Jakten Risk User Guide

Document ID: M17-CERT-001-D12

Audience: Risk analysts, estimators, project managers, and reviewers

Last updated: 2026-10-16

Applies to: Jakten Platform 3.0

1. What Risk does

Jakten Risk turns an Estimate Version into decision support for uncertainty and contingency. The integrated workbench combines estimate uncertainty, dependencies, line risks, discrete risk events, schedule-driven cost, quantity growth, Monte Carlo simulation, comparisons, and executive reporting.

Risk does not replace the estimate source. It analyzes the selected Platform Estimate Version and preserves saved analytical results as the basis for reports and cross-version comparison.

Jakten AI in Risk. When your company has a Jakten AI subscription and an administrator has opted in, Risk can draft risk events for you to complete (section 7) and draft the Full Report's analysis summary (section 11). AI drafts and a person decides: a drafted event never carries a probability, cost or schedule impact, and AI never runs inside a simulation or changes a saved result. Both AI actions need internet; everything else in Risk works without it. What each sends, and what is kept, is in Jakten Risk AI Data Handling.

2. Open the correct workspace

Before opening Risk, confirm:

• Jakten Risk is turned on for the company and in use on the project (chosen at project setup, or added under project Settings → Modules).
• You hold a Risk seat and have project access.
• The project Overview has the intended working Estimate Version selected.

Open the project and, under Open a workspace, select Open Risk. Risk opens on the last workbench tab used in the current session, or Settings on first integrated launch. Use Back to project (top left, beside the menu button) to return to the project. For a how-to question, select Ask Jakten in the top bar. It answers from Jakten's documentation when your company has turned on Jakten AI, and it never sends your risk model (see the Platform User Guide, section 9).

The Risk workbench tabs are Settings, Estimate, Quantity Growth, Dependencies, Line Exposures, Events, Schedule, and Results. Save and Run remain available across the workbench.

3. Recommended workflow

1. Review simulation and project defaults in Settings.
2. Load or review the estimate basis in Estimate, and set any quantity growth ranges on the Quantity Growth sheet.
3. Model relationships in Dependencies, including growth correlations.
4. Enter line-level uncertainty in Line Exposures.
5. Add discrete project risks and opportunities in Events, and mark the ones that drive completion.
6. Set the planned duration and time-related cost (house load) in Schedule.
7. Choose Run. Every run is the full analysis.
8. Review funding, exposure, sensitivity, finish dates, and comparison results in Results.
9. Save and export the appropriate report.

You can revisit any step. When a change invalidates saved analysis, Risk shows that a rerun is required rather than presenting old results as current.

4. Settings

Use Settings to confirm simulation behavior and project defaults before the first run. Project defaults include the risk-potential bands and dependency strengths used by editing tools.

Change defaults deliberately: they affect how new or updated modeling inputs are interpreted. Existing saved results remain evidence of the settings used when they were created.

5. Estimate

The Estimate tab is the cost basis for the simulation.

Available actions include:

• Add Row for manual estimate input
• Bulk Edit for repeated changes
• Import Estimate for spreadsheet or CSV input
• Export Estimate for a working extract
• Clear Filters to return to the full grid

In an integrated project, prefer the estimate basis Jakten Assurance publishes. The first time, the Estimate tab lists the estimates Assurance has published: select one, choose Preview comparison, then Use in this analysis. When a newer publication is waiting, review it and choose Accept estimate update. Review mapped columns, row identity, WBS, cost type, and uncertainty values before accepting an import. A completed Risk simulation is tied to its estimate basis; material estimate changes require a new run.

5.1 Quantity growth ranges

Pricing uncertainty and quantity growth are two different things, and the Quantity Growth sheet (its own workbench tab) keeps them apart:

• A pricing range is "what each unit will cost."
• A growth range is "how much more (or less) scope there will be."

Open the Quantity Growth tab. It has one row per classification (connected mode) or per WBS group (standalone) — the same groups the simulation draws for — with columns for WBS, Description, Cost, Growth Potential, Low % / Likely % / High %, Distribution, and Pricing and growth together. Set growth just as you set risk potential on the Estimate sheet: choose a Growth Potential band to fill Low / Likely / High from your project defaults, then edit the numbers directly if you like. A growth range applies to every cost type in the group — Labor, Material, Equipment, Subcontract, Other and Indirect alike — because more (or less) scope changes them all. Low may be negative to model a scope reduction, but it must stay above −100%: scope cannot shrink by 100% or more. To turn growth off for a group, choose Clear range.

A defined range is simulated — there is no separate confirmation step. Keep your pricing ranges (Low / High on the Estimate sheet) assuming fixed scope, so a growth range and a pricing range do not both count the same growth. The Pricing and growth together column shows the two combined for each group, so any double count between them is visible rather than hidden.

Growth Potential bands are a project default: edit them under Settings → Project Defaults → Growth Potential Bands, beside the risk-potential bands.

Growth never changes a published Likely and never writes back to the estimate. A republished Likely keeps the growth ranges you set against that classification.

6. Dependencies

Use Dependencies to model correlation among estimate packages or groups.

