Your capital, smarter

Capital decisions, sized against the bad year.

Works on top of the systems that run your projects and portfolios

  • Schedules
  • Cost control
  • ERP and ledgers
  • Portfolio monitoring
  • Board and IC papers
  • Market and supplier feeds

Where capital is committed

Refining & Petrochemicals LNG Power & Nuclear Renewables & Storage Transmission Data Centres Mining & Metals Corporate Capex Infrastructure Funds Real Estate Private Equity Portfolios

Every option · Every scenario · Every approval

Capibud runs schedule, cost, value and shared risk on one set of correlated scenarios, sizes each commitment against the risk appetite you set, and routes it to the person who signs. Nothing changes the plan until a person approves it.

Every project · Every dollar · Every decision

Capibud holds your whole capital programme or portfolio in one plan: schedule, cost and value, on one shared set of scenarios. It allocates the envelope to the projects and assets that create the most value. It re-balances when something changes and shows the likely outcome next to the bad-year outcome, so commitments are sized to hold either way. Each proposal goes to the person who signs it, with the reasoning attached, and nothing changes the plan until they approve.

Read→Simulate→Size→Approve→Replay

Where Capibud sits

A decision model on top of the systems you already run

Your schedule, cost, ERP and portfolio systems record what is happening on every project and asset. Capibud reads those records, decides where the next dollar does the most across the whole book, and sends the approved budget back. The systems stay as they are, and every update they record starts moving money to where it creates value.

Capital committee and boardSign the decision, with the reasoning attached
CapibudDecision model and budget allocator
  1. Readwith sources
  2. Simulateshared scenarios
  3. Allocateone envelope
  4. Approveyour authority matrix
  5. Recordreplayable

Capibud reads

  • Progress and float by activity
  • Forecast finish dates
  • Schedule exports (XER, XML)

Capibud sends back

  • Re-phased start dates the committee approved
  • Tripwires, e.g. critical-path float under 20 days
01

Reads what they record

Progress, cost to complete, commitments, actuals and memos, each value linked to the file or record it came from. File exports work today; API connectors switch on with your credentials.

02

Decides across the whole book

Every project, capex line and investment draws on one envelope and one set of scenarios, so the next dollar goes where it does the most, sized to hold in a bad year.

03

Sends the decision back

Approved budgets, re-phasing and tripwires go back to the systems and the people who run the projects, so the plan of record and the capital plan stay one plan.

Who it is for

Owners who build, and the investors who fund them

A refiner sizing its capex round and an infrastructure fund sizing its next commitment face the same problem: many claims on one pool of capital, risks that move together, and a committee that has to sign. Capibud runs both on the same engine, with the same record behind every decision.

Capital committees at owners and operators

Refiners and chemicals, utilities and grids, developers, manufacturers, data-centre and energy builders.

  • Which units to sanction, and in what phasesP50, P80 and the value of staging for every option at FID.
  • How to fund this year's capex roundHundreds of lines packed into one envelope, crews and shutdown windows.
  • Where recovery money goes when a project slipsRecovery options priced in money before anyone commits.
A coastal refiner's capital book →

What Capibud does, every day

Six capabilities on one plan: allocation that re-solves itself, files read as they are, connectors to the systems you already run, a brain that watches the market, projects built from a sentence, and the range behind every number.

01

Dynamic allocation

The envelope, always at its best.

Capibud solves your whole programme or portfolio for the highest expected value within the headroom you set. It re-solves whenever a cost, schedule or price moves, then proposes what to bring forward, re-phase or hold, with the value of each move and the confidence behind it.

02

Self-learning ingestion and ontology

Reads your files the way your team does.

Schedules, cost sheets, ledgers, registers and status reports arrive in any layout. Capibud classifies columns by meaning and turns narrative into typed facts, and every figure links back to the line it came from. Each correction teaches it your vocabulary, so the next file needs less help.

03

Connectors

Works with the systems you already run.

Capibud reads what your schedule, cost and ERP systems export today (XER, XML, Excel and CSV), and API connectors switch on with your credentials. Approved budgets and re-phasing go back the same way. Proposals can notify Slack, Teams or e-mail once you switch those channels on.

04

The always-on brain

Watches the world so your team doesn't have to.

Price prints, filings, supplier news and progress updates flow in continuously. When something moves, Capibud traces which projects it touches, prices the impact across the portfolio and drafts a re-allocation, then waits for a person.

05

Natural-language and voice project builder

Describe the project; Capibud builds the model.

Say or type "a 2 GW solar park in Rajasthan, EPC by Q3 2028", and Capibud assembles the schedule, cost lines and value drivers from the matching industry pack. Your team adjusts it in plain language and it joins the plan in minutes.

06

Portfolio outcome analytics

See the range, not just the number.

Every project and the whole portfolio show the bad year, the likely year and the good year: for value, P10, P50 and P90; for cost, P90 is the bad year. Your committee sees how confident to be before it commits, and whether the plan still holds when several things go wrong together.

Questions

Three questions decide whether a capital plan survives the year. Capibud answers all three on the same book, in the same run.

Which options deserve capital?

Rank by value for the risk taken, with the P80 doing the talking instead of the P50.

How much should each one get?

Size every commitment in dollars, inside the envelope you actually have.

What happens in the bad year?

Stress the whole book and read the chance of breaching your risk appetite before you commit.

