# Morning AI TL;DR — 2026-06-17

- **OpenAI launches Partner Network** — OpenAI’s new enterprise partner program is aimed at moving customers from pilots to production with consulting/SI support, which matters for builders selling AI implementation work. Sources: [OpenAI news](https://openai.com/news/), [Analytics Insight](https://www.analyticsinsight.net/news/openai-builds-global-ai-consultant-network-to-help-businesses-move-from-ai-ambition-to-results).
- **Anthropic publishes Claude Code usage research** — A privacy-preserving analysis of ~400,000 Claude Code sessions finds humans still make most planning decisions while Claude handles much of implementation, useful evidence for how agentic coding is actually being adopted. Source: [Anthropic](https://www.anthropic.com/research/claude-code-expertise).
- **Microsoft Work IQ APIs hit GA** — Microsoft says Work IQ APIs became generally available June 16, exposing Microsoft 365 context to custom agents in Copilot Studio, Foundry, and third-party platforms. Sources: [Microsoft Build blog](https://blogs.microsoft.com/blog/2026/06/02/microsoft-build-2026-be-yourself-at-work/), [Partner Center announcement](https://learn.microsoft.com/en-us/partner-center/announcements/2026-june).
- **Copilot Cowork is generally available** — Microsoft is positioning Cowork as a governed enterprise-agent layer for Microsoft 365 work, with cost/performance comparisons against external M365-connected agents. Source: [Microsoft 365 Blog](https://www.microsoft.com/en-us/microsoft-365/blog/2026/06/16/copilot-cowork-is-now-generally-available/).
- **OpenAI Codex had capacity/elevated-error issues** — Business Insider reported June 16 service strain for Codex, a reminder that coding-agent reliability and quotas are now production concerns, not just UX annoyances. Source: [Business Insider](https://www.businessinsider.com/openai-codex-elevated-errors-at-capacity-2026-6).
- **Anthropic export-control fight continues** — Reports say U.S. restrictions forced Anthropic to pull or limit its newest models for foreign access, sharpening the AI sovereignty risk for teams dependent on hosted frontier APIs. Sources: [AI News](https://www.artificialintelligence-news.com/news/anthropic-export-controls-ai-sovereignty/), [Semafor](https://www.semafor.com/article/06/13/2026/white-house-move-to-limit-anthropic-linked-to-concerns-about-chinese-access-to-mythos).

**Worth watching:** Enterprise AI is shifting from “which model?” to “which governed agent stack has reliable context, uptime, and policy insulation?”