1. OpenAI is trying to make ChatGPT the interface for everything
Their two big moves this week: Apps SDK and AgentKit.
Apps SDK will let apps run directly in ChatGPT.
Think asking Canva to create a slide deck and visuals appear right in your chat window.
Or asking ChatGPT to create a Spotify playlist that can be viewed and previewed in your convo.
No more jumping between tools. Everything happens right in the chat.
They’re also allowing developers to build their apps into ChatGPT.
Early days, but this will change how we'll use apps entirely.
AgentKit allows you to create automations and AI agents without knowing code.
Imagine creating a market research assistant to aggregate weekly news on your competitors.
Or adding new sales leads from chat to Google Sheets and updating the team over email.
There are plenty of tools like this already, but the learning curve can be steep.
AgentKit isn’t a magic wand for complex builds.
But it will allow more people to create simple automations.
It’ll also allow more people to get into the builder mindset (see Everyone Should Learn Vibe Coding).
2. Google's Gemini Enterprise joins the agent race
Google also launched a no-code AI agent tool: Gemini Enterprise.
Drag, drop, and deploy agents with access to your Google data.
You can build agents that:
- Monitor project deadlines, send reminders, and update team task lists inside Docs.
- Summarize and categorize emails, saving results to Google Sheets for team review.
- Create onboarding agents that guide new hires through paperwork in Google apps.
They're also launching an agent marketplace.
You'll be able to buy agents that other people built.
This move helps their AI features become more valuable.
Google owns the enterprise with Workspace (Gmail, Drive, etc.).
But they’re behind Claude and ChatGPT for enterprise AI use.
Pair this move with the upcoming Gemini model release...
And Google could make up serious ground in the enterprise race.
3. Google quietly cut off AI from 78% of the internet
Almost no one talked about this.
But it's a big deal.
Google removed the "num=100" search parameter.
That means crawlers and AI competitors can now only see the top 10 search results per query.
In plain English: AI model search tools just lost access to most of the web.
It's a strategic move.
Google protects its data advantage while limiting what competitors pull into results.
If its own AI models still have deeper access, that's a power play.
For marketers and comms teams, this means less organic discovery through AI.
And Google will control more of what AI learns next.
4. OpenAI CEO: "Expect a very high rate of change"
Sam Altman announced updates for their hot new video app, Sora.
His blog mentioned plans to add revenue-sharing and creator controls.
But the real signal was in his mindset:
“Please expect a very high rate of change from us; it reminds me of the early days of ChatGPT. We will make some good decisions and some missteps, but we will take feedback and try to fix the missteps very quickly.”
Translation: Your AI tools will keep changing weekly.
Don’t overwhelm yourself trying to keep up with everything.
Double down on learning the platform your company uses.
Keep up with it.
Know that it will change weekly.
New features, interface changes, buttons moving around…
Everyone is shipping at lightning pace.
They're testing in public and we’re the testers.
Stay close to one platform you trust, and adapt fast.
5. AI traffic already converts better than search
Here are some stats that stood out to me this week:
- ChatGPT traffic is outconverting search (AI: 11.4%; Paid Search: 9.3%, Organic Search: 5.3%)
- AI-assisted shopping is estimated to be up 520% this holiday season
- Social media use is down ~10% since 2022.
AI is the new traffic source.
Brands are already optimizing for visibility inside AI results.
Soon it'll be a paid advertising channel, just like search and social.
