Hi {{first_name}}!
Happy Friday! There's been a lot happening in the AI world this week.
I'm covering some fresh updates from Gemini that I've already started building into my routine, a big announcement from OpenAI that I think every business owner should pay attention to, and the usual roundup of what's new and noteworthy across the AI space.
This week we're covering:
Fresh updates from Grok, OpenAI, Anthropic, and more
Two new Gemini features I’m eager to share
What OpenAI’s latest announcement means for your business and your team
OK, let's get into it!
New and Noteworthy
Grok 4.20 is here and it runs on four AI agents working at the same time: Instead of one model generating your answer, four specialized agents (a coordinator, researcher, logician, and creative) work in parallel, debate each other, and hand you the consensus. The result: hallucinations dropped 65% in early testing because when one agent gets something wrong, another catches it. In a live stock trading competition, apparently it was the only profitable AI model in the field. Available free at grok.x.ai, with a $30/month SuperGrok plan for faster responses and a "Heavy" mode that scales up to 16 agents. Worth testing just to watch the agents think in real time.
OpenAI is building a smart speaker with Jony Ive: OpenAI’s first physical product is reportedly a smart speaker with a built-in camera and facial recognition, priced between $200 and $300, and expected to ship in early 2027. This effort kicked off after OpenAI acquired Io Products last year. The device is designed to observe its surroundings and nudge users toward useful actions, and prototypes like a smart lamp are already in development. AI glasses are also in the pipeline for no earlier than 2028. OpenAI entering hardware is a shift worth watching closely as larger players like Amazon and Apple push AI-first devices.
The U.S. is exporting its AI stack globally: A new “Tech Corps” initiative launched that’s essentially a peace corps for AI, with India as the first partner. The aim is to help other countries build sovereign AI infrastructure using American tools as an alternative to China’s influence. AI is becoming global infrastructure, and which countries lead that effort could shape how the world builds and governs AI.
Anthropic adds security scanning to Claude Code: Anthropic’s new Claude Code Security scans entire codebases to find vulnerabilities and suggests patches your team can review. It goes beyond simple pattern matching by tracing data flow and how components interact, catching logic flaws and broken access control that typical static analysis tools miss. In testing, it reportedly flagged hundreds of issues across open-source repos. Currently available in research preview for enterprise and team plans, with accelerated access for open-source developers. If you have a dev team, this is worth a look.
ElevenLabs adds A/B testing for AI voice agents: ElevenLabs released “Experiments,” a feature that lets you run controlled A/B tests on live AI voice traffic. If you’re building or running AI voice workflows, this gives you a structured way to improve performance instead of guessing what works.
Superpower launches an AI health partner: Superpower’s new AI chat tool analyzes lab results, blood tests, microbiome data, and over 100 biomarkers to deliver personalized wellness recommendations. The standout feature is memory, letting the tool build a health picture over time. At $17/month it’s positioned as a preventative tool, not a physician replacement. Early but a useful example of AI moving into personal health management.
Perplexity's Comet browser is coming to iPhone: Perplexity’s AI-native browser, Comet, is headed to iOS. It’s already making waves on desktop, and mobile availability puts an AI-first browsing experience in more people’s hands. One to watch as browsers start to look more like AI assistants.
Two New Gemini Features I Wanted to Share
I use Gemini every single day, and these two features just became available. I had to share them because they’re genuinely useful and both are completely free.
Scheduled Actions
Gemini now lets you create scheduled prompts that run automatically on a daily, weekly, or monthly basis. You name the action, write your instructions, pick a time and frequency, and Gemini handles the rest in the background.
Think about what that means practically. You could set up a morning digest that pulls together industry news relevant to your business before your day starts. Or a weekly check on competitors. You’re essentially building a lightweight assistant that shows up consistently, without you having to remember to ask.
Setup is simple. Inside Gemini, look for “Scheduled Actions” in the top left corner. From there you can write your own custom instructions or start from one of their built-in templates.
This is a feature worth building a habit around.If you set one up, reply and let me know how you’re using it. I’d love to share additional use cases inside the Ampra Circle community.
Video Templates
If you’ve ever wanted to create video content but felt like you were starting from scratch every time, this one’s for you. Gemini now has a library of pre-set visual styles you can browse and apply to your content ideas.
The real value isn’t just inspiration. It’s consistency. Once you find a style that fits your brand, you can reuse it as a template every time. Same look, same feel.
Both features are inside Gemini right now.
OpenAI's Announcement This Week Says About the Gap Between Having AI and Using It
This week OpenAI announced Frontier Alliance, multi-year partnerships with McKinsey, Boston Consulting Group, Accenture, and Capgemini. The goal is simple: move AI from isolated experiments into full-scale deployment across the organization. Not just pilots. Not one department. Real integration into how the business runs.
The announcement quietly reinforces what many leaders already feel. Most companies are experimenting with AI. Far fewer have scaled it in ways that drive measurable value.
The bottleneck isn’t the technology. The tools are here. They’re powerful. They’re accessible.
The real challenge is capability.
Teams need:
Clarity on where AI fits into real workflows so impact can be measured.
Collaboration across functions so learnings spread instead of staying siloed.
Curiosity and disciplined practice so AI becomes routine, not occasional.
As BCG put it, AI has to be linked to strategy, embedded in redesigned processes, and supported by the right culture. Otherwise it just collects digital dust.
Right now most teams use AI as an answer engine. Some are moving into using it as a thinking partner or even lightweight agents. Within a year, that will be table stakes.
The real competitive edge will come from using AI as an innovation engine. Designing new processes. Reimagining customer experiences. Creating new revenue streams. Changing how decisions get made.
This isn’t just an enterprise problem. SMBs face the exact same adoption gap, just with less room for error. The businesses that pull ahead won’t be the ones with the fanciest tools. They’ll be the ones whose teams know how to apply AI to real problems consistently.
That’s the difference between AI being a line item and AI changing how your company operates.
So where does your team stand?
You don’t need a massive consulting budget to close this gap. But you do need to invest in upskilling your people.
Teams need more than tool demos. They need shared language, practical frameworks, and a culture that encourages experimentation and cross-functional learning. Moving from using AI for answers to using it to invent new ways of working requires intention.
That’s exactly why I built our AI Training for Teams. Three focused modules, 90 to 120 minutes each, delivered by me and designed to meet your team where they are. Whether they’re complete beginners who need clarity fast, or already using AI daily and ready to push toward real innovation.
If you’re unsure where your team really stands, I created an AI Readiness Scorecard you’re welcome to download or print. It’s a practical self-assessment for business leaders who want clarity on where they stand and what to do next.
And If you’d like to chat about how this could apply to your team, I’m always happy to jump on a call, hear about your current situation, and share a few initial ideas.
OK that's it for this week! Curious which section resonated most with you. Hit reply and let me know.
And if there's a specific AI challenge you're working through in your business right now, I'd love to hear about it.
See you next week,
Julien
PS: If this was helpful, forward it to a business owner or team member who's still trying to figure out where to start with AI. They can subscribe at www.ampra.ai/join-our-newsletter.