- Everyday AI
- Posts
- EP 624: From Automation to Autonomy: Practical Steps for Enterprise AI Adoption
EP 624: From Automation to Autonomy: Practical Steps for Enterprise AI Adoption
OpenAIās Sora app tops the App Store, Google drops $4 billion on AI, Jeff Bezos says thereās an AI bubble
Everyday AI
October 03, 2025
š Subscribe Here | š£ Hire Us To Speak | š¤ Partner with Us | š¤ Grow with GenAI
Sup yāall! š
Sora Invite code giveaway results at the end of the newsletter, FYI.
Anyways, TONS of new AI features we can dive into more next week.
What do you want to see more of?
Perplexity Comet Release ā They just released their agentic browser to ALL users, not just paid.
NotebookLM Customizations ā now allows you to customize and personalize your chat experience.
Claude File Creation -- The ability to create Excel and Powerpoints in Claude has now been rolled out to all paid users. (Previously was $100-$200 Max plan)
Microsoft Excel Agent Mode ā This got a close second in our last poll. Do you want a closer look at the newly released Agent Mode in Excel?
What do you want to see more of?š³ļø VOTE to see live results š³ļø |
āļø
Jordan
(Add me on LinkedIn. Let me know youāre from the newsletter!)
Outsmart The Future
Today in Everyday AI
8 minute read
š Daily Podcast Episode: An Accenture leader told us how to prepare for four big AI trends. Give it a watch/read/listen.
šµļøāāļø Fresh Finds: A $40k/year school run by AI, reality doesnāt exist anymore, IBM drops new family of AI models and more. Read on for Fresh Finds.
š Byte Sized Daily AI News: NotebookLM goes more custom, OpenAIās Sora app tops the App Store, Jeff Bezos says thereās an AI bubble and more. Read on for Byte Sized News.
šŖ Leverage AI: Weāre taking what we learned from Accenture, an industry leader, and breaking down how AI-powered autonomy is shaping your workplace. Keep reading for that!
ā©ļø Donāt miss out: Did you miss our last newsletter? Did you miss our last newsletter? It included: Inside Sora 2, OpenAIās $500 billion valuation, Perplexity releases agentic browser to all, Microsoft takes on ChatGPT with new offering and more. Check it here!
How the Future is Being Shaped by AI-Powered Autonomy š
If most companies are using the same AI systems, how can they stand out and get ahead?
And as agentic AI becomes table stakes, what do enterprises need to keep in mind to make AI work?
And how can we even trust an AI-powered workplace when most people can't even explain the basics of AI?
Accenture's Mary Hamilton joined the the show to talk about building trust in an autonomous workplace, how we can prepare for the future of work, and four emerging AI trends you can't miss.
Also on the pod today:
ā¢ Autonomy and Enterprise AI Adoption š¢
ā¢ Agentic AI Models and Productivity Shifts š
ā¢ Continuous Learning Loops in Workplace AI š§āš«
Itāll be worth your 29 minutes:
Listen on our site:
Click to listen
Subscribe and listen on your favorite podcast platform
Listen on:
Hereās our favorite AI finds from across the web:
New AI Tool Spotlight ā Super Intern tries to bring people closer with AI, Strix is an AI agent that tests for vulnerabilities, Nyx is an AI-powered app to help you rest better.
AI Generated Everything ā Like we talked about a few months ago, the rest of the world is picking up on the whole āreality is goneā concept.
Context Engineering ā Anthropic says for AI agents, context engineering is better than prompt engineering.
AI Wars ā Gemini is gaining while Grok is losing ground to Perplexity.
GenAI Traffic Share Update
Takeaways:
ā Gemini's rapid ascent continues.
ā Perplexity catches up with Grok.šļø 12 months ago:
ChatGPT: 87.1%
Gemini: 6.5%
Perplexity: 1.7%
Claude: 1.7%
Copilot: 0.9%šļø 6 months ago:
ChatGPT: 77.2%
DeepSeek: 7.6%
Gemini: 5.5%
Grok: 3.2%ā Similarweb (@Similarweb)
8:10 AM ā¢ Oct 3, 2025
AI in education ā This school is run by AI and thereās no teachers. And it costs $40,000 a year?!
