Skip to content

Performance analysis

Know what changed. Know why.

Our AI agents investigate material movements across your financial and operational systems, find the drivers behind them, and prepare the explanation for finance review.

Go from variance detected to cause explained without manually tracing it across systems and spreadsheets.

See an agent investigate a varianceTalk to us
Illustration: the FP&A Agent's investigation of the question "Why are Q3 operating expenses $4.2M above budget?" for Q3 of FY26 (July to September 2026; the fiscal year is the calendar year), consolidated, forecast against budget, read from SAP S/4HANA, payroll journals and the FY26 operating plan. Operating expenses are forecast $4.2M, 10.0%, above a $42.0M budget. A waterfall bridges the budget to the $46.2M forecast through headcount +$2.1M, cloud infrastructure +$1.4M, marketing +$0.9M, professional services −$0.5M and other +$0.3M. The driver table gives each its current variance and, set apart as forward-looking, its Q4 impact and type: headcount +$1.8M and cloud +$1.1M recurring, the annual conference $0 and one-off, professional services −$0.3M recurring, other +$0.3M recurring, $2.9M in all. The agent's analysis: headcount and cloud explain 83% of the variance, and about $2.9M is recurring. The investigation is ready for review, with actions to view it or add it to the management report.

The work behind the variance

Seeing the variance is the easy part.

Your ERP, planning software and BI tools can show that actuals differ from plan. The work begins when finance has to explain why.

Illustration: the work of explaining a variance, two ways. By hand, finance exports account detail from SAP S/4HANA, opens the operating plan, checks payroll journals, looks through Salesforce, searches the transactions in an export, builds the bridge in a workbook and writes the commentary in a document: eight steps, several hours or days. With the agent, the variance is detected, the agent investigates it across the same systems, returns the drivers with their evidence and an explanation, and Megan Carter, FP&A Manager, reviews it.

From movement to explanation.

Five steps of one investigation, Q3 cloud infrastructure. Each leaves the output the next one starts from.

Illustration: the five steps of one investigation into Q3 cloud infrastructure, each with its output. Detect what matters: cloud infrastructure, +$1.4M, 21.8% over a $6.42M Q3 budget. Drill into the records: 6120 Cloud hosting, 31 invoices from 4 vendors. Find the drivers: the AWS commitment, amended on 1 August, with usage above it. Determine what persists: recurring, +$1.1M in Q4. Prepare the explanation: three sentences and four figures, each linked to its records.

Common investigations

The questions FP&A gets every month.

Six kinds of investigation, each starting from a movement and ending on its causes. Open one to see what the agent does.

  • Budget vs actualWhy are operating expenses above budget?Headcount and cloud explain 83% of +$4.2M
    What our agent does

    Finds the material lines, investigates the transactions and drivers behind them, and explains the movement.

    Typically reads SAP S/4HANA, Payroll journals, and FY26 Operating Plan.xlsx.

    Q3 opex +$4.2M · headcount +$2.1M · cloud +$1.4M · conference +$0.9M

  • Revenue performanceWhy did revenue miss forecast?−$0.3M: a renewal moved to October, and churn
    What our agent does

    Separates timing, churn, price, volume and mix across billing, pipeline and customer activity.

    Typically reads Salesforce, SAP S/4HANA, and Snowflake.

    EMEA subscriptions −$0.3M vs forecast · timing −$0.2M, a renewal moved to October · churn −$0.1M

  • Gross marginWhy did margin decline?−3.2 pts: freight, product mix, discounting
    What our agent does

    Quantifies each driver across revenue, cost of sales, logistics, mix and discounting.

    Typically reads SAP S/4HANA, Snowflake, and Salesforce.

    EMEA 41.6% → 38.4% · freight −1.3 pts · mix −1.1 pts · discounting −0.8 pts

  • HeadcountWhy is personnel expense above plan?+$2.1M: hires ahead of plan, merit six months early
    What our agent does

    Compares hiring dates, vacancies, compensation and contractors with the plan's assumptions.

    Typically reads Payroll journals, Headcount roster, and Headcount Plan.xlsx.

    Q3 headcount +$2.1M · 14 engineers from 1 Jul, budgeted from 1 Sep · merit 6 months early

  • SpendWhere is spending growing unexpectedly?Cloud $3.2M in September, 55% of it AWS
    What our agent does

    Breaks the change down by vendor, department, category and commitment.

    Typically reads SAP S/4HANA and Vendor invoices.

    September cloud $3.2M · AWS $1.76M · Azure $0.80M · Datadog $0.35M · other $0.29M

  • CashWhy is cash different from plan?−$1.4M against plan: collections and payment timing
    What our agent does

    Traces the difference through collections, payments, working capital and timing.

    Typically reads SAP S/4HANA and Cash Plan.xlsx.

