The product

Everything finance needs to control cloud and AI spend — in one place.

Five jobs, done end to end: see where the money went, give every cost an owner, forecast what's coming, approve and govern changes, and report to the board. Sixteen connected modules under the hood — but you only ever see one joined-up system.

See & forecast Map & budget Govern & report Operate
Group 1 / 4 · See & forecast

Understand what was spent — and where it's heading.

Two modules that close the gap between today's cloud bill and next quarter's forecast: Analyze for the now, Forecast for the next.
Analyze

See spend the way finance thinks about it.

A high-level executive cockpit plus an interactive Cost Analyzer that lets you slice spend by project, BU, resource type, tag, or any virtual dimension. Drill from a stacked area chart to the exact resource powering the spike.

  • Period-over-period trending with pie, bar, area, donut
  • Daily, weekly, monthly groupings — no SQL needed
  • Fuzzy-search filter bar for fast navigation
  • Full Resource Explorer with utilization metrics
Cost trend · grouped by project
Last 90 days
Forecast models in production
ARIMA / AutoARIMA
Classical time-series
Prophet
Seasonality + holidays
Holt's / SES
Exponential smoothing
TiDE
Deep-learning, multi-horizon
Auto-select: InstantViewAI runs all five models, scores them on your history, and picks the best fit per resource. You can pin a model per BU if your team has a preference.
Forecast

Five forecasting engines, one auto-selected answer.

We don't pick a single method and force you to live with it. Cost Forecasting runs five proven forecasting methods in parallel, scores them on your own history, and uses whichever fits best — per resource, per project, and across the whole org. No data-science team required. (For your data team: the five methods.)

  • Forecast vs. actual, side by side in chart and table views
  • Confidence intervals, not just point estimates
  • Runs as its own fast, isolated service — so forecasting never slows the rest of the app
Group 2 / 4 · Map & budget

Give every cost an owner — and a ceiling.

Allocate turns resource IDs into BU and project owners. Budget turns those owners into people accountable for what they spend.
Allocate

Map every dollar to an owner — even the messy ones.

Build a multi-level organizational hierarchy (5+ levels deep) and allocate cloud costs by direct resource, tag, virtual tag, project, or percentage split. Unallocated cost is always surfaced — never hidden — so you reach 98%+ attribution within weeks.

  • Drag-and-drop org chart, with merging and reorganization
  • Virtual Tags — group costs with your own plain-language rules, version-controlled
  • Versioned Cost Models — historical reports never break
  • Reusable cost Components for standardized attribution
Organisational hierarchy
Acme Corp$1.28M
Engineering$842k
ML Platform$304k
Data Platform$256k
Application Eng.$282k
Product$214k
Marketing$198k
Unallocated$26.8k · 2.1%
Budget · ML Platform FY26
Active
$1,200,000 / annual
$816k spent · 68%$384k remaining
Period type
Quarterly
Allocation
Front-loaded
Finance owner
finance.lead@
Ops owner
D. Okafor
Budget

Budgets that behave like actual financial controls.

Define hard or soft budgets in any currency, scope them precisely with include/exclude clauses, assign finance and ops owners, and choose how the total is distributed across periods. Budgets live through a real lifecycle — draft → pending → active → paused → archived — and every transition is audited.

  • Even, front-loaded, or back-loaded period generation
  • Period versioning — draft, submit, activate
  • Real-time budget cockpit with risk assessment
  • Scopes tied to versioned cost models — auditable
Group 3 / 4 · Govern & report

Run the controls a board expects. Pass the audit they require.

Govern runs approval workflows and writes every change to an immutable audit trail. Report turns it all into board-ready PDFs without copy-paste week.
Govern

Approval workflows. Delegations. An audit trail you can hand to an auditor.

Define approval policies, route requests through delegated approvers, monitor policy health, and log every state transition. When the board asks who approved a $400k spike — the answer is one click away, with timestamps, identities and prior states attached.

Approval policies

Rule-based — "requires approval if amount > $50k". Health monitored per policy.

Delegations

Time-bounded delegation of approval authority to colleagues. Auto-expires.

Approvals inbox

A single inbox per approver. Lifecycle: pending → approved / rejected / cancelled.

Immutable audit trail

Every material state transition is recorded — budgets, approvals, users, allocation.

Report

Drag, drop, schedule. Board-ready in five minutes.

The Report Builder gives you a drag-and-drop dashboard editor with 15+ building blocks — KPI cards, charts, tables, forecasts, utilization, insight panels, even section dividers. Export to PDF, set a schedule, and it lands in the board's inbox automatically.

  • Configurable widget sizing, filters and time periods per report
  • Pixel-perfect PDF export, securely stored and shareable
  • Cron-based schedules — create, pause, resume, trigger
KPI
Forecast
Donut
Area chart
Insights
Table
Bar chart
+ Add widget
Group 4 / 4 · Operate

Live where finance already works. Don't move them.

The AI assistant answers "why did this change" in plain English. Data & integrations plugs into the cloud, billing and identity systems your team already runs.
AI assistant

Ask why the bill changed. In English.

The built-in AI Assistant has live access to your cloud inventory — servers, storage, networking, databases and more. Ask "why did ML platform jump 38% last week?" in plain English and get a real answer tied back to the specific resources that caused it, not just a chart.

You: Why did ML Platform spend jump 38% last week?
Assistant: ML Platform increased by $94,210 week-over-week, driven by three changes:
  1. A new GPU node pool (8× A100) brought online May 4 — accounts for $62k.
  2. Vertex AI batch prediction jobs +320% — likely an evaluation run.
  3. Cloud Storage egress +$8.4k from the model artifact bucket.
Sources: 3 resources · cost model v17 · IT change log
Data & integrations

Built on top of your existing cloud, billing and identity stack.

A real data pipeline behind every dashboard — BigQuery ingestion, dbt transformations, BigQuery-to-PostgreSQL transfer for operational queries, GCP KMS envelope encryption for secrets, and Keycloak for identity with white-label login via Keycloakify.

GCP Billing
Resource costs + inventory via BigQuery
AWS · Azure
Multi-cloud cost ingestion
OpenAI · Anthropic
LLM token spend unified into $
GCP Utilization
CPU, memory, storage metrics
Cloud Build
IT change correlation with cost
dbt
Analytical model construction
Keycloak / OIDC
Identity per tenant realm
Mailjet
Transactional email + reports
Multi-tenant by design. Each customer is a Keycloak realm with isolated data in PostgreSQL (admin schema) and BigQuery (per-tenant dataset). Service-account keys and secrets encrypted at rest using GCP KMS envelope encryption.

Want to see all of this against your own data?

Book a 25-minute walkthrough on your own billing exports — no sales call required.