nPlan Case Study - AI at work!

The following document of a case study of nPlan is summarised below.

nPlan Case Study - How SCS is using AI-led forecasting and risk management to construct th.

Download • 3.92MB

From the nPlan.io website :

'Automatically forecast every activity in your schedule and surface the activities most in need of proactive mitigation.

Unlike traditional risk management software such as Primavera Risk Analysis, Safran Risk, or Deltek Acumen Risk, nPlan Insights Risk Professional calculates activity uncertainty using Machine Learning and a dataset of 750,000 past schedules - say goodbye to grueling quantification workshops, rolled-up schedules, and the rest of the pain that comes with traditional QSRA - and get used to generating unbiased forecasts for every activity in your schedule, and for every iteration of your schedule, at the click of a button.

Then watch as the productivity of your risk workshops and risk reviews--now focused on treating rather than quantifying risk--shoots up!'

A summary of the paper:

• Project Overview:

  SCS JV, a joint venture of Skanska, Costain, and STRABAG, is constructing the London tunnels for the high-speed railway project HS2, valued at £4.1 billion. The project involves 30,000 workers, extensive railway construction, and the use of six tunnel-boring machines (TBMs).

  
  

• Use of AI in Forecasting and Risk Management:

  SCS collaborated with nPlan, a leader in AI-led forecasting, to enhance planning, identify risks, and manage schedules for the HS2 project. nPlan uses deep learning on a large dataset of over 750,000 historical project schedules to predict risks and improve project outcomes.

  
  

• Key Objectives:

  1. Identify risks early to avoid unplanned costs and schedule overruns.

  2. Prioritize risk management efforts efficiently.

  3. Replace traditional backward-looking assurance with continuous live assurance at lower costs.

  4. Improve decision-making based on data from hundreds of thousands of historical projects.

     
     

• Workflow Process:

  SCS and nPlan work iteratively. SCS uploads the schedule to nPlan, which performs quality assurance (QA) and releases schedule forecasts. Insights generated help SCS mitigate risks and optimize decision-making.

  
  

• Results:

  • nPlan identified around 140 risk insights, potentially driving 250 days of delay at a cost of £120 million.

  • Utilized nPlan's visualization tools to quantify risks, improving reporting and risk management efforts.

  • Replaced traditional QSRA (Quantitative Schedule Risk Analysis) with AI-led Schedule Risk Analysis, reducing overheads and improving accuracy.

    
    

• Success Stories:

  • SCS successfully mitigated potential delays related to Park Royal Road sewer lining and River Pinn's RC deck installation using nPlan's insights.

  • SCS identified and avoided up to £9.5 million of additional costs at the Adelaide Road site by leveraging nPlan’s Driving Paths tool to quantify delay risks.

    
    

• Improved Decision-Making:

  • nPlan’s cost integration tool helps SCS forecast and control project costs more accurately, down to the activity level, increasing confidence in decision-making.

  • Regular AI-led forecasts allow SCS to manage scope, defer costs, and optimize resources for upcoming project milestones.

COMMENT

The power of Ai and its capabilities are truly amazing. A massive schedule database of global projects and Machine learning predicts likely forecast durations of your schedule, based on similar schedules.

The only issues I see are -

1. The necessity to upload potentially sensitive or 'Commercial in Confidence' information (project schedule) to nPlan's database. Some organisations may balk at this.

2. The cost - from the knowledge I have, it is a % of the Project cost. Of course, this can be negotiated.

3. Uploading your project data also improves nPlan's database, for improved Machine Learning. This also benefits nPlan. You are not able to uses nPlan, unless you upload your schedule.

4. Forecasting durations/delays assumes that your project scope/activity is 'similar' to what is in the nPlan database. For example, formwork durations on a nuclear power plant structure in a remote location (KSA for example) compared to formwork activities from hundreds of rail/road projects in the UK may skew results.

Definitely a different approach and methodology for understanding schedule risk and a preview of what's to come in the future.