Herne Credits Nodes & Links AI Scheduling With £14m Value

Herne Credits Nodes & Links AI Scheduling With £14m Value
Photo by Scott Graham on Unsplash

UK project consultancy Herne Group says it has generated £14m in what it calls acceleration value by applying Nodes & Links AI scheduling across its projects. The figure, reported by Construction Management on November 19, 2026, comes with a detailed account of a defence project where a client scope change threatened a fixed handover date.

According to Herne and Nodes & Links, the key numbers on that project were:

  • A 1,000-iteration probabilistic risk analysis put the chance of hitting the planned completion date at 0.2%.
  • Delay was costing roughly £80,000 per day.
  • The platform's first pass found scenarios worth up to 39 calendar days of potential recovery.
  • Implemented measures delivered 15 days of improvement by the following schedule update.
  • Within three months, the P50 completion forecast moved 10 weeks earlier.
  • Herne credits the work with eight weeks of recovery and more than £4m in acceleration-related savings.

How Herne used Nodes & Links AI scheduling

Herne queried its live programme for ways to claw back the slipped date. The software modelled options such as longer working hours, fast-tracking, revised logic links, earlier documentation and resequenced activities. The project team then filtered these against cost, resources, sequencing and client needs.

The top option was shifting selected critical commissioning work from a five-day to a seven-day week, which the client approved. Herne repeated the analysis at each update to track results and find further gains. Project controls director Andy Garnett said no other tool lets the team interrogate a live programme this way.

Our take: what to check before relying on AI schedule recovery

Where it fits

Nodes & Links sits on top of scheduling tools rather than replacing them. Planners still build programmes in Primavera P6, Asta Powerproject or Microsoft Project, and the AI layer interrogates that data. It competes for attention with other AI scheduling and forecasting products such as nPlan and ALICE Technologies, each with a different emphasis on risk forecasting versus generating alternative sequences.

The most useful detail here is not the headline £14m but the workflow: options were modelled, screened by people against real constraints, then re-tested at every update. That human filtering step is what turned a list of scenarios into something a defence client would sign off.

Questions the numbers leave open

Acceleration value is Herne's own metric, and it has not explained how it is calculated or whether it nets off the cost of the interventions. Moving commissioning to seven-day working carries real labour and supervision costs. The reported figures also vary, with eight weeks of recovery cited alongside a 10-week shift in the P50 forecast, which may reflect different measurement points.

What AEC teams should check

Output quality depends on the underlying programme. Firms with weak logic, open ends or poorly loaded resources will likely need cleanup before AI recommendations mean much. Buyers should also ask how auditable the reasoning is, since Herne leaned on that to satisfy assurance reviewers. Pricing and deployment terms were not disclosed. For contractors and consultants facing delay exposure on fixed-date programmes, a pilot on a live recovery problem could be the fairest test.

Ask Ace