- Everyday AI
- Posts
- Ep 648: How 74% of Enterprises Get Real AI ROI While Pundits See Failure
Ep 648: How 74% of Enterprises Get Real AI ROI While Pundits See Failure
Inside Wharton's AI ROI study, Apple hires Google Gemini for smarter Siri, OpenAI backpedals on government backing talk, Trump official says no to AI bailout and more.
Everyday AI
November 06, 2025
š Subscribe Here | š£ Hire Us To Speak | š¤ Partner with Us | š¤ Grow with GenAI
Sup yāall š
Two HUGE AI news stories today:
OpenAIās CFO (kinda) said there should be a government guarantee for AI financing. (then took it back)
Apple reportedly finally has a partner for a smarter AI Siri in a $1 billion a year partnership with Google.
Which of these should be our Hot Take Tuesday for Next week? (Gonna start prepping these now!)
Which recent AI news topic do you care more about?š³ļø Vote to see LIVE results š³ļø |
āļø
Jordan
P.S. - My wife said todayās show was āreally good.ā I think youāll agree.
Outsmart The Future
Today in Everyday AI
7 minute read
š Daily Podcast Episode: Is it 74% of companies who see an ROI on GenAI or 5%? Give todayās show a watch/listen for the answer.
šµļøāāļø Fresh Finds: Anthropicās valuation creeping to $350 billion, OpenAI warns of catastrophic AI, an open source LLM that beats GPT-5 and more. Read on for Fresh Finds.
š Byte Sized Daily AI News: Apple hires Google Gemini for smarter Siri, OpenAI backpedals on government backing talk, Trump official says no to AI bailout and more. . Read on for Byte Sized News.
š§ Learn & Leveraging AI: What can you learn from the recent (and well done!) Wharton study on AI? A lot. Keep reading for that!
ā©ļø Donāt miss out: Did you miss our last newsletter? We covered: Inside Google's new secret Presentation feature, IBM cuts thousands of jobs due to AI, Global stocks slide due to AI fears, Chrome users get big AI upgrade and more. Check it here!
Ep 648: How 74% of Enterprises Get Real AI ROI While Pundits See Failure
Is it AI failure or AI success? š¤
We see massive trillion dollar valuations for AI companies, yet constant āAI bubbleā bust stories.
And we see stories playing out in the media that say AI is both an enterprise boon and a complete waste of time.
Welpā¦. A new study from Wharton will hopefully put this to rest.
Among other things, it shows that 74% of companies see a positive ROI on AI and also lays out the roadmap to debunk all the naysayers along the way.
Also on the pod today:
ā¢ Is ROI at 74% or 5? š¤
ā¢ The secrets to AI success š¤«
ā¢ 5 Actions Steps from the study š¤«
Itāll be worth your 37 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 Dazl promises a new era of vibe coding, Ancher uses AI to stay informed and beat overwhelm, Everywhere brings AI capabilities outside of the browser.
AI Advancements ā OpenAI released a long blog post today, calling for more serious AI safety measures as they say AIās capabilities are growing faster than most people realize and that could carry catastrophic risks.
Social Media and AI ā Snap is bringing Perplexity's AI search to Snapchat, with a $400 million deal and answers coming straight from My AI.
Open Source AI ā Chinaās Moonshot AI just released Kimi K2 Thinking, an open-source model they say outperforms OpenAIās GPT-5, Claude Sonnet 4.5 and other rivals on key reasoning and coding benchmarks.
AI Valuations ā With its latest Google deal, Anthropicās valuation could be as high as $350 billion.
AI and Families ā A UK startup rolled out a way they say AI can predict IVF success with 90% accuracy and offers "no baby, no bill" guarantees.
ChatGPT Apps ā OpenAI released Peloton and Trip Advisor as available apps in ChatGPT.
@onepeloton and @Tripadvisor are now available as apps in ChatGPT.
