AI Gets a Business Operating System, Google Wants You to Optimize for AI, and Journalism's Burnout Crisis Gets Real Help

ANTHROPIC'S SMALL BUSINESS PLAY

Claude now ships with 31 ready-to-run business skills

The Rundown: Anthropic released what amounts to an AI operations team for small businesses — 31 pre-built skills that connect to your actual business tools and run real workflows.

The details:

• Skills span six categories: Finance, Sales/CRM, Marketing, Customer Support, Briefings, and Setup/Hiring/Legal
• The /content-strategy command generates a 30-day content roadmap; /canva-creator produces assets with captions already written
• Requires installing a 'Small Business' plugin through Anthropic's Cowork platform with 12 tool connectors including QuickBooks and HubSpot
• An onboarding command interviews you about your business and saves that context so every skill gets smarter over time
• Slash commands like /friday-brief and /invoice-chase trigger specific workflows — not conversations, actual task execution

Why it matters for us: For a small newsroom running on three people and a prayer, the /content-strategy and /canva-creator commands could legitimately cut your social media planning from half a day to half an hour. The catch? You'll need to test whether 'small business' assumptions translate to 'newspaper operations' — most of these skills assume you're selling products, not publishing news.

GOOGLE'S NEW PLAYBOOK

Google tells publishers exactly how to show up in AI answers

The Rundown: Google published official documentation on optimizing web content for AI-powered search — basically admitting that AI optimization is now a separate discipline from traditional SEO.

The details:

• The guide covers how AI systems like Search Generative Experience discover, process, and cite web content
• Technical recommendations include structured data, clear content hierarchy, and machine-readable formats
• Publishers can control AI access through robots.txt directives and specific meta tags
• The documentation addresses both traditional search rankings and AI-generated answer citations as distinct concerns
• This is Google's first explicit acknowledgment that optimizing for AI is becoming its own field

Why it matters for us: When someone in your county asks an AI "what's happening with the school bond referendum," your reporting should be what gets cited — not a stale PDF from the district website. This guide is worth an hour of your web person's time this week, honestly.

Read at developers.google.com →

THE BURNOUT RECKONING

New mental health helpline launches specifically for journalists

The Rundown: The Media Resilience Network launched a free, confidential, encrypted helpline for journalists — built on the insight that most of us need support after the story is filed, not during.

The details:

• A 2024 Reynolds Journalism Institute report found 80% of journalists cite burnout as a critical issue
• Women, gender-diverse, Latine, and Black journalists face disproportionate burnout due to discrimination, harassment, and caregiving burdens
• Experts recommend four longer shifts with real days off between them rather than continuous partial coverage during breaking news
• The Globe and Mail's 'happiness beat' focused on solutions journalism saw 43% year-over-year unique visitor growth
• Freelance journalists face particular vulnerability — no institutional support, no hazard pay, no follow-up care

Why it matters for us: If you're running a small newsroom, you probably can't afford to hire a wellness coordinator. But you can restructure surge coverage, create a safe channel for venting during sensitive stories, and explicitly permit breaks during deadline nights. Low-cost changes. Real impact.

Read at editorandpublisher.com →

THE COPYRIGHT CLARITY

Jane Friedman publishes the AI-in-publishing FAQ we all needed

The Rundown: A comprehensive FAQ tackles the questions writers (and newsroom staff) actually have about AI — from copyright protection to disclosure requirements to what's actually safe to paste into a chatbot.

The details:

• Purely AI-generated content doesn't get copyright protection, but human-authored work with AI assistance may still qualify depending on creative input level
• Major publishers are increasingly requiring AI disclosure in contracts — transparency is becoming an industry expectation
• Writers should be cautious about inputting unpublished work into AI systems due to potential data training concerns
• AI can legitimately assist with brainstorming, editing, research, and marketing without compromising authorship
• Professional reputation risks exist for authors who use AI without disclosure, especially as detection tools improve

Why it matters for us: Use this FAQ as a template. Your newsroom probably doesn't have a written AI policy yet — and honestly, you need one before someone pastes a sensitive source's quotes into ChatGPT. This gives you a framework to work from.

Read at janefriedman.com →

AUTONOMOUS AGENTS ARRIVE

The /goal feature lets AI run tasks for hours — or days — without you

The Rundown: Both Codex and Hermes now support a /goal feature that sets long-running autonomous tasks. You tell it what to accomplish, and it works in the background.

The details:

• The /goal command enables AI agents that run autonomously for extended periods, not just single conversations
• Users can prompt the AI to suggest goals based on accumulated context about their work patterns
• Once a goal is set, the AI reportedly works for hours or days toward completion
• The approach leverages conversation history to personalize what the agent works on

Why it matters for us: Here's where my skepticism kicks in. The promise is compelling — imagine an agent that monitors county commission agendas every week and flags newsworthy items. But I haven't tested this in a newsroom context, and "works for days autonomously" is exactly the kind of claim that sounds better in a tweet than in practice. Worth watching, not worth betting on yet.