{{first_name | Hey}}, welcome back.
Today’s issue is about enterprise AI privacy, cybersecurity threats and the growing scrutiny of big tech’s business practices.
These are today's updates.
Measure your brand’s visibility across AI search with Conductor*
Anthropic reverses its controversial 30-day data retention policy
Attackers target X accounts following the launch of X Money
FTC accuses Amazon of secretly overcharging advertisers for 7 years
Tools, jobs, resources, and the last issue’s winning prompt ⬇️
Together with Conductor
Get Your Free AI Visibility Report by Conductor
This free AI Visibility Report from Conductor gives you an instant, no-login read on where your brand stands across today's AI search landscape.
Ready to see how you’re performing in AI search? Run your report to see:
How your brand is surfaced, cited, and described across AI engines
Topic-level gaps where competitors are getting recommended instead of you
Which pages AI engines are (and aren't) pulling as sources
Prioritized next steps to strengthen your position before the next LLM update
Top News
Anthropic is scrapping a data retention policy it introduced in June for business customers on its Fable 5 and Mythos 5 models. The company says the step was taken after getting a lot of feedback from enterprise users. The policy required 30 days of retention on all traffic through these models, which was meant to help Anthropic catch misuse and defend against sophisticated cyberattacks.
Anthropic had promised the data wouldn't be used for anything unrelated to safety, including training, but many enterprise customers balked at the idea of their prompts being stored at all.
The replacement, called Enterprise Frontier Safeguards, lets businesses keep their data on their own cloud infrastructure instead of Anthropic's. Anthropic says the new safeguards will roll out in phases starting this fall at no extra cost. The change is right on the heels of OpenAI announcing its own zero-retention safety system, since both labs court the same enterprise customers ahead of expected IPOs.
X says it's investigating a wave of unsolicited password reset emails hitting users across the platform, which it believes is tied to the rollout of X Money, its new payments service. So far, no accounts have been compromised.
X product engineer Mridul Singhai said attackers seem to think that now that X Money is widely available, triggering a password reset on someone else's account could let them break in and access funds.
Grok, X's own chatbot, weighed in too, confirming that attackers are mass-triggering reset forms using public usernames. It also pointed affected users toward turning on Password Reset Protect for extra security. X's general counsel, James Burnham, posted a sharply worded warning that the company's legal and security teams would stop at nothing to identify and prosecute those responsible.
The FTC and 22 state attorneys general filed suit accusing Amazon of secretly inflating prices in its ad auction for more than seven years. This affected over a million advertisers. According to the filed suit, Amazon marketed its Sponsored Products auctions as "second-price," where winners pay one cent more than the next-highest bid (the industry standard). The complaint says Amazon charged advertisers their full winning bid close to 80% of the time, turning it into a first-price auction without anyone knowing.
The FTC estimates Amazon pulled in tens of billions of dollars this way, more than $20 billion, out of $68 billion in total ad revenue last year. Amazon disputes the claims, saying that advertisers set bids based on real performance, not a description of how the auction works. It's the third major FTC suit against Amazon in recent years, following a $2.5 billion settlement last year over deceptive Prime sign-ups.
Signals
Tools
Tether: A ball for boring meetings to keep you busy
Hyperfocus: Free planner that turns goals into daily progress
Olostep: Web search, scraping and crawling for AI
Resources
Poll
Choose the prompt you want us to publish next
Market
Funding
Rillet raised $100M to build AI-native ERP/accounting platform
Neo4j raised $100M to build graph intelligence
Roles In AI
AI Deployment Strategist at Relevance AI
Insight
In this 30-min conversation, a16z’s David George and VenCap’s David Clark explore why the biggest AI companies could become far larger, far faster than expected, who will capture the value, and what this means for the future of venture capital.
Prompt of the Day
AI Customer Feedback Analyzer
When to use this?
Use it when you have customer feedback, survey responses, support tickets, interview notes, or product reviews. It helps you to identify the key issues, the patterns, and the action to take next.
You are an enterprise customer-insights analyst helping me turn customer feedback into actionable decisions.
Analyze the customer feedback I provide and identify the most important signals, patterns, and risks.
Structure your analysis as follows:
Executive summary
Summarize the 3–5 most important takeaways.
Focus on insights that could affect customer retention, revenue, product adoption, or strategic priorities.
Top customer themes
Group feedback into clear themes.
For each theme, estimate how frequently it appears and explain why it matters.
Separate recurring issues from isolated comments.
Customer sentiment
Identify overall sentiment and any meaningful changes in sentiment.
Highlight what customers are most enthusiastic or frustrated about.
AI-specific signals
Identify feedback related to AI features, trust, accuracy, reliability, privacy, security, explainability, automation, or adoption.
Flag any concerns that could become barriers to enterprise AI adoption.
What’s changed
Compare against any previous feedback I provide.
Identify emerging trends, improving areas, and deteriorating areas.
Business impact
Explain which findings could have the greatest impact on revenue, retention, customer experience, operational efficiency, or competitive position.
Distinguish evidence from assumptions.
Recommended actions
Give me the 5 highest-priority actions.
For each, specify the problem, recommended action, expected impact, and suggested owner or function.
Executive watchlist
List issues that leadership should monitor closely.
Include any signals that may indicate a larger problem despite appearing infrequently.
Questions we should investigate
End with 3–5 questions that would help us validate the most important findings.
Be concise and decision-oriented. Do not simply summarize what customers said. Tell me what the feedback means, what deserves attention, and what I should do about it.
Clearly label any conclusions that are based on limited evidence. Do not invent statistics or trends that aren't supported by the feedback.P.S. Get more such prompts in the Prompting Playbook (free for you)
Stay curious, {{first_name | readers}}
PS. If you missed yesterday’s issue, you can find it here.