AI Models ā IBM has released its latest models, the Granite 4.0 family of hybrid LLMs
AI Training ā Has AI run out of data to train on? Maybe.
Sora 2 ā Sora 2 Pro is rolling out with 10 and 15 second duration options
Sora 2 Pro looks like rolling out with 15 second duration
ā Jordan Talks Everyday AI (@EverydayAI_)
6:49 PM ā¢ Oct 3, 2025
1.Sora tops Appleās App Store, edging out Google Gemini and ChatGPT š„
OpenAIās invite-only Sora app has jumped to No. 1 on Appleās App Store despite iOS-only access and code-based entry, according to CNBC. Powered by OpenAIās new Sora 2 model, the app lets users generate short AI videos, remix othersā clips and post to a shared feed, showcasing increasingly realistic scenes and sounds.
The early surge highlights how fast consumer video generation is moving into the mainstream, but safety and legality questions are already bubbling up, including a controversial clip depicting Sam Altman shoplifting, with OpenAI touting controls over user likeness and active iteration in response to feedback.
2. Bezos calls AI an āindustrial bubble,ā but says the tech is undeniably real š«§
Speaking at Italian Tech Week in Turin, Jeff Bezos said the AI sector looks like an āindustrial bubble,ā where stock prices and funding are racing ahead of business fundamentals, according to CNBC and Reuters.
He argued bubbles can still be productive, noting that past biotech hype ultimately delivered lifeāsaving innovations, and predicted AI will transform every industry despite todayās froth. Bezos cited unusual funding patterns, like tiny teams landing billions, while figures such as Sam Altman and Goldman Sachs CEO David Solomon warn a reset may follow the current bull run.
3. Nvidia and Fujitsu team up on AI infrastructure in Japan by 2030 š¾
According to the Taipei Times, Nvidia and Fujitsu announced a collaboration to build an AI infrastructure in Japan aimed at enabling smart robots and applications across healthcare, manufacturing, the environment, next-gen computing and customer service.
The partnership will lean on Nvidiaās GPUs and Fujitsuās local expertise, with robot maker Yaskawa cited as a potential project partner, though no investment figures or concrete pilots were detailed. The timing matters because executives say the AI industrial revolution is underway, and Japanās push could position it as a leader in robotics while addressing labor shortages that are already shaping the economy.
4. Google drops $4B on Arkansas AI data center šļø
According to Alphabetās latest Q2 2025 earnings and a company statement, Google is building its first Arkansas data center campus in West Memphis by 2027, paired with a new 600-megawatt solar project and 350-megawatt storage to cover its power needs. It has a $4 billion price tag.
The company also launched a $25 million Energy Impact Fund and will offer free Google AI courses and certifications to residents, signaling a talent pipeline play alongside infrastructure. With Google Cloud revenue up 32 percent year over year and Alphabet targeting roughly $85 billion in CapEx this year, this move underscores how big tech is racing to scale AI capacity while trying to keep energy costs in check.
5. The GPT-5 haters were off base, Sam Altman said š
Sam Altman says the messy August GPT-5 launch was misread, with glitches and stiff tone overshadowing what was a steadier, incremental upgrade rather than a leap to sci-fi.
He argues that people expected instant AGI, but GPT-5 was designed to play the existing game better, which explains the Reddit backlash and the āsame gameā critique. Altman is reframing the narrative from hype to execution, signaling that reliability, accuracy and safer behavior are the near-term goals, not dramatic fireworks.
6. NotebookLM adds customizable chat, aiming at smarter learning š£ļø
According to NotebookLM, users can now tailor response length and conversation style, a timely push as AI chat tools race to feel more personal and useful.
The update introduces a Learning Guide that tests and deepens understanding of your materials, signaling a shift from passive chat to active coaching. For anyone growing a career or company, this could speed up training, onboarding, and content digestion by turning documents into guided lessons that are easier to retain and apply.
š¦¾How You Can Leverage:
High-end photo editing used to require years of training.
Adobe Photoshop demanded hundreds of hours to master one task like a seamless background removal.
Now?
You literally just say "swap out this background" and it automagically happens.
Same professional results. Zero technical expertise required.
Mary Hamilton said the same level of AI autonomy is happening everywhere.
And we've all gotta be prepared.
Mary is a Managing Director at Accenture, Leading the Connected Innovation Centers Globally. So, she knows a bit about workplace innovation.