    September cash $41.6M vs $43.0M plan · collections −$0.9M · payment timing −$0.5M

Worked example

A 3.2-point margin decline, explained.

EMEA revenue held at $8.4M in September, yet gross margin fell from 41.6% to 38.4%. One question, investigated across three systems and tied back to the books.

Illustration: the FP&A Agent's run on the question "Why did EMEA gross margin fall in September?", six steps: it pulled revenue and cost of sales from SAP S/4HANA, isolated EMEA, examined the freight postings against Snowflake shipments and carrier invoices, compared product mix with August, pulled quarter-end discounts from Salesforce, and tied the result to the trial balance.
Illustration: the answer, EMEA gross margin from 41.6% in August to 38.4% in September, −3.2 points on $8.4M of revenue in both months. Three drivers: freight −1.3 points, product mix −1.1 points and discounting −0.8 points. Freight is open: freight cost rose $109K, from $509K to $618K, because carrier rates on Rotterdam lanes rose in August and reached September invoices; $109K on $8.4M of revenue is 1.3 points, read from SAP S/4HANA, Snowflake shipments and carrier invoices. The result ties to the trial balance, with nothing unexplained, and is ready for review.
EMEA gross marginSeptember 2026 vs AugustReady for review

41.6% → 38.4%−3.2 pts

Revenue $8.4M in both months

Drivers

  • Freight cost increased $109K. Carrier rates on Rotterdam lanes rose in August and reached September invoices.

    5310 Freight out, EMEA · $618K − $509K = $109K · $109K ÷ $8.4M revenue = 1.3 pts

    SourcesSAP S/4HANASnowflakeCarrier invoices
  • Total−3.2 pts

Nothing unexplainedTies to the trial balance

Don't stop at the explanation.

An investigated variance feeds the rest of FP&A: the part that recurs changes the forecast, and the explanation becomes the report's commentary.

Illustration: what follows an investigation. The Q3 variance of +$4.2M is investigated and verified: $2.9M of it recurs and $1.3M is one-off. The recurring part becomes a forecast update, +$2.9M proposed as payroll +$1.8M and cloud infrastructure +$1.1M. In the September Management Pack the commentary is drafted and ready for review.
Review the forecast updatesOpen the September pack

Every explanation opens onto the evidence.

Each statement opens onto the calculation behind it, the account and its postings. Open the highlighted posting to see the record it came from.

Illustration: the freight statement from the worked example traced to its source. Statement: freight cost increased $109K. Calculation: $618K − $509K = $109K, divided by $8.4M of revenue, 1.3 points. Account: 5310 Freight out, EMEA, $618K in September. Transactions: 212 September postings, the largest DHL Express $48,620, Maersk $41,200 and Kuehne+Nagel $36,870. The DHL Express posting is highlighted and opens onto its source record: invoice 88-4410, 26 September, Rotterdam to Hamburg, $48,620, supplier invoice PI-30412 in SAP S/4HANA, with its evidence: read from the SAP S/4HANA general ledger, last synced 30 September 2026 at 08:40, calculated as September less August over EMEA revenue, owned by Megan Carter, FP&A Manager, included in the September Management Pack, from the investigation of 30 September at 08:49.
  1. StatementFreight cost increased $109K.
  2. Calculation$618K − $509K = $109K · $109K ÷ $8.4M revenue = 1.3 pts
  3. Account5310 · Freight out · EMEA · $618K in September
  4. Transactions212 September postings
    • MaerskMLS-31408$41,200
    • Kuehne+NagelKN-55120$36,870
    • 209 more postings

    Select the highlighted posting to open its source record.

Questions

Where our agents start and stop.

  • Do you replace our BI or planning software?

    No. Those tools keep showing and storing financial performance. Our FP&A Agent investigates the movements they surface and prepares the explanation.

  • What data can your agents investigate?

    Financial and operational data in the systems you connect: the ERP and its payroll journals, the CRM and the warehouse, and the plans and spreadsheets your team works from.

  • Can an explanation be traced back to individual transactions?

    Yes. Each investigation keeps the calculation, the accounts, the postings and the source records behind every figure, so any line of the explanation opens onto what it rests on.

  • How do your agents decide which variances to investigate?

    Your finance team sets the materiality thresholds and the dimensions that matter, such as an amount, a percentage, an entity or a department. Our agents investigate what crosses them.

  • Does finance review the explanation?

    Yes. Our FP&A Agent prepares the investigation and the commentary; your team reviews them before they go anywhere.

The productFP&AInvestigate performance, prepare reporting, update forecasts and model decisions across the finance stack you already use.Explore FP&A

Bring us the variance your team spent last week explaining.

We'll show you how the agent investigates it, what evidence it finds and what it would prepare for review.

Book a working session