
Hi {{first_name}}!
Happy New Year! I hope you had a great holiday break.
This week we're covering some significant AI updates worth knowing about, a cautionary tale from Salesforce that every business owner should pay attention to as we head into 2026, and a tool that's been saving me real time every single day.
This week we're covering:
New AI developments from Google, OpenAI, Nvidia
Lessons from Salesforce's AI failure and what it means for your 2026 planning
Wispr Flow - the free dictation tool that saves me hours
OK, let's get into it!
New and Noteworthy

ChatGPT's New Formatting Interface: OpenAI quietly rolled out a feature that transforms ChatGPT's layout to match whatever task you're working on. When you highlight text, a mini editor toolbar appears so you can format documents directly inside the chat interface. This makes ChatGPT feel less like a conversation and more like an actual workspace tool. Small change, big impact on daily usability.
OpenAI's "Head of Preparedness" Role: Sam Altman posted a job listing for someone to plan for increasingly advanced AI models, including systems that can self-improve. This signals OpenAI is thinking seriously about the long game as models become more autonomous. For businesses, it's a reminder that AI capabilities are moving faster than most organizations can keep up with.
Google's Year in AI Research: Google published their 2025 recap covering AI reasoning breakthroughs, scientific applications, disaster preparedness tools, and progress on robotics and energy-efficient chips. I dropped the full report into NotebookLM and created a breakdown focused on what matters for business owners and busy professionals. Here’s the 8-minute overview that pulls out the key developments that might show up in business tools in 2026. Check it out here.
Nvidia's $20B Groq Acquisition (Note: Groq the chip company, not Grok the AI chatbot): In their largest deal ever, Nvidia is paying roughly $20 billion for a non-exclusive license to Groq's breakthrough inference technology. This brings Groq founder Jonathan Ross (the original architect of Google's TPU) and key engineers directly into Nvidia's ranks. The move signals Nvidia is securing its position in the AI inference market while neutralizing potential competition from Google's chips.
Nano Banana 2 Flash Coming Soon: Google is about to release Nano Banana 2 Flash (codenamed Mayo), the efficient sibling to Nano Banana Pro. This Flash variant delivers near-Pro image quality at lower costs and significantly faster speeds. Perfect for high-volume creators and businesses scaling visual content with lower overhead. Early leaks show outputs rivaling the top tier image generation models.
ChatGPT Ad Monetization Plans: OpenAI plans to embed sponsored content and ads directly into ChatGPT responses, including personalized ads based on your chat history. This comes despite CEO Sam Altman's previous concerns about advertising in AI tools. I’m guessing this will mean the free tier might start feeling less useful over time as commercials creep in.
NotebookLM's Major Updates: Google's research assistant just got significantly more powerful with several updates worth noting. You can now upload notebooks directly into Gemini to combine multiple sources seamlessly, pick up conversations across mobile and web with chat history syncing (and delete them anytime for privacy), send study guides and briefing docs directly to Google Docs or Sheets, and organize scattered insights with the new data tables feature that exports to Sheets (rolling out to free users soon). The real power here is continuity of context across tools: NotebookLM for grounded research with cited references, Gemini for creative expansion. This is how AI tools should evolve: not competing, but composing together. Note that the Gemini integration appears to be available for individual accounts but not workspace accounts yet, which seems backward for business users.
Google's CC Productivity Agent: Google Labs launched CC (as in “Carbon Copy”), an experimental AI assistant that connects directly to your Gmail, Calendar, and Drive. Every morning it sends you a personalized briefing called "Your Day Ahead" covering your calendar, important emails, and relevant files. You can also email CC directly (at [email protected]) to get help with tasks, add to-dos, or summarize email threads. It only emails you privately, never sends to others. Currently in the US and Canada only with a waitlist, but Google AI Ultra subscribers get priority access. This is Google's play at building an AI assistant that lives inside your existing workflow rather than requiring you to learn a new app. I actually really like this if for no other reason than it feels more natural and feels like you’re emailing an assistant for support.
Getting Your Team Ready for AI: Lessons From Salesforce's Failure
As we head into the new year, there's a story worth paying attention to. Salesforce just admitted something most companies won't: they moved too fast with AI and it backfired.
The company cut over 4,000 customer support jobs in 2025, replacing skilled workers with AI systems. Initially, they celebrated the move as a productivity win that could "handle customer conversations at scale." But reality hit hard. AI struggled with complex cases, complaints increased, and service quality dropped. Leaders now acknowledge they overestimated what AI could do and underestimated the value of human judgment.
This isn't a story about AI failing. It's about implementation without preparation.
The businesses winning with AI right now aren't replacing anyone. They're upskilling and reskilling their teams to use AI as a tool in order to increase productivity. Customer service reps who use AI to draft responses but apply their own judgment. Marketing teams who let AI handle the first draft so they can focus on strategy. Operations people who automate the boring stuff (like onboarding or taking hours to write an SOP or job description) and instead spend their time on what actually matters.
The pattern is clear: AI works when people know how to work with it.
As you're planning for 2026, this is worth considering. Not whether you should use AI, but whether your team is ready to use it well. Because six months from now, your competitors will have figured this out. And the gap between companies whose teams know how to leverage these tools and companies whose teams are still figuring out the basics is going to get really wide, really fast.
We built our AI Training for Teams program specifically for this. Three sessions that take your team from "we just use ChatGPT to help draft emails" to "everyone uses multiple AI tools in our daily workflow to save 5-10 hours a week." Practical, hands-on, and customized to how your business actually works.
The Salesforce story isn't just a cautionary tale. It's a roadmap for what not to do. Learn from their expensive mistakes instead of making your own.
If you're thinking about how this fits into your organization's 2026 plans, hit reply and let’s chat! I’m here to to help as a resource and a guide.
Tool Spotlight: Wispr Flow

I've been using Wispr Flow for over a year and it's become one of those tools I use without even thinking about it anymore.
Wispr Flow accurately dictates what I say way better than native dictation on my phone or MacBook. It automatically cleans up the "ums" and formats properly as you speak.
Where this actually helps business owners:
You're reviewing a proposal in Google Docs and want to leave comments for your team. Speak them instead of typing each one out.
You're in back-to-back meetings all morning and your inbox is piling up. Between calls, you dictate quick responses to keep things moving instead of letting messages sit for hours.
When I'm working with AI tools like ChatGPT or Claude, I talk through my prompts instead of typing them. Way faster than composing written prompts from scratch.
I dictate at roughly 120 words per minute versus typing at 45. This tool matters when you're sending a high volume of messages daily across email, Slack, and text.
Available on Mac, Windows, and iPhone.
Any questions about team training or how to think about AI implementation for your business? Just hit reply.
Looking forward to what 2026 brings,
Julien
PS: If you know someone who's thinking about how to get their team ready for AI in 2026, please forward this to them. They can subscribe at www.ampra.ai/join-our-newsletter.