Select Apply Dependencies after review. Resolve diagnostics before relying on a full simulation. A dependence matrix that is incomplete or invalid can materially change the distribution.

The Growth view is a separate matrix from the line-by-line one, headed Growth dependence. It carries Spearman rank correlations (from −1.00 to 1.00) between growth ranges only: a growth range can be correlated with another growth range, not with a line or with the schedule. Scope growth is usually systemic, so groups that grow together should say so — drawn independently, the upper tail is understated. The view offers only the growth ranges the run would actually sample, so it stays empty until two or more sampled ranges exist. Edit the upper triangle only; the diagonal is always 1.00.

7. Line Exposures and Events

Use Line Exposures for escalation and productivity exposure attached to estimate lines. You can bulk edit, import, export, and filter the grid. The grid's Schedule-Related Cost Exposure column is retired under the current model — see section 13; price schedule cost on the Schedule tab instead.

Use Events for discrete project risks and opportunities that are not represented adequately as line uncertainty. Select Add Event, enter probability and impact information, and keep the event description specific enough for a reviewer to understand the modeled basis.

An event's schedule delay is entered as a range — Delay if triggered (days): Low / Likely / High — with a Delay distribution. Turn on Drives completion for an event that pushes the finish date out; leave it off for a delay that runs beside the critical work and costs time without moving completion. With Drives completion on and a schedule set up for the scenario (section 8), those delay days are priced at the scenario's time-related rates and land in the cost simulation.

Avoid counting the same exposure in both Line Exposures and an event unless the two entries intentionally represent different effects.

Proposed events. Beside the events table, Proposed events lists events waiting for a decision. Nothing in it is part of the risk model, or of any simulation, until a person adds it:

• From Quote. When a Jakten Quote award bound to the project published risks that were accepted or are awaiting clarification, each appears here marked Recorded in Quote, with its title, type, severity and any priced exposure.
• Drafted by AI. Draft risk events reads the scenario's estimate (for the scenario bound to an estimate Version, as Assurance published it), the Knowledge Assurance published for that Version, and the Quote risks above, and proposes events the model does not already hold, each with notes on its cause, a WBS element and the sources it was drawn from. It never proposes a number, and Risk leaves out any proposal that carries a figure its sources don't state, a date, an owner or advice, or that repeats the name of an existing or proposed event.

Accept opens the event editor with the name, notes and WBS filled in and every number empty: enter the probability and the cost impact, the schedule delay, or both, then Add event. Reject discards a proposal, and Cancel stops an accept you started. While someone else is adding a proposal, its card says Another person is adding this event. A proposal drawn from Knowledge or a Quote award that has since been replaced or withdrawn is marked so the next time someone who can edit the project opens the page. An event that has neither a cost impact nor a delay blocks Run until it has one.

8. Schedule and time-related cost (house load)

The Schedule tab holds the scenario's schedule model — a planned duration, how far it may vary, and what each day of that variance costs. This is house load: the time-related cost a project incurs for every day it runs beyond its planned finish. Risk models only the change from the planned duration; it is still not a CPM schedule model, and there is no activity logic or critical path.

House load starts off. The panel, titled Schedule and Time-Related Cost, reads "House load is off. This scenario has no schedule model" until you choose Set up a schedule. To turn it off again, choose Remove this schedule.

When house load is set up, enter:

Two rules the panel states on screen:

• Every duration is in calendar days. A week is 7 days and a month is 365.25 / 12 days. A working-day rate has to be converted before you enter it; no conversion is made for you.
• Do not enter a cost the published estimate already carries as a direct line. Assurance publishes directs only, so planned general conditions are not in Risk's base — but equipment rental carried as a direct line is, and entering it here counts it twice.

What you enter here is priced in the cost simulation: every iteration samples the duration range and the rates, and the cost lands in the total, the percentiles, and the funding contingency. Run the simulation again after you change it.

If the project's start or finish dates change after you set the planned duration, the panel shows the change and offers Use the new dates. Risk never applies a date change to the planned duration for you, because the planned duration is the scenario's own modeling assumption.

9. Run analysis

On Results, choose Run. It is the only run command: every run is the full analysis, which simulates the estimate and refreshes contingency attribution, sensitivity, comparisons, schedule outcome, quantity growth, and insights together.

Risk performs readiness checks before analysis. If the review panel identifies blocking inputs, open the first affected tab, correct the issue, save, and run again. A growth range whose Low / Likely / High are out of order, or that shrinks scope by 100% or more (a percent at or below −100%), is flagged there and navigates to the Quantity Growth sheet.

When inputs change after a run, Results keeps the last run's figures on screen with a notice that they changed since the last run. Choose Run again before relying on them.