ā OpenAI (@OpenAI)
6:57 PM ā¢ Nov 6, 2025
AI Improvements ā Annoyed having to wait 5-15 minutes to give more details to thinking models? OpenAI is rolling out a way to interrupt long-running queries. (Low key, this is a big deal for power users!)
AI in Nature ā AWS is funding a digital overhaul of world-renowned conservationist Jane Goodallās research, turning decades of handwritten notes and film into a searchable AI-powered database
1. Apple Taps Googleās Gemini for Siri Overhaul Next Year š¤
Apple is reportedly finalizing a $1 billion annual AI deal with Google to bring Geminiās 1.2 trillion-parameter AI muscle to a revamped Siri, targeting a spring 2026 launch, according to Bloomberg.
While Appleās own models lag behind in scale, Googleās tech will specifically boost Siriās ability to summarize and plan, but not take over all features. In a privacy-focused twist, Apple will keep user data on its own servers, ensuring Google wonāt see it.
2. Nvidia Slides as Trump Official Says No to AI Bailout š
Tech stocks sank Thursday after Nvidia dropped 3.6%, spooked by a Trump officialās flat-out denial of any federal bailout for the AI sector. The White Houseās David Sacks swiftly shot down speculation after OpenAIās CFO floated the idea of government-backed financing for pricey AI chips, prompting clarifications from top execs.
With chipmakers already uneasy over sector overvaluation and Nvidiaās recent $100 billion investment commitment to OpenAI, the market is on edge about the industryās tangled web of partnerships.
3. OpenAI Backtracks on Government Guarantee Talk š¤
OpenAIās CFO Sarah Friar quickly walked back comments suggesting the company wanted a federal guarantee for its massive AI spending, sparking a whirlwind of damage control just hours after her initial remarks at The Wall Street Journalās tech conference.
CEO Sam Altman then weighed in, clarifying that OpenAI isnāt seeking government bailouts and that failing companies shouldnāt get taxpayer rescues. The controversy drew a swift rebuke from White House officials, who said thereād be no federal safety net for AI firms.
4. OpenAI Reveals Sky-High Revenue and Trillion-Dollar Data Center Ambitions š¤
In response to the āgovernment guarantee-gateā above, OpenAI CEO Sam Altman made headlines today by announcing the company expects to top $20 billion in annualized revenue this year, with a staggering $1.4 trillion pledged for data center investments through 2033. The companyās ambitious roadmap includes new enterprise offerings, consumer devices, robotics, and even direct AI cloud computing services.
Altman also teased ventures into scientific discovery, while signaling OpenAI could pay for its expansion by selling more equity or taking traditional loans.
5. Microsoft Bets on "Humanist Superintelligence" Over AGI š¤
In a bold move this week, Mustafa Suleyman announced Microsoftās new AI division is building "Humanist Superintelligence" (HSI), aiming for practical, human-serving technology instead of chasing Artificial General Intelligence.
Suleymanās team rejects the race for limitless, autonomous AI and instead focuses on systems that solve real-world problems while keeping humans fully in control. This approach directly challenges OpenAIās ambitions, signaling a major shift in how Big Tech thinks about the future of AI development.
6. Google Unveils Its Most Powerful AI Chip Yet šŖ
Google just announced its seventh-generation Ironwood TPU chip and new Arm-based Axion processors, marking a major leap in AI infrastructure as demand shifts from training models to serving billions of real-world users.
Anthropic is betting big, committing to access up to one million of these chips in a deal worth tens of billions, making it one of the largest AI infrastructure agreements ever. Google is pushing custom silicon to challenge Nvidiaās dominance, aiming for better performance, reliability, and economics as the industry faces skyrocketing power and cooling demands.
š¦¾How You Can Leverage:
MIT released a viral study a few months back claiming 95% of enterprise AI pilots fail, and it dominated headlines for weeks while sending executives into panic mode.
The methodology?
52 interviews measuring success on the books at six months.
