Why Canary on Counterpoint

The one-sentence version

Counterpoint is the transaction engine. Canary is the agent-native operating layer that makes a Counterpoint installation intelligent — accessible to both the operator (analytics, LP, OTB) and the customer's own AI agent (via ALX over MCP).

Two surfaces, one product

Analytics spine: 13 ARTS modules that turn Counterpoint's Document history into operational intelligence — loss prevention, distribution recommendations, demand forecasting, OTB enforcement, commercial account risk. The operator surface.

ALX / VSM: The store's MCP endpoint. Answers customer Claudes in real time. Diagnoses plant symptoms, checks inventory, settles transactions to account. The customer surface. See Agent Network.

What Counterpoint doesn't surface

Gap Canary module / agent
Loss prevention detection — rules over the audit log Q + Fox (case management)
Multi-store transfer loss reconciliation D
Demand forecasting and OTB enforcement J
B2B commercial account risk scoring C
Multi-store distribution recommendations D
Analytics via natural language Owl
Customer agent interface (MCP) ALX / VSM
Cross-store device heartbeat A (Module A) + back office hub

The co-sell story for VARs

You already own the Counterpoint relationship. Your customers trust you for implementation, training, and support. Canary adds the enterprise analytics layer AND the agent interface — the things your customers couldn't build themselves, available through you.

No additional hardware. No change to the POS workflow. Canary reads the API and adds intelligence above it.

The market moment

Stat Source
$3–5T agent-mediated consumer commerce by 2030 McKinsey
$15T B2B purchasing via agents by 2028 Gartner
45% of consumers already use AI during buying journeys IBM/NRF (18,000 respondents, 23 countries, Q3 2025)
97M MCP SDK downloads per month March 2026

Shopify launched MCP support in Summer 2025. Walmart built Sparky/Associate/Marty on the same principles. Google + Shopify launched UCP at NRF Jan 2026.

Why NCR won't build this

NCR's batch-first architecture cannot deliver the millisecond responses MCP requires. Leadership is financial (not AI-native). The VAR channel model means they'd have to cannibalise the partners who reach mid-market retail if they tried.

NCR cannot control what VARs build on top of Counterpoint. The agent layer is yours to build.

Full analysis: NCR Context

The opportunity window

NCR is in a financial transition (stock -40%, revenue -13 to -18% projected 2026). The garden center vertical carries 130+ years of domain knowledge that has not been machine-readable — the Armstrong proof case makes this concrete.

Layer 4 in the Counterpoint VAR channel is the place where the agent interface gets built; the substrate already exists.