I Shared My AI Chat With My Boss. Should You?

Hello,

This is Simon with the latest edition of The Weekly. In these updates, I share key AI related stories from this week's news, list upcoming events, and share any longer form articles posted on the website.

As we get deeper into using AI in the office, we keep discovering new ways of working. This week I've started using a genuinely useful technique: sharing a chat with colleagues. When you're having a conversation with Claude or ChatGPT, you can share the response. This doesn't just let a colleague see the entire conversation, including every follow-up question, but they can also continue it themselves and ask whatever's more relevant to them. (I wrote about the benefits of asking follow up questions and pushing deeper in chats in a recent edition.

This came in handy this week as I headed off on holiday. I asked Claude to pull together a set of numbers my manager needed for a presentation. Rather than just copying and pasting the answer across, I shared the entire conversation with him. That way, if he needed other figures or wanted to clarify a point, he could simply pick up where I'd left off. Genuinely useful, and a real time-saver.

That said, a word of caution. You may have seen the news this week that some users' Claude conversations and AI-built tools have been turning up in Google search results. Here's how it happens: sharing a conversation creates a public link, and Anthropic says it doesn't hand these links over to search engines directly. But once a link gets posted somewhere public, e.g. a forum, a social media post, an open web page, search engines can find and index it the normal way, just as they would any other page. So it's not that anyone's guessing your link; it's that a link posted publicly can end up crawled and searchable.

In a work setting, it's unlikely your specific link ends up on the open web this way. But this isn't a one-off scare story — it's happened before, and this week's episode reportedly affected a meaningful number of shared conversations and tools, some containing sensitive material. So it's worth staying mindful of what you put into an AI chatbot, particularly personal or sensitive information, and thinking twice before sharing a conversation you wouldn't want to become public.

If you've shared a link previously and would rather it wasn't public anymore, go to Settings, then Privacy, and look for "Shared chats" — you can unshare any of them from there.

The best assistants don't multitask their attention across a hundred tools. Neither does Catch. It's an AI agent that focuses on one thing — the admin work you'd rather not touch — and does it exceptionally well.

Scheduling, flights, restaurants, follow-ups, vendors, clients. You hand it over; Catch handles the back-and-forth and comes back with it done.

No context-switching. No dropped balls. Just your admin, quietly cleared — so your focus stays on the work only you can do.

Meet the agent built for admin, and it'll be ready to work before your next meeting.

Get started at catchagent.ai — and give your attention back to what matters.

Start now for free

In this section, I’m going to bring to you a real world example of AI use.

Uber rolled out Anthropic's Claude Code to its engineering team in late 2025, and by April 2026, the company had exhausted its entire annual AI budget. Monthly API costs per engineer ran between $500 and $2,000 as adoption spread to 84% of the engineering workforce. The problem, according to COO Andrew Macdonald, is that despite the spending, he cannot draw a clear line between token consumption and meaningful improvements for customers. A problem many other companies are also facing.

Why it matters: Uber's experience is a preview of a reckoning now arriving across large enterprises. Global AI spending reached $2.59 trillion in 2026, yet fewer than a third of organisations can demonstrate financial returns. The shift from measuring AI usage to measuring AI outcomes is the defining business challenge of the second half of this year.

Coinbase told investors in July that 95–100% of its codebase is now AI-generated, up from 40% in February. The company's Head of Platform said the AI agents are doing the equivalent work of 1,200 employees. In May, Coinbase cut approximately 700 staff — around 14% of its workforce — with the shift to AI-assisted development cited as a direct factor. Senior engineers are now managing multiple AI agents rather than writing code themselves.

Why it matters: Coinbase is among the first major public companies to put a specific headcount figure on AI's displacement effect and tie it directly to a restructuring. The pattern of "AI does the work, humans supervise the agents" is moving from aspiration to operational reality faster than most workforce planning assumed.

A July 2026 industry analysis found that 72% of enterprises now have AI agents running in production environments, with Gartner forecasting 40% of business applications will include autonomous agents by year-end. But only 21% of organisations have a mature governance model for those agents, and 52% cite data quality as the biggest deployment blocker, according to research compiled by Technology Radar. The 2026 International AI Safety Report warned that the most pressing risks may come not from the models themselves, but from the complex business systems built around them.

Why it matters: The gap between deployment speed and governance maturity is becoming the defining risk of the current phase of enterprise AI. Most organisations are applying the spending controls and oversight they use for cloud compute only after something goes wrong. As agents are given more autonomy over real business processes — customer triage, compliance checks, procurement — that lag becomes a material liability.

Thanks for reading, and see you next Thursday.

Simon,

Was this email shared with you? If so subscribe here to get your own edition every Thursday.

Enjoying Plain AI? Share it and get a free gift!

If you find this newsletter useful, why not share it with a couple of friends or colleagues who would also benefit? As a special thank you, when two people subscribe using your unique referral link, I’ll send you my "Exclusive Guide: Supercharge Your Work with NotebookLM." It’s a practical, no-nonsense guide to help you turn information overload into your secret weapon.

You can also follow me on social media: