are material, which departments need management attention, and what is driving the change from the prior period.
For CFO, FP&A, Controllership, and Finance Transformation leadership.
Revenue increased. Why did gross profit decline?
Synthetic single-product example. Separate from the executed variance proof; not a customer deployment.
| Amount | S-01 | S-02 | Movement |
|---|---|---|---|
| Revenue | 10,000 | 11,440 | +1,440 |
| Cost | 6,000 | 7,920 | +1,920 |
| Gross profit | 4,000 | 3,520 | −480 |
Gross-profit bridge
- Volume+400
- Price+440
- Cost rate−1,320
Contribution scale. Full width is 1,320 synthetic monetary units. Opening and closing amounts are levels, not bars.
4,000 + 400 + 440 − 1,320 = 3,520
Reconciled · residual 0
Supporting evidence
Linked to the cost-rate contribution.
Purchase record PR-S
Supplies the unit-cost input, 30 then 36, and stays linked to the −1,320 cost-rate line. The link does not confirm why the cost changed.
Approved adjustments
None applied. Contribution 0.
Business explanation BE-S
Context only: a supplier list price increased before S-02. It does not change the formula, and it is not a confirmed cause.
Result and evidence inform the proposed interpretation
Proposed interpretation
Gross profit fell by 480 synthetic monetary units while revenue rose by 1,440. The cost-rate contribution (−1,320) exceeded volume (+400) and price (+440). Purchase record PR-S supplies the unit cost and stays linked to that line. Explanation BE-S is context still to be assessed. This proposal does not change the calculated amounts.
Finance review
No decision is recorded. A reconciled bridge does not select one.
- Confirm the explanationNot recorded
- Correct the explanationNot recorded
- Request further investigationNot recorded
Calculation method
One sequential method: volume at the opening unit margin, then price, then cost rate. There is no product-mix contribution. The three lines explain gross-profit movement, not revenue movement. This is one specified attribution method, not a unique causal explanation.
S-01: 200 physical units, selling price 50, unit cost 30. S-02: 220 physical units, selling price 52, unit cost 36. Prices and costs are synthetic monetary units. Quantities are physical units.
- Volume. (220 − 200) × (50 − 30) = +400
- Price. 220 × (52 − 50) = +440
- Cost rate. −220 × (36 − 30) = −1,320
Volume changes revenue by +1,000 and cost by +600, which is the +400 gross-profit effect. Residual: 4,000 + 400 + 440 − 1,320 − 3,520 = 0. Tolerance is 1 synthetic monetary unit.
Preparation
The example is already aligned to S-01 and S-02, one entity, one currency of synthetic monetary units, and one product identifier. Accepted source records can supply inputs. Narrative context does not redefine the formula.
It is not an executed warehouse, an Alteryx workflow, or a deployed Orisdale system. Illustrated host: Snowflake. Databricks and BigQueryare alternative hosts for the same calculation, not extra layers. The host follows the customer's selected platform and configuration.
What happens if reconciliation fails?
This example does not take the failure path. If the residual exceeds the tolerance, stop at calculation exception review, correct the calculation or the input, and reconcile again before an explanation proceeds.
Illustrative failure, kept out of the bridge: residual 25 against a tolerance of 1.
Sample executive questions
Example executive questions
- What are the largest material variances?
- Which departments require management attention?
- What changed from the previous period?
- What are the primary unfavorable drivers?
- What should the CFO review first?
Related notes
Interpretation boundary
Why finance AI should explain governed variance fields, which controls to require, and what a matching number still does not prove.
Executed synthetic calculation example
A worked variance example computed locally from published synthetic inputs, with the contract, results, claim checks, and validation record available to download. It was not run in Alteryx or BigQuery.
Finance Intelligence reference architecture
The governed finance-to-decision architecture used for this solution.
Capabilities
- Budget-versus-actual intelligence
- Material variance detection
- Management-attention classification
- Executive commentary
- Department and category analysis
- Conversational finance questions
Governed logic
- Actual amount and budget amount
- Variance amount and variance percentage
- Amount and percentage threshold rules
- Material flag and variance category
- Management-attention classification
Reference architecture
- 01Actuals and Approved Budgets
- 02Governed Finance Logic
- 03Semantic Finance Intelligence
- 04Enterprise Data PlatformFoundation — hosts data, compute, and access control for every stage above
- 05CoreAgentic AI Runtime
- 06Executive Finance Experience
Trust and control
Deterministic logic
Calculations, thresholds, and classifications are defined, reviewed, and controlled before any AI interpretation.
AI interpretation
The AI runtime explains and contextualizes governed outputs within approved tools and permissions.
Human review
Consequential decisions retain human approval, with monitoring for quality, access, cost, and behavior.
- The AI layer interprets governed finance outputs.
- It does not recalculate variance, redefine materiality, invent a KPI, override the source of truth, or create unsupported financial values.
Reference deployment patterns
01
Enterprise sources
Budgets, actuals, and operational records start in the systems that own them.
02
BigQuery source tables
Source data is ingested and staged inside the BigQuery data foundation.
03
Governed preparation
Alteryx One defines the preparation and the approved business rules. Supported operations run as BigQuery queries.
04
Published governed results
Governed tables or views and their semantic context are published for permitted access.
05
Runtime interpretation
The runtime reads approved results and context through approved tools. The business experience presents the explanation.
06
Pending human review
A person reviews consequential findings. Rule changes go to the rule owner. Nothing is approved automatically.
Snowflake
Orisdale designs governed business logic and semantic intelligence for Snowflake-based Enterprise Intelligence Systems.
Databricks
Governed logic, semantic preparation, and Unity Catalog governance support Orisdale solution patterns on Databricks.
Google Cloud
Alteryx One prepares governed logic, BigQuery holds the analytical outputs, and Gemini Enterprise is the proposed business-facing agent experience.
Alteryx One
Alteryx One serves as a governed data preparation, analytics, workflow, and business-logic layer that prepares trusted, AI-ready datasets for enterprise platforms.