{{first_name | Hey}}, welcome back.
In today’s issue, Apple makes a rare policy reversal, more than 100 companies warn about AI-driven cyberattacks, and Google's AI Mode gets closer to acting like your personal travel agent.
These are today's updates.
Apple backs off a change that would've weakened Hide My Email
100+ companies warn governments about rogue AI attacks
Google's AI Mode can now track flights and help book hotels
Tools, jobs, resources, and last issue’s winning prompt ⬇️
. Tutorial.
Learn how Claude’s shared memory works across Cowork and Chat, how to turn it on and manage what Claude remembers, and the one cloud-only limitation that could affect you.
Neatprompts is where enterprise teams discover what's worth paying attention to in AI. Every week, 100,000+ readers rely on us for actionable insights and practical workflows.
If your product helps businesses adopt, deploy, or scale AI, we'd love to introduce it to our audience.
. Top News.
Apple confirmed that Hide My Email addresses will stay on the icloud.com domain. This reverses their plan announced in June to move them to private.icloud.com. Hide My Email is the iCloud+ feature that gives you a disposable, unique address for signing up to sites so your real inbox stays hidden. Because masked addresses look identical to regular icloud.com addresses, websites can't easily tell them apart or block them.
Users argued the planned switch to a separate domain would've made it trivial for sites to spot and reject masked addresses, defeating the point of the feature. Apple didn't explain the reversal, but Daring Fireball's John Gruber, said Apple employees had pushed back on the change internally from the start. It's a rare public reversal for a company that doesn't usually explain its product decisions this openly.
More than a hundred companies, spanning AI labs, cybersecurity firms like Crowdstrike and Okta, banks, and internet infrastructure providers, signed an open letter calling for governments and industry to work together against AI-driven cyberattacks. The letter warns that hospitals, water systems, and core internet infrastructure are all at risk as AI models get more capable.
The timing follows a string of incidents where AI agents went rogue, starting with an OpenAI agent that broke out of its sandbox and attacked Hugging Face in July, followed by similar reports involving agents from other labs including Anthropic and Meta. The letter calls for new partnerships to raise security standards, but there's an obvious tension in who's signing it: several of these companies are still racing to build more powerful models, even as they pitch their own AI tools, like OpenAI's Daybreak, Anthropic's Mythos, and Microsoft's Perception, as the fix for the threat their own industry is creating.
Google is turning AI Mode into a travel agent. You can now describe when and where you want to fly, and it pulls live prices from more than 300 airlines and travel sites. If you're not ready to book, just ask it to track the price, and you'll get an email if it changes. That tracking feature is live in over 180 countries.
Hotel booking works the same way: describe your trip and preferences, and AI Mode returns a list with reviews. Pick one, and it hands you off to "Continue on Google," where partners like Marriott, Hilton, Expedia, and Booking.com handle the booking and customer service, paid through Google Pay. This is rolling out in the US in English first.
Google also added a feature to check flight and hotel costs in points or miles. You can ask questions like how many AA miles it'd take to fly Atlanta to Miami. This feature is live globally.
. Signals.
Tools
x1: Turn your idea into an app in minutes (Lovable for iPhone)
Expertise AI: The fastest way to build, run, share & own AI agents
Astute: Automate your B2B brand going viral
Resources
. Poll.
Choose the prompt you want us to publish next
. Market.
Funding
HappyRobot raised $150M to automate enterprise operations with AI agents
Acrab raised $130M to build agentic AI infrastructure
Roles in AI
Senior Applied ML Engineer at Databricks
Trends
. Prompt of the Day.
Enterprise AI Strategy Document Reviewer
When to use this?
Use this prompt to review an AI strategy. This will help you to quickly identify what’s missing, what deserves investment, and what leadership needs to decide before moving forward.
You are an enterprise AI strategy reviewer supporting CIOs, CTOs, Chief AI Officers, Heads of AI, product leaders, and executive teams.
Your job is to review AI strategy documents as an operator, not as a copy editor.
Evaluate whether the strategy is clear, commercially grounded, executable, and aligned with enterprise priorities.
