AI FOR CAPITAL PROJECT DELIVERY

Know which dates will slip before they do, and fix them.

nPlan's AI has watched more than a million real projects play out. It shows you where yours will lose time, why, and which interventions buy the most of it back, months ahead.

Planners and risk teams: Schedule Studio is free to start.

Works with your planning tools and your AI assistant.

Oracle logoPowerproject logoMicrosoft Project Plan logoClaude logoMicrosoft Copilot logoChatGPT logoProcore logoMicrosoft SharePoint logo
NPLAN RESEARCH

Independent forecasts of major live projects.

Will the Chapelcross data centre campus finish on time?

Eight 120MW buildings in Scotland.

44 days late
Planned finish
7 Mar 2035
nPlan forecast
20 Apr 2035

32% chance of hitting the planned date

Spread of likely finish dates
What could move the date
Fix every risk we found
43 days
Substation stress test
2 days
Council waits for national planning guidance
96 days
Scottish Ministers take over the planning decision
342 days

Fixing every risk still leaves the campus a day late, so holding 7 March 2035 means changing the plan itself.

Read the full study
How we did it
What we did
01

Built the plan. From the planning application and public filings.

02

Projected every key date. Against more than a million real projects.

03

Found the reasons. Site set-up alone adds 86 days to the date.

04

Ran six scenarios. Three fixes, a stress test and two consent delays.

A forecast of nPlan’s reconstruction of the plan from public documents, not the developer’s own schedule.

Will Disney’s new Cars Land open in 2030?

Disney’s new Cars-themed land at Magic Kingdom, Florida.

347 days late
Planned opening
14 Mar 2030
nPlan forecast
24 Feb 2031

1% chance of hitting the planned date

Spread of likely opening dates
What could move the opening date
No change to the plan
Feb 2031
Biggest change to the ride sequence
Second half of 2030
Phased opening
First areas in 2029

On the current plan the land opens in the first half of 2031, so a phased opening is the only way to open in 2030.

Read the full study
How we did it
What we did
01

Built the plan. Eight phases from permits, filings and site reports.

02

Checked it against progress. Two dates reached, both as we projected.

03

Found the reasons. Ride testing and certification add 224 days.

04

Built a recovery plan. Eight interventions, each with its effect.

A forecast of nPlan’s reconstruction of the plan from public documents, not Disney’s own schedule.

Will Darlington’s first small modular reactor be running by the end of 2030?

The first small modular reactor in the G7, in Ontario.

840 days late
OPG’s working plan
Apr 2030
nPlan forecast
12 Aug 2032

1% chance of hitting the working plan date

Spread of likely operation dates
What could move the date
No change to the plan
840 days
Commissioning team ready early
127 days
Tunnelling crews and key trades
73 days
All six decisions together
283 days

OPG’s end-2030 target leaves too little buffer. Six decisions win back 283 days, and four of them are due before 1 December 2026.

Read the full study
How we did it
What we did
01

Built the plan. From OPG’s filings with the energy regulator.

02

Projected every key date. Against more than a million real projects.

03

Checked it against progress. Nine dates reached. Ours was nearer on six.

04

Tested six decisions. Each with what it buys back and when to decide.

A forecast of nPlan’s reconstruction of the plan from public documents, not OPG’s own schedule.

Will your project finish on time?

Send us your plan and the date that matters most, and we’ll run the same forecast on your project.

How it works
01

Share your plan in whatever format your planners already use, along with the date you care about.

02

We project every key date, with a first look ready in about 30 minutes.

03

We show what’s driving any delay, ranked by how much each reason moves your date.

04

We run the scenarios with your team, so you can see how much time each option buys back.

Your plan remains private to you, and nothing you upload is used to train the model unless you opt in.

Trusted by
Our dataset

The nPlan difference - a vast archive of past project data

The foundation of everything we’ve built is our dataset of more than 750,000 past project schedules - the largest collection of past programmes in the world, representing every capital project sector.

Why does it matter?

Past project schedules for retrospective analysis

With experience building projects for:

highway icon

Highways

railway icon

Rail

building icon

Public buildings

skyscraper icon

Commercial Buildings

bridge icon

Heavy infrastructure

valve icon

Oil & Gas

building icon

Utilities

database icon

Data Centers

Our AI has $500bn of projects under active management
Our AI has $500bn of projects under active management
Our AI has $500bn of projects under active management
Our AI has $500bn of projects under active management
Increase your Knowledge

Join the nPlan Academy

We’ve created a series of courses to enable current and future nPlan users to familiarise themselves with our software and the fundamentals of our unique technology. Master the new world of AI project controls and delivery in a few short hours, and gain new career-enhancing certifications in the process!

Learn more
rocket icon

Gain recognised certifications to boost your career.

school icon

Go ‘from zero to hero’ in approximately 2 hours

book icon

Enjoy full access for the duration of your employer’s engagement with nPlan.

Knowledge base

Deepen your forecasting and risk management expertise -and find out what makes us tick - on our blog👇

As seen in
book a demo icon
Talk to us

How can we help you today?

Follow us on LinkedIn

Why does it matter?

Because of something called 'Mean Absolute Error' or MAE. As more schedules become available for AI-led forecasting and risk management, MAE falls dramatically, meaning more accurate forecasts.

To understand more about the importance of MAE, read our Model Performance 101 white paper.

Why nPlan - dataset