10. Read the Results workspace

Results can include:

• Cost distribution and cumulative confidence curves
• Selected funding position and exceedance probability
• Contingency bridge and source attribution, with layers including Quantity Growth and Schedule Duration
• Sensitivity and major uncertainty drivers, including entries named "Quantity growth:" followed by the classification name
• Dependency, line exposure, and event exposure
• Schedule outcome: the simulated finish, delay days, and schedule-driven cost
• Quantity growth: how scope growth changed the cost (a reduction shows as a negative figure)
• Scenario comparisons
• Estimate-Version evolution and changed-line drivers

Finish-date percentiles. When a scenario prices a schedule, the Schedule and Delay panel shows the Simulated finish at P50 and at the funding percentile, alongside the planned duration and the expected delay. The finish is reported as a date when the project has a start date, and as elapsed calendar days from start otherwise. A Delay against time-related cost scatter plots one point per sampled iteration. If a scenario marks events as driving completion but prices no time-related rate, the panel reports the delay days and says plainly that the delay is not costed.

Schedule-driven cost and quantity growth in the contingency. Both are part of the simulated total, so both are inside the funding contingency. The contingency breakdown carries a Schedule-driven cost (house load) row and a Quantity growth row when a run priced them. The Quantity Growth panel shows the mean growth and the growth dollars at the funding percentile; because classifications do not peak together, the per-classification figures at the funding percentile do not sum to the layer total, and the panel says so.

When a result or comparison says Unavailable, the required saved analytical basis is missing. Run and save the necessary analysis; do not interpret unavailable as zero.

11. Export and share

From Result Actions, choose:

The Executive Brief adds a finish-and-lateness line when a run priced a schedule, and the Full Report carries the schedule outcome and quantity growth sections.

Analysis summary (Jakten AI). On the Results workspace, the Analysis summary panel holds the paragraph the Full Report prints on page 9 for this result. Draft summary drafts it from the saved result's figures and contingency drivers, as the save sealed them. Every figure it states is one of those, printed as the report prints it, and Risk checks the draft before it is shown: a figure the result does not hold, a date, an owner, an action or advice is refused. Edit it if needed, and Accept as my text; your edits pass the same check. The summary prints only for the result it was drafted from: after a new run is saved, the panel shows the last summary as out of date. Wait for the export to finish before closing Platform. Review the project name, Estimate Version, scenario, target confidence, and analysis date before distributing a PDF.

12. How Risk models cost

Risk runs a Monte Carlo simulation over the estimate: it draws many possible outcomes and reads the funding contingency off the resulting distribution. Each draw combines several sources of uncertainty:

• Pricing ranges. Each line carries a Low / Likely / High range for what its work will cost. This is the core of every simulation.
• Quantity growth. A growth range says how much more (or less) scope a classification may carry. Growth is drawn once per classification each iteration and applied to every cost type in that classification, before escalation and productivity, so escalation applies to the grown scope. A negative range lowers the scope. It never changes a published Likely.
• Correlated growth. When growth ranges are correlated on the Dependencies Growth view, classifications that tend to grow together do so in the simulation, which widens the upper tail appropriately.
• House load (schedule-driven cost). Each iteration samples the project's delay — the duration range plus the delay from events that drive completion — and prices the time beyond the planned finish at the entered time-related rates. The time-related costs of the planned duration itself are not in Risk's cost base and are not priced here: house load prices only the change from the planned duration.
• Driving events. A triggered event that drives completion adds its sampled delay; its cost and its delay come from the same trigger. Delays from driving events add in sequence.
• Dependencies. Correlation among estimate lines is modeled with a dependence matrix so related lines move together.

The funding contingency Risk reports, and that Jakten Assurance reads, includes all of the above — pricing spread, quantity growth, schedule-driven cost, dependency effects, and discrete risk events.

Limitations, in plain words:

• Risk is not a CPM, critical-path, or network schedule model. Delays from events that drive completion add in sequence, which is conservative.
• The time-related rates do not rise with delay, so very late outcomes may be understated.
• Quantity growth is independent of the schedule, and independent across classifications unless you set growth correlations.
• Discrete risk events are independent of one another.
• Contingency attribution is sequential and order-dependent: each layer is measured after the ones before it.
• Standalone group exposures are not an additive allocation of the project's funding contingency, and they exclude schedule-driven cost (house load).
• The reconciliation residual in the contingency breakdown can include overlap between layers, order effects, methodological differences, and Monte Carlo noise.

13. What "retired under v3" means for existing projects

Earlier releases carried a per-line Schedule-Related Cost Exposure factor on Line Exposures. The current model prices schedule cost through house load and driving events instead, so that factor is retired: stored values stay readable on Line Exposures, but nothing samples them. The Line Exposures column reads Schedule-Related Cost Exposure (retired) and is read-only.

The first time you run a project that still carries a nonzero legacy factor, Risk shows a one-time notice titled Schedule-Related Cost Exposure is retired. When a previous run recorded what the factor added, the notice reports the lines and the amount from that stored result — never a recompute. To keep that exposure, enter house load on the Schedule tab, or mark the events that drive completion. The notice does not block the run and does not return once shown.

Because the retired factor is no longer applied, an existing project's contingency can change when you rerun it under the current model. The direction depends on the project: removing the retired factor lowers the figure where it was inflating every line, while house load itself raises the figure where real time-related cost was not previously priced. Review the new result before relying on it.

Old saved results are never recomputed. A result saved under an earlier model keeps its stored figures and its original wording; only a new run uses the current model.

14. Common problems

Related documentation