AI Office Automation Daily Briefing — June 10, 2026

Covering the latest developments from DingTalk, Lark/Feishu, WeCom, Microsoft 365, Google Workspace, and the AI Agent landscape.

1. DingTalk

1.1 Alibaba Employee Posts 70,000-Word Postmortem on DingTalk ONE Project Before Leaving

On June 5, 2026, a 70,000-word departure postmortem on the DingTalk ONE project was published on Alibaba internal network. The author witnessed the entire lifecycle of this flagship AI product. ONE was designed to proactively aggregate work messages, peaking at ~3 million DAU. The project contracted due to multiple contradictions: shifting positioning, short-video card designs ill-suited for offices, employee pushback against auto-read receipts. Under a high-pressure one-release-per-day cycle, the team prioritized surface features while core infrastructure remained underdeveloped. The project was absorbed into Wukong. 1

1.2 DingTalk Integrates Multiple Mainstream AI Models

DingTalk has integrated Sora2, NanoBanana, Gemini 3 Pro, and ChatGPT, supporting video generation, image editing, and 3D particle system creation. Tests show DingTalk AI performs on par with original models. Sora2 videos feature coherent plots; NanoBanana performs precise image modifications. DingTalk AI also offers enterprise knowledge Q&A. 2

2. Lark / Feishu

2.1 Feishu 2026 Spring Launch Introduces aily Smart Partner

Feishu launched aily (Smart Partner) in a rare full-demo format. aily activates in under a minute, accessing user documents, calendars and chat history upon authorization. Unlike OpenClaw, aily securely retrieves context without exposing data in group chats. All major domestic LLM providers have completed Feishu OpenClaw plugin integration. 3

3. WeCom

3.1 WeCom CLI Goes Open Source

Tencent released wecom-cli as open-source, exposing seven core office capabilities to AI Agents including Claude Code and Codex. Supports one-click meeting creation and message summarization. 4

3.2 Weisheng Launches Claw AI Agent for WeCom

Claw AI Agent embeds natively into WeCom with scheduled task dispatch and 100% local data deployment. Employee efficiency improved by over 50%, setup takes 5 minutes. 5

4. Microsoft 365 / Copilot

4.1 Microsoft 365 Copilot Gets Major Redesign

Microsoft acknowledged the forced floating Copilot button disrupted workflows, allowing users to dock it to toolbar or sidebar. Paid adoption rate: 3.3%. Edge Copilot now supports cross-tab info aggregation. Microsoft ended OpenAI exclusivity, introducing Claude model in Office. 6

5. Google Workspace / Gemini

5.1 Google I/O 2026 Unveils Major Workspace Upgrades

Google announced new Gmail Live for inbox queries, Docs Live for voice brainstorming, and Google Pics for image editing with Nano Banana. 7

5.2 Gemini Upgraded Across Docs, Sheets, Slides

Gemini gains near-human spreadsheet capabilities with auto-fill and web search for missing data. Rolling out globally in spring. 8

6. AI Agent Landscape

6.1 Wukong AI: 5x Efficiency Gains in Commercial Deployment

Zhejiang Youkela deployed Wukong in 2 weeks, boosting efficiency 5x and revenue 30%. Key apps: 10-min analysis of 5000+ reviews raised launch success from 60% to 92%; real-time Arabic/Spanish transcription. Suzhou Guangxian imported 1M records into Wukong for NL query analysis. Four enterprises signed strategic agreements at 2026 DingTalk Summit. 9

6.2 WorkBuddy Enterprise Edition Launches

Tencent Cloud launched WorkBuddy Enterprise Edition with 23 partners. WorkBuddy is China largest desktop office agent. Per-user token consumption grew 10x in 3 months. 10 11

6.3 Anthropic Expects $10.9B Q2 2026 Revenue

Anthropic expected Q2 2026 revenue of $10.9B with ~$559M operating profit. Claude Code at $2.5B annualized revenue holds 54% market share in AI coding tools. 4% of GitHub public commits by Claude Code. 8 of Fortune 10 are Claude customers; KPMG deployed across 276K employees. 12

6.4 AI Agent at Mid-2026: From 88% Failure to 54% Production

88% of pilot projects never reach production, but 54% of enterprises have deployed AI Agents in core ops. Median payback: 5.1 months. Gartner predicts 40% of enterprise apps will integrate task-specific AI Agents by end of 2026. 13

References

1. https://3g.china.com/act/news/10000169/20260607/49534493.html
2. https://post.m.smzdm.com/zz/p/aqkl25xp/
3. https://new.qq.com/rain/a/20260319A084CV00
4. https://www.toutiao.com/article/7645106542434386470/
5. https://college.wshoto.com/a/311651.html
6. https://www.toutiao.com/article/7645181150839570959/
7. http://view.inews.qq.com/k/20260521A002IC00
8. https://m.zhiding.cn/article/3180813.htm
9. https://caifuhao.eastmoney.com/news/20260528120131813296660
10. http://www.dsb.cn/p/01ktgpw0b94f47gvmdtyzm13w4
11. https://m.163.com/dy/article/KUTSJMM60511AQHO.html
12. https://www.toutiao.com/article/7644485192337326627/
13. https://www.hlxxi.com/ai-agent-tech/ai-agent-production-mid-2026-turning-point/