When project change outruns project control.
Fluor reported $108 million of cost growth on legacy Urban Solutions projects in 2025. In the first quarter of 2026, one mining project added another $37 million impact.
Cost growth is recognized in finance. The operating signal usually appeared earlier—in a design revision, subcontractor response, field report, or schedule dependency.
Independent analysis of public information. Fluor is not a Fitzroy client.
A project loses margin one unresolved change at a time.
Large projects do not run in one system. Design revisions sit in document control, commitments in procurement, progress in field reports, forecasts in cost systems, and entitlement in correspondence. A change can be known everywhere and owned nowhere.
Fluor’s 2025 disclosures make the mechanism concrete: three infrastructure projects absorbed a $54 million net impact from subcontractor design errors, schedule effects, and price escalation. Full-year legacy-project cost growth reached $108 million.
The Fitzroy proposition is a governed change ledger across the existing project stack. It identifies the first conflicting signal, quantifies the affected work, assigns a commercial owner, and preserves the evidence behind every approved forecast movement.
- $108M
- Cost growth on legacy Urban Solutions projects reported for full-year 2025.
- $25.7B
- Backlog at March 31, 2026; 82% was reimbursable.
Give every material change a chain of custody.
The layer captures structured cost and schedule changes alongside controlled project documents, validates versions and identifiers, detects conflicting forecast signals, and prevents a material movement from disappearing between design, commercial, and field teams.
See the first movement.
MSK Connect captures cost and schedule changes while controlled files and extracted document facts preserve the commercial and design context around them.
Quantify before escalation.
Glue validates versions and project identifiers. Lambda applies explicit variance rules so the workflow carries affected scope, value, and schedule—not another unstructured alert.
Close the commercial loop.
Step Functions assigns design, project-controls, and commercial decisions. DynamoDB holds current state; Object Lock retains the approved record.
Price early control against late recovery.
Fluor’s reported project impacts demonstrate the pattern, but they are not treated as recoverable savings. The planning case below uses a separate, transparent portfolio model.
What the platform costs to run.
Modelled for 12 major projects: 2 million controlled documents, 50 million cost and schedule changes, 25 million variance executions, 40TB retained evidence, and 100 control users per month.
| Service | Basis | Monthly |
|---|---|---|
| Change and document ingress | MSK + Connect, Transfer Family | $3,000 |
| Document interpretation | Amazon Bedrock, controlled extraction | $2,400 |
| Quality and retained evidence | Glue Data Quality, S3 Object Lock | $1,700 |
| Variance and change control | Lambda, Step Functions, EventBridge | $1,500 |
| Ledger, queries, reporting | DynamoDB, Athena, QuickSight | $1,300 |
| Security and observability | CloudWatch, KMS, endpoints | $2,100 |
| Total | $12,000 / month |
What it is modelled to return.
Fitzroy planning case for a global engineering and construction portfolio; not a result reported by Fluor.
| Lever | Assumption | Annual |
|---|---|---|
| Design and scope variance | $250M affected work; 1% avoided | $2.5M |
| Schedule and controls effort | 50 FTE at $170K; 20% capacity released | $1.7M |
| Claims evidence and rework | $15M annual effort; 15% avoided | $2.3M |
| Annualized opportunity | $6.5M |
AWS list-price planning estimate as of August 2026; implementation and support excluded. The impact model is illustrative and is not a reported result for a company named on this page.
The project-control case, expressed as operating economics.
- First-year net
- $2.3M
- Benefit-to-cost
- 2.9×
- Annualized opportunity
- $6.5M
Fifty-five percent of annualized benefit in year one, less the $1.1M build case and $144K first-year run cost.
First-year gross benefit divided by implementation and first-year run cost.
Avoided exposure and released capacity—not revenue, guaranteed savings, or a result achieved for Fluor.
The annual run case is 2.2% of the modelled opportunity. The case depends on controlling ordinary project movements earlier—not assuming that any disclosed charge could have been eliminated.
Facing a similar project-control challenge?
Start with an Enterprise Systems Assessment.