FinHelm is a unified probabilistic finance platform organized around one architecture, six Applications, and three commercial tiers. The user encounters one product. The architecture compounds beneath it.
The legacy FP&A stack — Adaptive, Planful, Anaplan, Vena, Mosaic, Cube — was built around a single-point forecast that everyone knows is wrong but no one measures. FinHelm replaces the point estimate with the probability distribution as the unit of planning. Six pillars of intelligence around one calibrated forecast. Four user-facing metrics. A closed loop that gets better every period.
This page is the operational reading. The canonical methodology, the scientific foundations, and the four-axis intelligence frame live at /methodology. Here, the architecture is rendered as a product: six Applications across two categories, three commercial tiers, distributed where the CFO already works.
The argument for FinHelm fits on a single line: the data the platform needs already exists in the ledger, the math has been peer-reviewed for three centuries, and the technology to run the math in milliseconds converged in 2024 and 2025. Before that, probabilistic FP&A required a quant team and a six-figure budget. After it, it requires a CFO and twelve minutes. FinHelm is what closes that gap.
FinHelm runs on a three-layer commercial architecture over a single canonical IP foundation. Layer 1 is the Helm Stack — six pillars, patent-protected, built once. Layer 2 is the surface — six Applications in two categories. Layer 3 is the commercial structure — three tiers from $49 to $2,500+.
FIG. II.a · The Probability Helm Stack — the canonical Layer 1 architecture. Six pillars (UES™ · FSI™ · AMI™ · Reflection Engine™ over the Monte Carlo and ProbabilisticCell™ substrates). The full reading is at /methodology. On this page, the diagram is the foundation everything else assembles on.
The platform separates what the user pays for from what the user receives from what the platform actually does. This separation is not cosmetic. Each layer compounds independently — Core IP through accumulated calibration history (Reflection Engine™ gets better with every actual that lands), Applications through tier-by-tier feature density, Tiers through the upgrade narrative from Beacon to Compass to Command.
Three commercial tiers that determine which Applications a user can access and at what depth. Beacon includes the entire Probabilistic Analysis category. Compass adds Probabilistic Planning. Command adds Reflection Engine™ closed-loop calibration depth on every Application, multi-entity rollups, and PE-specific capabilities.
Probabilistic Analysis (Dashboard · Variance Analyzer · Probabilistic Cash Runway) and Probabilistic Planning (Scenario Builder · Planning Grid · Driver Sensitivity). Each Application has its own URL, its own page, and its own job in the FP&A workflow.
UES™ (Magnitude) · FSI™ (Stability) · AMI™ (Attribution) · Reflection Engine™ (Calibration) over the Monte Carlo simulation substrate and the ProbabilisticCell™ temporal data substrate. The mathematical and architectural foundations every Application depends on. The deeper this layer runs, the harder displacement becomes.
FIG. II.b · The three-layer architecture. The user encounters one unified product. The architecture compounds beneath it.
Six Applications organized into two categories that map to discrete phases of the FP&A workflow. The categories are stable. The third category — Probabilistic Reporting — is deferred to roadmap until the methodology can support it.
Be the signal.
Beacon · $49 / moThree Applications for month-end close, board preparation, and covenant monitoring. The Analysis category answers the questions “what happened” and “why did it happen.” Every Beacon user gets all three Applications immediately at signup — no upgrade gate, no demo timer, no feature throttle.
Single canonical view of the methodology running. UES™ score, Monte Carlo confidence band, ranked variance drivers, calibration history snapshot — everything a CFO needs before the board meeting.
Decompose variances into structural bias and stochastic noise. Rank drivers by contribution. Read narrative explanations of why the actual landed where it did. The end of “we missed by 12%” with no why.
Cash runway as a confidence band, not a single date. P10 worst-case, P50 base, P90 upside. The number to tell the board, the number to tell the team, and the number to tell yourself — with explicit uncertainty.
Find your direction.
Compass · $499 / moThree Applications for the budget cycle, scenario modeling, and driver analysis. The Planning category answers the question “what should happen next.” Compass users access all six Applications — the entire Analysis category plus the entire Planning category — without per-feature gating.
Build named scenarios with explicit driver assumptions. Compare scenarios against each other and against the base forecast. Every scenario carries its own UES™ score, its own confidence band, and its own narrative of what changed.