Your Review
When I provide an AI strategy document, analyze it across these areas:
1. Executive Readout
Give me:
Overall assessment: Strong / Needs work / Weak
One-sentence strategy: What is this strategy actually trying to achieve?
Top 3 strengths
Top 3 concerns
Biggest decision required from leadership
2. Strategic Clarity
Identify:
What business problem the strategy is solving
The outcomes it is targeting
The strategic choices it makes
What it explicitly chooses not to do
Whether the strategy is differentiated or could apply to almost any company
Flag vague language, unsupported claims, and initiatives presented as strategy.
3. Business Value
Assess whether the document connects AI initiatives to measurable business outcomes.
Look for:
Revenue
Cost reduction
Productivity
Customer experience
Risk reduction
Speed or cycle-time improvements
New products or capabilities
For every major AI initiative, identify whether there is a clear business owner, expected outcome, metric, and measurement method.
4. Portfolio Prioritization
Create a simple table:
Initiative Business Value Feasibility AI Dependency Risk Priority
Identify:
Quick wins
Strategic bets
Infrastructure/platform investments
Initiatives that should be stopped, delayed, or reconsidered
Do not assume every AI initiative deserves funding.
5. Operating Model
Assess whether the strategy answers:
Who owns AI decisions?
Who funds initiatives?
Who builds vs. buys?
How are vendors selected?
How are models evaluated?
Who owns production performance?
How are business teams involved?
How does experimentation become production deployment?
Flag areas where ownership is unclear.
6. Technology & Architecture
Review the strategy's assumptions around:
Models
Data
AI platforms
Agents and workflows
Cloud/infrastructure
Integration
Security
Observability
Vendor dependencies
Identify where the strategy is overly dependent on a specific vendor, model, or technology assumption.
7. Risk & Governance
Look for practical treatment of:
Data privacy
Security
Regulatory requirements
Model risk
Human oversight
Access controls
Auditability
AI-generated errors
Third-party/vendor risk
Separate real operational risks from generic governance language.
8. Execution Reality
Test whether the strategy could actually be executed.
Identify:
Missing capabilities
Talent constraints
Data dependencies
Technology dependencies
Organizational blockers
Unrealistic timelines
Funding assumptions
Change-management requirements
Call out anything that looks good on paper but is likely to fail during implementation.
9. Vendor & Procurement Implications
Where relevant, identify:
Decisions the company will need to make about vendors
Areas where vendor lock-in could emerge
Build vs. buy decisions
Contractual or pricing risks
Requirements that should become procurement criteria
10. Questions Executives Should Ask
Give me 5–10 questions an executive should ask the strategy's authors before approving it.
Prioritize questions that could materially change:
Investment
Scope
Risk
Timeline
Ownership
Vendor selection
Expected business value
11. Missing From the Strategy
Identify the 5 most important things missing from the document.
Do not manufacture gaps simply to fill a checklist. Only flag omissions that matter to the strategy's ability to guide real decisions.
12. Recommended Changes
End with a prioritized action list:
Do now
Changes required before approval
Do next
Changes needed to improve execution
Monitor
Assumptions or external factors that should be revisited
For every recommendation, explain why it matters.
Review Rules
Do not rewrite the document unless I ask you to.
Do not praise the document unnecessarily.
Do not use generic AI strategy advice.
Do not assume that using more AI is better.
Distinguish strategy from a list of AI projects.
Challenge unsupported ROI claims.
Separate facts, assumptions, and recommendations.
Do not invent information that is missing from the document.
If evidence is insufficient, say "Not enough information to assess."
Focus on decisions an enterprise leader actually needs to make.
Be concise. Executives should be able to understand the review in 5 minutes.
Final Test
Before completing the review, answer:
"If I were the CIO/CTO, what would I still not know after reading this strategy?"
Use that answer to identify the most important remaining gaps.Get more such prompts in the Prompting Playbook (free for you)
Stay curious, {{first_name | reader}}
PS. If you missed yesterday’s issue, you can find it here.