Hi!

Happy Friday! I can't believe it’s already October….Spooky how fast this year is going by! 

I was hoping to send this out earlier in the week but it’s been a whirlwind. As we head into October, we’re seeing AI news and model updates come fast and furious with an amazingly capable new version of Claude, a new version of Gemini expected this month, and lots of new features and updates expected from ChatGPT in the coming week, (with some already here)!

Beyond the updates mentioned above, this week I’ll  also be spotlighting NotebookLM,  easily one of my top 5 AI tools. If you’ve seen one of my presentations you know what it does, but I wanted to showcase one of my personal use cases.

In addition, I’m also tackling the question I get at almost every presentation and meeting I have: “How do we handle AI privacy and data security?”

So in summary… we’re covering:

• New and Noteworthy AI updates

• Tools worth sharing

• NotebookLM - a creative use case

• AI Privacy and Data Security

• Celebrating 100+ members in the Ampra Circle community!

OK, let's get into it!

New and Noteworthy

Recent AI Developments

• Google Mixboard Launch: New AI-powered moodboard tool that creates visual concepts from simple text prompts. Perfect for planning events, exploring design ideas, or creating mood boards without design skills.

• Replit's Agent 3: Just launched with 10x more autonomy, running for over 3 hours straight without human input. This AI agent can build apps, test code, fix bugs, and click through interfaces like a real user, representing a major leap in autonomous software development.

• Claude Projects Gets Bulk Chat Management: You can now move chats in and out of Claude Projects in bulk, a feature many of us have been requesting since Projects launched.

• ChatGPT Pulse: This is OpenAI's new daily news digest that arrives on your phone or app every morning. It's personalized based on your chat history and connected apps like your calendar. You tell it what you want to focus on, and it curates relevant updates for you each morning. Think of it as your personal AI research assistant that keeps you informed without the inbox overload.

• Microsoft adds Claude to 365: Business users can now switch between ChatGPT and Claude in Copilot, giving teams more AI options for different tasks.

• Photoshop Embraces Nano Banana: Adobe integrated Google's Nano Banana directly into Photoshop Beta. Instead of fighting the competition, they're giving users choice. You can now select between Firefly, Flux Context Pro, and Nano Banana right inside Photoshop's generative fill tool.

• Alibaba's Qwen3-Max: Over 1 trillion parameters, designed for coding and autonomous agent tasks, claiming to outperform Claude Opus 4 on several benchmark

• Google Opal: No-code platform for chaining AI models into custom workflows. Create multi-step apps that combine research, content creation, and media generation without coding. This just came out a few days ago, and it is still in the early experimental phase, but it looks cool. Basically, it seems like a chatbot agent builder on the front end with the ability to make technical edits to the workflow on the backend. 

Tool Spotlight: NotebookLM

I've been using Google's NotebookLM since it first came out in 2023, and I love it for research, learning and content creation.

Here's what makes it special: You can upload company documents, industry reports, videos, URL’s or client feedback into a “notebook” and NotebookLM becomes an expert on your specific content. It doesn't just summarize, it connects ideas across documents and identifies patterns and insights you might miss. It also generates visual presentations, audio overviews, study guides and mind maps!

Real use case: I had 18 episodes from the "My First Million" podcast sitting in my "watch later" list only a few of which I had listened to. Instead of spending 18+ hours listening, I fed all the YouTube links into NotebookLM and asked it to pull out business trends and AI insights across all episodes.

What came back was incredibly useful: a visual mind map that connected themes I wouldn't have spotted listening to individually, plus organized reference material I could actually use and a 30 minute audio overview I could listen to while in my car.

Other NoteBookLM business applications:

• Upload quarterly reports and get strategic insights for board presentations

• Feed it competitor websites and receive comprehensive market analysis

• Drop in client feedback and generate action plans for service improvements

AI Privacy and Data Security

How can you handle privacy in the age of AI? This is the most common question I get at every presentation, and for good reason. Let's clear up some of the confusion.

The Reality: Most AI privacy concerns stem from misunderstanding how these tools work. 

Key Privacy Guidelines

Business vs. Personal Accounts:

• ChatGPT Business and Enterprise don't train on your data

• Claude for Business keeps conversations private and separate

• Google Workspace AI follows your existing data governance policies

• Ultimately, it’s a best practice to always use business accounts for work. They operate very differently than personal accounts

Data Handling Best Practices:

• Never upload client data with personally identifiable information

• Remove names, addresses, and financial details before uploading, and when in doubt, anonymize first

• It’s important to have well structured and well organized data

• Be extremely careful connecting AI tools to Google Drive, Dropbox, or other storage (consult with IT before doing so). 

• Accidental connections can expose sensitive data across your organization

• Only grant necessary permissions and regularly audit system access

Create Internal Policies:

• Define what can and cannot be shared with AI tools

• Train your team on proper usage

• Document your organization's specific guidelines

For most businesses, following the guidelines above will provide the necessary protection with cloud-based AI tools. But some industries require absolute data control, especially those handling HIPAA-protected health data, attorney-client privileged documents, financial documents or other highly proprietary business data.

That's why we recently partnered with Trusted Intelligence which has a cloud solution and an on-prem hardware solution that is enterprise-grade and allows you to run fine tuned LLM’s completely on your infrastructure with zero data retention and unlimited token usage. Click here to learn more.

You don't have to choose between AI innovation and data security. Whether cloud-based tools with proper safeguards or private on-premise solutions, there's an option that fits your business.

Let me know if you are curious or have questions about how this works!

Ampra Circle Community Milestone

On a final note, I'm excited to share that we've hit 100+ members in the Ampra Circle community! The conversations and resource sharing have been incredible. Thank you to everyone that has joined thus far!

Any questions feel free to reply to this email directly and I’ll get back you!

Julien