What Happened This Week
1. Google Workspace Studio: Build Automations In Your Apps
Google launched Workspace Studio this week.
It's a no-code tool that lets you build AI automations inside Gmail, Drive, Chat, and Calendar.
These automations handle multi-step workflows.
From simple stuff like managing emails and approvals to more complex multi-step, multi-tool processes.
Explain what you want to build to your AI assistant.
Either start from scratch with your idea or use templates for common workflows like status reports, reminders, and follow-ups.
Then it helps execute your workflow.
The signal?
Google calls these agents and they do some agentic work.
But think of them like smart automations that get you thinking like a builder.
Tools like this expose non-technical people to systems thinking.
An easy way to learn the six building blocks of AI systems in practice.
How data flows, where logic and triggers fit, how prompts work, where AI belongs in a workflow, governance, and feedback.
This is automation training wheels.
Don't mistake these simple workflows for the ceiling of what agents can truly do at scale.
Just like how early AI chatbot work typically starts with writing emails.
Then people discover more complex use cases.
Same pattern here.
Start with everyday workflows.
Then expand your understanding as you go.
One barrier to call out is corporate access.
Data and privacy policies may limit who can use this inside your org.
If you have access to Workspace Studio (for Google Workspace users) or to another no-code automation platform, the opportunity is the same:
Experiment, ideate, and start to think bigger about your work.
2. Amazon Launches AI Models, Agents, and Custom Training Tools
Amazon Web Services (AWS) rolled out a wave of AI announcements at re:Invent this week.
New models, agents, custom model training, and chips.
You might not think of Amazon as an AI model provider.
But if your company runs on AWS (and many do), these tools are important.
What launched:
- Nova models: Amazon has its own family of AI models for reasoning, coding, voice, and multimodal work.
- Nova Forge: A service that lets companies train custom models using their own data combined with Amazon's.
- Frontier agents: Pre-built AI agents for coding (Kiro), security, and DevOps that can run for hours or days.
- Nova Act: A service for building browser automation agents.
- Agent-building platform: Upgraded tools for creating and managing AI agents.
The signal?
AWS isn't a flashy AI name like OpenAI or Anthropic.
But they're a huge player in enterprise tech.
And they're leaning into agents like other big players (Salesforce, Google, Microsoft), creating robust tools and profit streams around AI, knowing they already power so much of business tech.
If your company uses AWS, your tech teams are probably exploring these tools.
The broader point: don't only track the big three or four model providers.
Companies like AWS are building solutions your org may use.
Less splashy, more practical given existing tech stacks and relationships.
3. OpenAI Goes "Code Red" After Google's Gains
OpenAI is rushing a new reasoning model to counter Google's Gemini 3.
Internal reports say the company declared "Code Red" after Gemini started gaining users and benchmark leadership.
Sam Altman told staff the new model outperforms Gemini 3 in internal tests.
Launch planned as early as next week.
To focus resources, OpenAI is pausing several projects: ad rollout plans, new autonomous agents, and its "Pulse" briefing service.
They're prioritizing core ChatGPT upgrades and the reasoning model release.
In addition to Gemini, competitive pressure is coming from other AI organizations too:
DeepSeek released another high-quality, cheap reasoning model that anyone can use and modify.
Mistral dropped new models strong at reasoning and coding, backed by partnerships with Nvidia and Microsoft.
The signal?
Even the dominant AI player has to react fast.
Google's Gemini release weeks ago triggered immediate strategic shifts at OpenAI.
That's how competitive this space is.
4. ChatGPT Ads Are Coming
Although temporarily on pause, developers found ad-related code in ChatGPT's Android app this week.
References to "search ad," "carousel," and ad modules.
No formal announcement from OpenAI yet, but infrastructure is built and ready when they pick it back up.
This is a natural next step.
AI tools are expensive to run, and OpenAI needs revenue.
More importantly: they have unprecedented personalization data.
AI companies know your personality, what's important to you, the problems you're solving, and what you're researching.
That's advertiser gold.
And brands will always pay for ad space that's likely to convert.
AI will be able to personalize ads better than anything we've seen before.
The signal?
This is where all AI platforms are headed.
But building an ad business is complex.
You need infrastructure for brand partnerships, measurement, targeting, compliance, and data processing.
Google already has all of this from decades of running ads.
They can turn on ads in Gemini whenever they want.
OpenAI is starting from scratch.
Good that they're building now, but they have a long way to catch up to Google's established ad business.
For marketers: watch how this plays out. It's definitely something to test when it's available.
5. Numbers Worth Knowing This Week
- AI-powered shopping exploded this holiday season.
- Black Friday and Cyber Monday hit record U.S. online sales, with AI-directed traffic up 760% from Nov 1 - Dec 1.
- 47% of U.S. consumers used AI for holiday shopping tasks like price comparison and product research.
- Sources: Adobe Analytics, Visa
- MIT study: AI can already replace 11.7% of the U.S. workforce.
- That covers about $1.2 trillion in wages, with high exposure in HR, finance, logistics, and admin roles.
- Source: MIT Iceberg Index
- Anthropic's engineers rely on Claude for 60% of their daily work.
- They report a 50% productivity increase, and a quarter of AI-assisted tasks wouldn't have happened without it.
- Source: Anthropic