Like photo retouching, she said many professions requiring decades of expertise just essentially became accessible to everyone because of large language models.
Mary just laid out the blueprint from Accenture's 25th annual Technology Vision report, laying out the path forward for AI-enabled workplaces.
Read the whole thing for free here, or a quick preview:
Development costs hit zero. Robots gained conversational intelligence. Trust-building determines who wins the autonomy race.
And most importantly ā the companies implementing these changes right now will dominate their industries within 18 months.
Here's what ya need to know:
1 ā Development Costs Hit Zero, App Building Goes Wild š
Enterprise software creation just became stupidly cheap.
Mary calls this the Binary Big Bang. Development costs that used to run six figures now cost basically nothing.
Here's what's happening.
Marketing teams are cranking out custom analytics dashboards during their lunch breaks. Sales folks build automated proposal systems between client calls. Operations peeps design workflow tools while waiting for Zoom meetings to start.
All functional applications. All built in hours instead of quarters.
Mary told us how Accenture's Fortune 500 clients are pumping out dozens of micro-applications daily. Traditional IT approval processes can't even keep up.
The game changer? Natural language replaced coding as the development interface.
When language becomes programming, every single employee becomes a potential builder. Companies that unlock this capability will absolutely steamroll competitors still waiting for IT tickets to get approved.
Try This:
Identify your team's five most annoying daily tasks right now.
Test Zapier Central or Microsoft Copilot Studio against each problem this week.
Document what gets automated versus what doesn't work. You'll quickly map your automation readiness and spot the natural builders on your team.
2 ā Robots Just Got Conversational Smarts š§
Physical automation became genuinely intelligent overnight.
Mary explained how large language models are giving robots human-like reasoning abilities. Instead of painful programming for specific tasks, workers now just tell robots what they want in plain English.
This changes everything about manufacturing.
Robots used to require complex coding for every single function. Now they understand context, adapt to changing situations, and work naturally alongside humans.
Mary shared how Accenture partnered with Keon Group to deploy AI-driven robots in warehouse operations.
The results? Faster order fulfillment, way lower costs, and massively improved worker safety.
But here's the kicker most companies are missing. This isn't about replacing humans with robots. It's about supercharging what humans can accomplish.
Robots handle the physical precision and endurance stuff. Humans bring creativity and complex reasoning. Together they absolutely crush either working alone.
Mary emphasized how multimodal AI gives robots three-dimensional understanding of real environments. They can see, process visual info, and make smart decisions instead of just following pre-programmed responses.
Try This:
Walk through your most labor-intensive processes and spot tasks requiring both physical work and real-time decisions.
Research robotics companies offering natural language interfaces for those specific applications.
Book demos this month. Early adoption creates serious competitive advantages before this becomes standard.
3 ā Trust Tech Beats Training Programs Every Time šŖ
Trust determines autonomy success way more than fancy technology.
Mary shared the real secret behind enterprise AI adoption. Trust gets built through tiny moments of successful interaction, not boring training sessions explaining how systems work.
Most companies are doing this completely backwards.
They're teaching employees about tokenization and large language model architecture instead of just letting them experiment safely. Mary actually coaches her AI assistants on communication style and personality, not just accuracy. She tells them to cut the snark and fine-tune responses.
This builds genuine working partnerships.
The winning approach? Verify first, trust gradually. Start with low-stakes scenarios where AI screwing up creates minimal damage but success shows clear value.
Mary described how Accenture built verification technology right alongside AI deployment. Knowledge graphs provide business context for data interpretation. Cross-checking mechanisms catch potential errors before they mess up operations.
These safeguards speed up trust development like crazy.
The competitive advantage compounds every month. Organizations building systematic trust get enterprise-wide adoption while competitors battle resistance and fear-based paralysis.
Try This:
Start trust journals with three employees using AI daily.
Have them rate response quality and track confidence levels for different tasks.
After two weeks, analyze which interactions build trust fastest. Design your rollout around natural trust patterns instead of forcing adoption down people's throats.
š Sora Invite Code Giveaway š
Right now, Sora 2 is free but invite only. Weāre giving away a few codes this week.
Alexie O. is todayās winner.
Alexie, hit us with a reply to this email and weāll send your Sora code.
Reply