Driver-based probabilistic planning. Every cell shows uncertainty. Every formula propagates distributions. The spreadsheet feel without the single-point delusion — built on the ProbabilisticCell™ primitive that distinguishes the platform. Includes a P&L view.
Tornado charts in burgundy. Which assumptions move the forecast most. Which drivers are signal and which are noise. The visualization that turns a CFO’s intuition about “where the risk lives” into evidence.
FIG. III.c · Two categories, six Applications, three-and-three symmetric. The tier mapping is one-to-one with the categories — Analysis is Beacon, Planning is Compass — which makes the upgrade narrative crisp.
A seventh experience surfaces outside the six-Application grid: the Forecast Health Check Agent. Available free at finhelm.ai/assessment — twelve minutes from arrival to first probabilistic insight, no login required before the score reveal. The agent runs UES™ Lite on the last six months of budget-versus-actual data and returns a score, a band, and the four-component breakdown. It is the free entry point for Beacon.
Command, the third tier at $2,500+/mo, does not unlock new Applications. It unlocks depth — Reflection Engine™ closed-loop calibration on every Application, multi-entity rollups for parent-plus-PortCo structures, the five PE-specific capabilities, SSO/SAML, and dedicated CSM. The Compass-to-Command upgrade is not an addition. It is a deepening of every Application the user already has.
FinHelm runs above the ledger you already have. No migration. No new system of record. No IT project to start. The data flows in, the analysis runs, and the forecast is recalibrated to the actuals as they land — all without disturbing the system your finance team relies on to close the month.
Integrations are data adapters, not partnerships. The platform never depends on a single ledger. If your finance team operates across multiple systems — a holding-company QuickBooks, a portfolio NetSuite, a sister-entity Sage Intacct — FinHelm reads from all of them and produces one unified probabilistic forecast across the entire group.
QuickBooks Online
Live
DualEntry
Live
Rillet
Live
Sage Intacct
On roadmap
NetSuite
On roadmap
SAP Joule
On roadmap
Campfire
On roadmap
Xero
On roadmap
FIG. IV.a · FinHelm reads from any general ledger in the FP&A workflow. The architecture is integration-agnostic by design — the same Probability Helm Stack reads the same way regardless of which system the data originated in. CSV upload is supported for ledgers not yet on the active integration list.
For finance teams running on two or more ledgers, the multi-source read is one of FinHelm’s practical differentiators. The unified forecast carries the same UES™ score, the same Monte Carlo confidence band, and the same ranked driver list across every entity in the rollup. The board reads one number with one calibration history. The work happens once.
The three commercial tiers are not three sizes of the same product. Each tier answers a different question the finance function is asking. Beacon reads. Compass plans. Command runs the closed loop across multiple entities.
$49 / month
Be the signal.
The entire Probabilistic Analysis category. Dashboard, Variance Analyzer, Probabilistic Cash Runway. Forecast Health Check Agent. For finance teams who need to read what is happening — and explain it.
$499 / month
Find your direction.
Beacon plus the entire Probabilistic Planning category. Scenario Builder, Planning Grid, Driver Sensitivity. For finance teams running the full UA-FP&A™ discipline — analysis and planning, every period.
$2,500+ / month
Close the loop.
Compass plus Reflection Engine™ depth on every Application, multi-entity rollups, PE-specific capabilities, SSO/SAML, dedicated CSM. For organizations who treat calibration as governance.
FIG. V.a · Pricing detail, feature matrix, and the free Beacon entry option live at /pricing. The tier definitions above are canonical; the pricing page is the commercial reading.
Three entry points, depending on which seat you are sitting in.
Run the Forecast Health Check.
Get your UES™ Lite score in twelve minutes. Six months of budget-vs-actual is enough. No login required.
finhelm.ai/assessment →Read the canonical methodology.
Six pillars. Four metrics. The closed-loop intelligence model. The Forecast Quality Matrix. Three centuries of mathematical lineage.
The Methodology →Request the investor brief.
Category geometry, the architecture, and the acquisition thesis. Sent under password-protected link.
hello@finhelm.ai →Twelve minutes. No login. The free Forecast Health Check at finhelm.ai/assessment is the fastest way to read where your forecasting system actually is.