This article was originally published on LinkedIn.
Thirty years ago I coauthored a paper on quality management. Google Scholar currently lists 3,348 citations.
Ahire, S., D. Golhar, M.A. Waller. Development and Validation of TQM Implementation Constructs, Decision Sciences, 27(1), (1996), 23–56.
I didn’t write it thinking about combat operations. I should have.
Most current investment pours into sensors and AI, with far less going toward the quality architecture needed to make those tools reliable when it counts.
That is not just a data and AI challenge. It is a quality system challenge.
The Army needs speed. Speed makes quality non-negotiable.
Because in a contested environment:
If your demand signal is corrupted and nobody catches it, your resupply plan becomes a convoy chasing a ghost. If suppliers fail under surge, readiness collapses. If processes drift under pressure, rework happens under fire.
Those aren’t theoretical risks. They are the exact failure modes that turn “precision sustainment” into fragility.
The enemy does not have to destroy your logistics to defeat it. They can let friction do the work, if your system cannot trust its own data and processes.
In permissive environments, quality problems are expensive. They create waste, rework, and delays.
In contested sustainment, quality problems create signature, exposure, and missed windows.
Quality failure becomes survivability failure.
That is why quality management belongs in the contested logistics conversation.
The pivot most people skip
The core lesson from quality management is simple:
You do not inspect quality into a system after the fact. You design it in so it holds together under stress.
Precision sustainment is asking for machine-speed decisions across echelons. But speed without quality discipline simply accelerates error propagation.
Start with governance
In quality management, top management commitment is not symbolic. It determines whether quality is strategic or cosmetic.
In contested sustainment, the equivalent is governance.
Senior leaders must treat data integrity, allocation discipline, and supplier reliability as readiness drivers.
If the enterprise tolerates “good enough” data, AI will amplify noise rather than reduce it. If leaders treat data trust as mission-critical, the system gets designed differently from the start.
Define the customer correctly
Military logisticians have heard every version of the “four rights.” The idea is sound.
Here is the contested sustainment translation:
Quality is when a weapon system returns to mission-capable before the next fight. Quality is when casualty care does not pause because the supply chain guessed wrong.
A warehouse can look efficient on paper and still fail the customer.
In contested environments, the customer is not a spreadsheet. It is the unit that either keeps moving or stalls.
Everything upstream either serves that outcome or it is waste.
What a quality defect looks like when the enemy gets a vote
A brigade requests a repair part that should be in theater. The system shows it on-hand at an intermediate node. A convoy is routed. The maintenance platoon plans around that delivery window.
But the part was mislocated days earlier during a surge. It is physically present but not where the system thinks it is.
Now you do not just have a data error.
You have second-order effects:
The convoy ran for nothing. The unit missed a repair window. The equipment stayed non-mission capable longer. You generated signature and exposure correcting a mistake you did not know existed.
The failure did not happen at the point of need. It happened upstream, invisibly, because no one was managing quality at the process level.
AI didn’t cause it. AI amplified it.
Treat supplier quality as a combat function
The defense industrial base is not just an economic system. It is a combat power system.
Surge demand is real. Substitution is limited. Lead times are uncertain. Adversaries can disrupt supply and transportation simultaneously.
When a single supplier of a critical repair part cannot scale production in a major contingency, the effect does not stay in the supply chain. It moves forward.
Maintenance backlog becomes readiness gap. Readiness gap becomes lost maneuver options.
Supplier quality management in this context is not procurement administration.
It is resilience design.
Build resilience into process, monitoring, and people
This is where design quality, statistical process control, and employee empowerment converge and where most sustainment planning falls short.
A sustainment system that works in permissive conditions and collapses under denied conditions is poorly designed.
Design quality means asking:
What happens when connectivity drops? When data arrives late or corrupted? When a node disappears?
Once processes are in place, variation must be monitored before it becomes failure.
In manufacturing, statistical process control detects drift before catastrophe.
In contested sustainment, the same discipline applies. Monitor picking accuracy, cycle times, inventory location errors, supplier variability, and demand volatility before they cross thresholds that generate forward failure.
SPC here is early-warning radar for sustainment drift.
But monitoring assumes someone can act.
When communications degrade, centralized control collapses. The sustainment personnel closest to the problem must be trained, trusted, and empowered to flag suspect data, stop bad processes, and trigger local corrections within guardrails.
Feedback loops are how complex systems learn faster than they fail.
In a denied environment, the person at the point of friction is the only quality control left.
Quality culture becomes combat resilience.
Design for imperfection
Precision sustainment will not be achieved by AI alone.
AI accelerates decisions. Quality architecture determines whether those decisions are trustworthy under denial.
If we design sustainment systems that assume perfect data, perfect suppliers, and perfect communications, they will fail the moment those conditions disappear.
Quality is how you design for imperfection.
And in contested environments, imperfection is guaranteed.
If you are working in contested logistics, defense AI, or industrial base resilience, I would welcome your perspective.