What changed, and what should move?

Every forecast carries the band the committee agreed to. Capibud checks actuals against it every day. When a line leaves the band, it flags the move, traces the projects it touches and drafts the re-allocation, with the value of acting now beside it.

Illustrative example, figures invented. The band is the P10–P90 range agreed at approval.

Capital committee

The capital decision journey

Eight steps from an envelope and an appetite to a vote the committee can defend, on the real committee screens. Illustrative data for an invented coastal refiner.

Explore the committee

Frame

Model

Prioritise

Act

Frame01 / 08

Set the envelope and the appetite

The committee fixes the money and how bad a year it will accept: $600M over three years, with the 1-in-20 book value held above $700M.

0:00 / 1:04

The brain

Watch a signal move
through the book

The monitoring layer of the decision model. It reads what your project systems record and what market and supplier feeds report, finds every exposed project, prices the effect in money and drafts a re-allocation. The engine checks the draft; a person decides.

Explore the brain

Decision record

Decisions you can replay, years later

The systems that run your projects record what was spent and when. Capibud records the decisions: what the committee knew, what it chose and why, and the run it relied on. When someone asks why a project was funded, the answer is one click away.

Explore the record

Illustrative decision. Arrow keys move between beats.
Hash chain · last six of 1,519 records sha256(prev_hash ‖ canonical row)

Select a record to see what it holds.

Every claim on the envelope,
sized and stress-tested in one book.

Pick a stress scenario and watch the whole book re-price at once. When the bad-year value falls below the floor your committee set, Capibud proposes the smallest change that restores it. Click any claim to see its range.

Capital committee›Stress & reverse stress Illustrative data
4,000 shared scenarios · floor: bad-year value ≥ $700M
Risk appetite met. Bad-year value $815M, above the $700M floor.
Capital committed$470Menvelope $600M
Value after hurdles$1,180Mexpected, all scenarios
Bad-year value (1-in-20)$815M floor $700M · met
Smallest shock that breaks itCopper +2.1σreverse stress
Claims on the envelopeClick a claim to inspect it
ClaimDecisionBudgetP50 returnBad yearP10 returnVerdict
Polypropylene unitEvaluate · growthFund

Illustrative figures, not client data. Returns are internal rates of return across 4,000 shared scenarios.

Four decisions, one book

Which plants to build, which assets to back, how to recover projects in construction and which capex lines to fund this year. They draw on the same money and the same risks, so Capibud runs them on one set of scenarios.

Built for the committee, not the spreadsheet

You don't have to take a recommendation on faith. Every number traces back to a file, a parameter or a run, and every run replays to the same result.

Ranges, not points.

Every option is priced as a P50, P80 and P90 range for cost, schedule and value, never a single number dressed up as certainty.

Shared risks, modelled as shared.

One supplier, one EPC crew or the copper price hits every exposed project in the same scenario, so concentration shows up before it bites.

One envelope, not ranked lists.

Every candidate from the four decisions is placed on one envelope and one risk appetite, instead of each budget holder funding down their own list.

Single-point plans Risks treated as independent Ranked lists, not sized bets Numbers nobody can replay
Book value across 4,000 scenarios Same projects, same estimates · illustrative $0.6bn$1.4bn$2.2bn

P50 · P80 · P90 ranges Correlated scenarios Sized against your appetite Every run replays

Stress-tested by default.

Alternatives, stress tests and reverse stress tests (the smallest shock that breaks the plan) sit beside every recommended plan.

Approval before any change.

Changes are previewed, routed by your delegation-of-authority matrix with maker-checker rules, and can be undone.

Audit-ready.

An append-only, hash-chained log stores each run's inputs, seed and engine version. OCR runs locally, so documents stay on your server.

What the evidence shows

Megaproject replay on public information. Completed megaprojects replayed quarter by quarter, using only what was public at each date, then scored against how they actually finished.

QuarterlyPoint-in-time data TestedNo look-ahead MissesShown, not hidden

The replay and its results are shown in a guided walkthrough after sign-in.

How a replay works

A backtest built to break us. Our P80 against single-point forecasts on 210 projects, then on four generators that violate our assumptions.

77%P80 cost held 84%P80 finish held 29%Single-point

Target 80%. Under our own generator; the four adversarial generators are scored too.

How we measured

Why large programmes overrun. Single forecasts, risks treated as independent, money allocated pot by pot. The pattern holds across countries and sectors.

91.5%Over budget or late 0.5%On cost, time & benefit

Flyvbjerg's database of 16,000+ projects in 136 countries.

Industry briefs

Start from researched ground. Every project opens from an industry pack: sourced parameters, price feeds, the factors that usually move value, and a project template, pinned to the decision.

37Industry packs 662Sourced parameters 651Cited sources

Every number links to its source, or says plainly that it is an estimate.

Industry briefs

Three steps from files to decision

1ConnectPoint Capibud at the exports your project, cost and ERP systems already produce, plus the memos and options on the table. No templates, no re-keying.
2SimulateThe engine ranges every estimate, links the risks projects share, and sizes the book against your risk appetite.
3DecideWalk into committee with sized recommendations, the shock that would break the plan, and a record that replays.

The fastest way to evaluate Capibud is to point it at a decision you are making this quarter.

Bring us one real decision.

One working session on a decision you are making this quarter.

Write to hi@capibud.com
Talk to us →