The decision layer most capital project software is missing
Primavera tracks schedules. SAP tracks spend. Risk registers catalog risks. All three are necessary, none of them answer the questions the capital committee actually asks: which projects do we fund, is the sanctioned plan realistic, how much contingency is defensible, and what happens to the portfolio if macro assumptions change?
Capital Project AI sits at that decision layer. It reads from the tools you already use, applies probabilistic and portfolio math on top, and returns answers in the language of the capital committee — priority lists, trade-off tables, sensitivity walk-throughs. Four engines, each designed around one decision type.
Forecast — the probabilistic view of any plan
Deterministic plans lie. A capital project plan that says "this will cost $200M and take 24 months" is asserting something that almost never happens. Forecast replaces the single-point plan with a full distribution, calibrated against your firm's historical performance and a reference class of similar completed projects.
What it does
Ingests your schedule and cost model (Primavera P6, MS Project, Excel, ERP exports), converts each activity into a probability distribution, and runs Monte Carlo to produce the full outcome range — P10, P50, P80, P95 — plus the variance drivers behind each. The output includes reference-class comparison, optimism-bias adjustment, and a variance decomposition that tells the project team which risks are actually worth buying down.
When to use it
At sanction, at every stage gate, and any time macro or execution assumptions shift materially. Forecast is the engine behind probabilistic project planning, Monte Carlo project simulation, and AI project scheduling.
Optimize — portfolio decisions as actual optimization
The standard capital allocation process — rank projects by IRR, fund the list until capital runs out — leaves enormous value on the table. It ignores how risky each project is, how correlated they are with each other, and how much downside the portfolio can absorb before it's in trouble. Optimize treats allocation as a constrained optimization problem: maximize risk-adjusted value subject to capital, schedule, and resource constraints.
What it does
Runs expected-value math across thousands of scenarios for each project, applies fractional Kelly position sizing with drawdown constraints, handles correlations between projects explicitly, and solves for the optimal allocation under any combination of capital, engineering-hour, or strategic constraints. The output is a ranked priority list with explicit substitutions: drop project D, accelerate project A, and portfolio NPV moves by $42M.
When to use it
Annual capital planning cycle, mid-cycle rebalancing, any time the portfolio gets a new candidate project. Optimize is the engine behind project portfolio optimization with AI, capital allocation software, and capital efficiency software.
De-risk — stress-test the sanctioned plan
Every sanctioned plan carries assumptions that haven't been stress-tested: about oil prices, steel costs, vendor delivery, regulatory timing, productivity. De-risk runs the sensitivity analysis in seconds and surfaces the assumptions the plan is most exposed to, so management attention goes to the risks that actually move the answer — not the ones that made it onto the register.
What it does
Computes sensitivity of cost, schedule, and NPV to every material input variable, identifies the 3-4 drivers that typically explain 70%+ of variance, quantifies correlation between seemingly independent risks, and produces a buy-down analysis showing which risks are cheapest to mitigate and by how much.
When to use it
Post-sanction stress testing, board risk reviews, before committing management reserve. De-risk is the engine behind AI capital project risk management and megaproject risk management.
Deliver — track execution against the sanctioned plan
The gap between sanctioned plan and delivered outcome is where most capital projects lose money. Deliver tracks execution against the probabilistic plan — not the deterministic one — and surfaces the earliest signals of slippage, cost drift, or scope creep. It is not a project-controls dashboard; it is the decision-layer view on top of whatever project-controls stack you already run.
What it does
Ingests earned-value data, updates the probabilistic model with actuals, and flags when the live forecast materially diverges from the sanctioned P70 outcome. Recommends corrective actions with quantified impact — not just "you're behind schedule," but "accelerating structural steel fabrication costs $1.4M and buys back 6 weeks at P80."
When to use it
Monthly and quarterly project-review cycles, any time the sanctioned outcome comes into question. Deliver is the engine behind capital project management software and AI for EPC contractors.
How the four engines work together
Forecast produces the probabilistic view of any single project. Optimize allocates capital across projects using those probabilistic views. De-risk stress-tests the sanctioned plan for each funded project. Deliver tracks execution against the probabilistic plan in real time. Each engine feeds the others: Forecast outputs are Optimize inputs; De-risk insights feed back into Forecast calibration; Deliver actuals sharpen Forecast for the next planning cycle.
The result is a single decision loop — plan → allocate → sanction → execute → learn — that runs continuously, not just at annual planning time.
See all four engines on a real portfolio
Open the dashboard and run a portfolio through Forecast, Optimize, De-risk, and Deliver in a single session.
Launch the Dashboard →Who uses the platform
- Capital committees and CFOs — for the annual planning cycle, mid-cycle rebalancing, and stage-gate decisions. The outputs are the priority list, the trade-off tables, and the sensitivity analysis they've been asking for.
- Capital planners and project controls leads — for probabilistic schedule and cost analysis, buy-down quantification, and execution tracking.
- EPC contractors and engineering firms — for bid evaluation, schedule realism checks, and resource bottleneck detection against the active backlog.
- Private equity and infrastructure investors — for portfolio construction, position sizing, and correlation-aware allocation across capital-intensive assets.
Review specific use cases at Use Cases, or see pricing at Pricing.