A few years ago, I walked through the full case study of what “People, Process, then Technology” actually looks like in practice. It’s the Department of National Defence story: how asking “What time do orders come in?” led to mapping four different agents — technicians, suppliers, couriers, and customs — and why no technology solution could have surfaced the root cause.
Watch the video here:
MODEL 5 ASSESSMENT
A recent Hammerspace post caught my attention. They wrote:
“What’s really slowing down your AI pipeline? If you’re blaming the GPUs, you may be looking in the wrong place. With most AI projects derailed by data-readiness issues, the way you store and move data is often the real choke point.”
This prompted a question: Where do platforms like Hammerspace and Tealbook fit within an agent-based procurement architecture?
Tealbook:
Single strand focus: supplier data enrichment and consolidation
Creates a “source of truth” for supplier intelligence
Static data aggregation — who are your suppliers, what can they do, how do they perform
Essentially a supplier master data platform
Hammerspace:
Enterprise-wide data orchestration
Unifies unstructured data across multiple systems, locations, and protocols
Dynamic — data moves to where it’s needed, when it’s needed
Agent-integration capabilities (they specifically mention AI agent connectivity)
In the DND Model Architecture:
Recall the agents in that 1998 system:
Service technicians — creating orders (sandbagging at 4 PM)
The self-learning system — connecting all agents, weighting performance
The Critical Distinction:
Both Tealbook and Hammerspace are data plays. Neither is a behavioral play.
Tealbook would have given DND better supplier data — but wouldn’t have revealed the sandbagging pattern
Hammerspace would have unified the data from all agents into an accessible plane — but wouldn’t have asked the question that surfaced the root cause
The DND self-learning system did something neither does:
It connected agents behaviorally, not just informationally
It weighted performance based on delivery, quality, price, and geography
It learned from outcomes and adjusted
It was built on the insight that came from asking “What time do orders come in?” — a human question, not a data question
The Bottom Line:
Tealbook is a strand. Hammerspace is a data plane. Neither is a diagnostic.
The DND system achieved 97.3% delivery accuracy not from better data infrastructure — but from understanding agent behavior first, then building the system to support it.
Hammerspace would have accelerated the data access. Tealbook would have enriched the supplier strand. But neither would have told you that technicians were sandbagging orders because of incentive misalignment.
“What time do orders come in?”
No data platform would ask that question. That question delivered 97.3%.
Procurement Insights
Where Data Platforms Fit — And Where Behavioral Diagnosis Must Lead
Posted on December 8, 2025
0
A few years ago, I walked through the full case study of what “People, Process, then Technology” actually looks like in practice. It’s the Department of National Defence story: how asking “What time do orders come in?” led to mapping four different agents — technicians, suppliers, couriers, and customs — and why no technology solution could have surfaced the root cause.
Watch the video here:
MODEL 5 ASSESSMENT
A recent Hammerspace post caught my attention. They wrote:
This prompted a question: Where do platforms like Hammerspace and Tealbook fit within an agent-based procurement architecture?
Tealbook:
Hammerspace:
In the DND Model Architecture:
Recall the agents in that 1998 system:
The Critical Distinction:
Both Tealbook and Hammerspace are data plays. Neither is a behavioral play.
The DND self-learning system did something neither does:
The Bottom Line:
Tealbook is a strand. Hammerspace is a data plane. Neither is a diagnostic.
The DND system achieved 97.3% delivery accuracy not from better data infrastructure — but from understanding agent behavior first, then building the system to support it.
Hammerspace would have accelerated the data access. Tealbook would have enriched the supplier strand. But neither would have told you that technicians were sandbagging orders because of incentive misalignment.
“What time do orders come in?”
No data platform would ask that question. That question delivered 97.3%.
-30-
Share this:
Related