(Lolz)
Meanwhile, Wharton just dropped a three-year study tracking 800 actual enterprise decision makers with real budgets and found that 74% are getting legitimate ROI from GenAI investments. (read it here.)
It's about who's measuring reality versus who's selling you their solution disguised as research.
So on today's show, we're torching that MIT study's flawed methodology again (sorry) but more importantly breaking down Whartonās findings and four specific things separating the 74% of winners from everyone else still debating whether any of this technology actually works.
This is an episode you canāt afford to miss.
Letās get it.
1 ā Canvas mode beats every slide tool šÆ
The Wharton study created effectiveness scores for different AI task types across 800 companies.
Highest scores? Data analysis and document summarization. Legal contract review and HR recruitment showed massive returns too.
AI agents scored 58%.
That's it. 58% effectiveness while everyone dumps their entire budget trying to deploy agents.
The boring stuff nobody wants to talk about is where the actual money is. But here's the part that matters for you right now.
82% of leaders use AI weekly and 46% use it daily. Usage tripled since year one of the study. But investment in training dropped 8 points while leader confidence in that training collapsed 14 points.
Companies are spending more on AI while training less.
49% say recruiting AI talent is their biggest challenge. But they won't train internally. Organizations spent millions fine-tuning GPT-3.5 back in early 2023 with zero training budget for employees.
They assumed people would just figure it out.
Try This: Pick three repetitive tasks you do weekly. Time each one manually. Next week, use ChatGPT or Claude with prompts describing your exact workflow and output format. Multiply your time saved by 52. When you see 200+ hours recovered annually from three tasks, scale it across your team and measure the impact.
2 ā The perception gap kills momentum šØ
Talk about a gap.
56% of VPs are highly optimistic about AI adoption.
28% of managers share that optimism.
Half the confidence between the people approving budgets and the people actually using the tools. VPs see shareholder value and profit. Managers see job replacement and skill obsolescence.
Top-down AI implementation fails every single time because of this exact gap. Your executives experience the capabilities firsthand and get excited. Your frontline workers get licenses and confusion.
The companies winning? They're treating this as organizational transformation, not software deployment.
72% now formally measure GenAI ROI. That's good, but it should be 100%. If you're not measuring specific time savings, cost reductions, and revenue impacts with actual dollar amounts, you're flying blind.
Try This:
Stop all new AI tool purchases for 30 days.
Redirect that budget to training immediately. Find three employees who actually understand AI and have them run weekly 15-minute demos showing specific workflow improvements. Make it mandatory and require attendees to share their own use case by the end. Build internal expertise instead of trying to recruit external unicorns who don't understand your business.
3 ā 2026 separates winners permanently š
88% expect AI budget increases in the next 12 months based on results so far.
11% are already cutting legacy IT and HR budgets to fund more AI.
But here's what actually separates the winners from losers. Clear metrics tied to dollars. Focus on boring back-office automation. Strong executive alignment with frontline workers. Continuous training that never stops.
Not the model you're using.
Every model will be good enough soon. The gap is about execution and people, not technology.
But there's a hidden crisis even winners face. 89% say AI enhances their skills. 43% simultaneously fear skill atrophy from over-reliance.
The vanishing ladder problem is real.
AI automates the baseline tasks juniors practice on to build expertise. Your future senior talent never develops the domain knowledge to catch AI failures. When models hallucinate, nobody recognizes it.
You can't just bolt AI onto existing workflows. Upskilling and reskilling are the wrong frameworks entirely.
You need to unlearn everything and rebuild as AI-native from scratch.
Different processes. Different roles. Different metrics. Companies trying to add AI to old workflows are the ones failing in every study.
Try This:
Implement monthly AI-off days where teams work manually to maintain core expertise. Document your top five use cases with formal ROI tracking tied to revenue, time, or cost in actual dollars. Share these numbers with your entire org quarterly. Create fallback workflows for when tools fail or new releases break everything. Your operations need to survive GPT-6 sucking.
Reply