Tagged
RCS for Business shipped 31 March 2026. A US Rich text segment lists at $0.01194 on delivery. A rich card lists at $0.02949. CLI 2.35.12 is the first release with send-rcs-message.
Since 1 July 2025 WhatsApp on AWS is billed per message. The AWS fee is $0.005 outbound and $0.001 inbound. India Authentication and Utility are $0.002. Meta's fee is extra.
Ten real lookups and three must-escalate cases, re-run before every release. A green latency chart is not correctness. Below 16 out of 30 readiness, do not add a write while you tune the set.
Send one workflow, a deny-list, and a pass bar. The fill-in RFP is 14 sections. Below 16 out of 30 readiness, the correct award is do not build.
Score the vendor on 11 lines before a pilot. Below 16 out of 30 on your own readiness, do not let any vendor attach a write. About $791 a month at 50,000 sessions is a platform floor, not proof the vendor is cheap.
There is no universal price. A July 2026 first-party benchmark put one AgentCore support silhouette near $791 a month at 50,000 sessions — platform plus a small model, before integration or a reviewer.
On 23 Sep 2026 AWS added Salesforce and Zendesk connectors. The managed Salesforce crawl is 4 entity flags, not the classic preview's 21 CRM objects. New customer-managed Salesforce connectors stop on 30 Sep 2026.
On September 21, 2026 the Strands Agents team reported 28% lower token cost across six benchmarks, and 77% less than Claude Code with Fable 5 on Terminal-Bench. That figure is not an AgentCore bill.
Managed Knowledge Base Web Crawler, GA with the service on 17 Jun 2026, follows up to 300 URLs per minute from 10 seeds or 3 sitemaps. A 10,000-URL help center is at least 34 minutes at that cap, and the crawler still has no per-page ACL.
On September 22, 2026 AWS shipped Claude Opus 5.5 — 1M-token context, thinking always on, ZDR by default on Bedrock. Anthropic lists $4/$20 per million tokens. Here is the routing call and what breaks from Opus 5.
On September 22, 2026 AWS made GPT-6 Sol and GPT-6 Luna generally available on Bedrock. OpenAI’s internal eval found Sol makes about half as many factual mistakes as GPT-5.6 Sol. Here is the lane routing.
Assemble the hold reason, aging, and promise-to-pay. Finance releases the order. The agent does not. The account-manager list still caps at 5 priorities — a hold is one of those 5, not a button.