Which signals feed the forecast?

Revenue side — pipeline (weighted by stage and rep), cohort retention, ramp curves, seasonality, FX rates. Cost side — headcount roster plus open reqs, run-rate vendor spend, project-burst calendars, contracted commitments, inflation indices. Plus any custom signal you wire — usage data, marketing-qualified-lead pace, customer-success health scores.

How is driver attribution done?

When the forecast moves between cycles, the agent compares the inputs that changed. Pipeline shifted by 12 million — that explains 4M of the revenue swing. Hiring accelerated by 8 heads — that explains 200K of the cost line. Each driver named with its dollar contribution, ordered by magnitude. The narrative writes itself.

Can it run multi-scenario branches like best, base, downside?

Yes. Each scenario carries its own assumption set — pipeline conversion, ramp slope, hiring pace, cost inflation. The agent runs all three in parallel each cycle. Side-by-side comparison surfaces what changes between branches. Hand the trio to the board, no Excel reconciliation.

New

Forecast Agent

Re-Models. Detects Drivers. Updates. Surfaces

Monthly rolling re-forecast for revenue and cost using ML on historical data and live business signals — pipeline, hiring, project starts, ramp curves. Surfaces driver-level changes. Replaces the static annual plan with a continuous forecast. ±5% accuracy at the quarter, ±10% at the year.

Activate This Agent View Demo

Forecast Agent. Re-Models. Detects Drivers. Updates. Surfaces.

By the numbers

Forecast Agent by the numbers.

±5%

Quarter accuracy

Monthly

Refresh

Driver

Attribution

<5 min

To activate

The use case

Throw out the annual plan. Live with a forecast that moves.

Annual plan obsolete by Q2. The FP&A team manually rebuilds every quarter — 5 days of model rebuild for a number that lasts 8 weeks. Sales says forecast moves by 20 million. Why? Nobody knows. CFO wants 'how is the year going to land' for the board. Best the team has is gut feel plus last quarter's slides.

Forecast Agent rebuilds revenue and cost forecasts monthly using ML on historical data and live signals — pipeline, hiring, ramp curves, project starts. When the forecast moves, the driver-detection engine names the cause. Multi-scenario branching gives base, best, and downside in parallel. ±5% at the quarter is the steady state — not the ambition.

For who

Head of FP&A

“Annual plan obsolete by Q2. Manually rebuild every quarter.”

FP&A Director

“Sales says forecast moves by $20M. Why? Nobody knows.”

Finance Business Partner

“Re-forecast is 5 days of model rebuild monthly.”

CFO

“Board wants 'how is the year going to land'. Best I have is gut feel.”

Capabilities

What Forecast Agent actually does.

01Capability

ML-Based Revenue Forecast

Combines cohort retention, pipeline-weighted opportunities, ramp curves, and seasonality. Trained on your last 36 months. Re-trained monthly as actuals refresh.

02Capability

Cost Forecast With Live Signals

Headcount roster plus open requisitions, run-rate vendor spend, project-burst calendars, and cost-inflation indices. Cost line moves the day a req is approved, not the month after.

03Capability

Driver-Detection Engine

When the forecast moves, the agent attributes the change to specific drivers — pipeline delta, ramp slip, hiring acceleration, FX move. No more "why did the number change" reconciliation.

04Capability

Multi-Scenario Branching

Base, best, downside scenarios run in parallel. Each branch with its own driver assumptions. Compare side by side, version the lineage, hand to the board with confidence.

Continue exploring

More on Templates.

ISO 27001:2022 · GDPR · EU-hosted · Full audit trails · Trust Center →

See Forecast Agent on your stack.

Start free. See your first workflow run before the next quarter close.

Activate This Agent View Demo

ISO 27001

GDPR

CyberVadis

PA Certified

Peppol