ai-agents

Five program levels from FAQ assistant to multi-agent ops. Level 5 is optional. McKinsey found only 23% scaling in at least one function.

Score org-wide readiness out of 30. Below 16 out of 30, do not let it change orders. McKinsey found 62% experimenting and 23% scaling — still a minority per function.

Most store AI failures are IDs, APIs, ownership, and evals — not the model. Plan against a published ~$791/mo platform floor at 50K sessions — that is cost, not savings. Run the pre-model gate first.

Nine a.m. Slack is holds, stockouts, and missing GTINs. Pick one of ten back-office tasks this week. A person still signs every write — the sample refund ceiling is $75.

A shopper asks ChatGPT for a blue tee in medium. Adobe's July 2026 cohort says 39% of retail homepages still aren't machine-readable — and size, stock, and price live on the PDP, not the slogan. GEO is product facts, not a rank guarantee.

Human approval is a queue with order id and tool trace — not a prompt. The public worksheet uses a $75 refund ceiling above which a person must click Approve. Never auto-approve when the SLA expires.

LangChain composes Bedrock models, tools, and retrieval; AgentCore hosts production agents. Reuse the B2B CRM Gateway canary (~180→95 ms) and 50K-session ~$791/mo silhouette — do not give the model the warehouse.

LangGraph belongs on AgentCore Runtime, not Harness. Reuse the B2B CRM Gateway canary (~180→95 ms) and the 50K-session ~$791/mo platform silhouette — then make graph state explicit.

Your org chart wants five agents before one lookup works. If you are under ~8 tools and one audience, stop — one agent. The ~$791/mo platform cost floor at 50,000 sessions is one support-shaped mix, not five specialists.

Do not staff all five specialists in week one. An open-ended "handle this order" fans out to every agent. Hop cap 2. The ~$791/mo platform cost floor at 50,000 sessions is one support-shaped mix, not a fleet budget.

AgentCore Harness went GA June 17, 2026. Use n8n for process orchestration and @aws/n8n-nodes-agentcore to invoke Harness — pin the ARN in production. Reuse the Gateway ~180→95 ms canary; do not treat n8n as the agent runtime.

Store-connected agents fail at secrets, PII, and payment tools — not at prompts. Support-shaped pilots tracked roughly 3× spend when a browser tool stayed on for lookup turns. Delete payment tools.