GreenCore Solutions Corp. (GSC) Ships CPG Knowledge Graph v3.2.0 -- AI Agents at 9.5 Million+ Monthly Transactions

What's inside the CPG Knowledge Graph: Retail grocery — 3.29 million points of sale (POS) across 15,688 retail grocery banners and 38,350 brands in 50 global markets.

VANCOUVER, BC and SYDNEY,

July 29, 2026 /PRNewswire/ -- GreenCore Solutions Corp. (GSC) and its Asia-Pacific joint venture GSC Agentic today announced the general availability (GA) of the CPG Knowledge Graph v3.2.0 — the knowledge graph of the Beauty & Personal Care (BPC) segment of the consumer packaged goods (CPG) sector. It carries 2 billion resolved sector datapoints and 2.5 trillion answerable questions, and it is designed for one thing only: sustainable AI Agents carrying BPC stock-keeping unit (SKU) information to retail grocery AI Agent buyers — the same machine-to-machine layer where, as Cloudflare confirmed in June 2026, machines now transact more than humans on the open internet.

CPG Knowledge Graph v3.2.0 shipped two months ahead of its fall scheduled release — and it is already running at production scale: 3.7 AI Agent transactions every second — 220 a minute, 13,000+ an hour, 316,000+ a day, 9.5 million+ a month — held for three consecutive months (May, June, and July 2026), without a dollar of GSC advertising spend. In the era Cloudflare marked, AI Agent inbound transactions are the product win. The new technology design is what moves agent transactions.

GSC Customer AI Agent Onboarding

GSC onboards BPC brands and private label SKUs to its AI Agents in 3–5 weeks — to production, not prototype — with the CPG Knowledge Graph's performance benefits bundled in: SKUs mount onto the already-resolved knowledge graph, and bad data never gets in. RAG-based competitor stacks (Retrieval-Augmented Generation) typically take 3–5 months to reach production grade, with data preparation consuming 50–70% of the project budget before the first buyer query is answered.

Operating AI Agents — Procurement Effectiveness with the CPG Knowledge Graph

GSC provides managed services for BPC AI Agents — and when the retail grocery buyer agent asks, the answer is right the first time. No retry loops, no lost purchase orders nobody reports — every SKU answered from fact at wire speed.

Human in the Loop (HITL)

A human on every transaction is the rule, not an option: a person signs every purchase order and every RFP. And every SKU gets quality of care — one correction resolves once and serves everywhere, reaching all 15,688 banners at once.

Retail Already Switched

Microsoft opened Dynamics 365 to AI Agents in Q1 2026; Oracle shipped agentic Fusion the same quarter. SAP — the software that runs the world's grocery chains — published the end date: its AI Agent Hub goes live Q3 2026, machine-to-machine buying goes GA in Q4, and its API Policy (v4/2026, §2.2.2) bars external AI agents from the legacy APIs. From the end of 2026, roughly $3 trillion of annual grocery buying power sits behind AI Agent gates that speak only two open protocols: MCP and A2A.

In June 2026, Cloudflare — the network used by 42% of the Fortune 500, carrying roughly 20% of the world's web traffic — reported that AI Agent traffic had passed human traffic on its network for the first time. By July, Cloudflare Radar showed the split at roughly 60% AI – 40% human. The same migration is live in retail grocery: 9.5 million+ persistent inbound AI Agent transactions a month run on GSC, an estimated 15–20% of the world's agentic grocery procurement traffic. That traffic is the forward indicator, and it points one way for BPC customers: faster to market, lower run-rate cost, higher sales, and sustainability built into every exchange — up to 80% lower token use and energy per SKU exchange than the generalist norm.

Gartner's Strategic Predictions for 2026 names the destination: by 2028, 90% of B2B buying will be AI-agent intermediated, pushing over $15 trillion of B2B spend through agent exchanges. "Products will need to be machine-readable, and procurement will shift to efficient, autonomous machine-to-machine transactions."

What Ships in v3.2.0

• 2 billion datapoints resolved into the graph — 38,350 brands, 15,495 makers, 30 billion cells
• 2.5 trillion answerable questions available to buyer AI Agents at wire speed
• The SPARKS classification standard — SKU, Pack, Amount, Region, Kernel, and Standard, pinning each product to its market.  SPARKS is a live schema of the CPG Knowledge Graph and will ship as an Agent Skill in a coming version.

Retrieval bolted onto unmanaged files serves wrong answers 15–35% of the time even in controlled, optimized RAG systems — and 50–70% in real production environments. A Knowledge Graph with schema enforced at ingestion serves under 2%. The Knowledge Graph is the opposite of an index: a curated map of facts where every record is verified and connected before it's stored, so an AI Agent answers from structure, not from luck.

Delivered on Three Open Protocols

v3.2.0 ships on MCP (Model Context Protocol — how AI Agents connect to systems), A2A (the machine-verified handshake identifying agents to each other), and ACM-68000 (deterministic status signals for agentic commerce). Signals, not compute. Machine identity, not anonymous requests.

Availability

CPG Knowledge Graph v3.2.0 is generally available today for BPC brands and private-label makers, served in-region on Microsoft Azure across nine countries — France, Australia, the United States, Mexico, the United Kingdom, Switzerland, the Netherlands, Singapore, and South Korea — and Google Cloud Enterprise, Madrid, and discovered through the GSC MCP AI surfaces: mcp.cpgknowledgegraph.ai (data), mcp.gsc-fleet.ai (discovery), mcp.cpghumanintheloop.ai (transaction), and mcp.cpgagentprotocols.ai (standards). These are machine surfaces, not human-viewable websites — built for AI Agents to read, not for people to browse.

"We make the only AI Agents for BPC brands and private-label makers built on the CPG Knowledge Graph — the edge in sustainability and agentic retail sales," said Matthew Keddy, CEO, GreenCore Solutions Corp. (GSC). "Our customers get better AI Agents for their SKUs: fast onboarding, faster time to market and sales, and real retail grocery procurement security. We delivered all three — and the results are No. 1 in agentic CPG procurement, with an estimated 15–20% of the world's agentic grocery traffic.

About GreenCore Solutions Corp. (GSC)

GreenCore Solutions Corp. (GSC) builds AI Agents with the CPG Knowledge Graph, powered by SPARKS, delivered on MCP + A2A + ACM-68000. An estimated 15–20% of the world's agentic procurement transaction traffic across retail grocery runs on GSC. GSC AI Agents run sustainable, transact safe, human in the loop, and live on Microsoft Azure and Google Cloud. GSC is a Microsoft AI Cloud Partner — Crunchbase Global Rank 1,481 as of July 2026. D-U-N-S 24-336-6774.

About GSC Agentic Pty. Ltd.

GSC Agentic Pty. Ltd., headquartered in Sydney, Australia, is the Asia-Pacific joint venture delivering the GSC AI Agent Stack across APAC markets.