Product
The product.
One loop, run by agents: observe, answer, show the reasoning, keep testing.
A standing view
Agents re-read capacity, operators, demand and catchments on a schedule. The picture is current when you open it.
Ask in plain words
The agent picks the tools, assembles the case and resolves the site to build, upgrade or walk away. The reasoning is the output.
The case holds. The capacity does not.
Two operators added capacity nearby and the weekday peak moved 90 minutes later. Utilisation is flat; the shortfall sits in hours this site cannot cover.
Competing capacity · 90 d · market.observe
+12 units within 4 km
Peak movement · 6 mo · peak.profile
Peak +90 min later
Catchment mix · 12 mo · catchment.compare
Through-traffic up
Evidence behind this decision| Evidence | Change | Window | Tool |
|---|
| Competing capacity | +12 units within 4 km | 90 d | market.observe |
|---|
| Peak movement | Peak +90 min later | 6 mo | peak.profile |
|---|
| Catchment mix | Through-traffic up | 12 mo | catchment.compare |
|---|
Illustrative example.
Cases are re-tested
An approved case is re-run as the market moves. When one stops holding, the agent says so, with the evidence that changed it.
Who it is for
- Network development
- Ranking and defending the next sites, with the reasoning attached.
- Investment committees
- Evidence that survives challenge in the room.
- Strategy and market intelligence
- Knowing the market moved before the quarterly review does.
Physical infrastructure
The method applies wherever capacity is fixed and demand moves. The same tools are open to your agents over MCP.