Animal Spirits Podcast — "Talk Your Book: The Two Biggest Stories of the Year: AI & Tariffs"

Date: November 8, 2025
Hosts: Michael Batnick & Ben Carlson
Guest: Ara Kharazian, Economist at Ramp

Episode Overview

In this episode, Michael Batnick and Ben Carlson are joined by Ara Kharazian, economist at Ramp and author of "Ramp Economics Lab." They dive deep into the two most dominant and influential stories of the year: the explosion of artificial intelligence (AI) adoption in business and the realities of tariffs and their economic impact. Drawing from Ramp’s unique business spend data, Ara brings fresh, data-driven insights on how AI is being adopted across sectors, discusses discrepancies with traditional government reporting, and unpacks what’s really happening with tariffs versus media narratives. The conversation is candid, analytical, and peppered with actionable takeaways for market watchers and investors.

Key Discussion Points & Insights

1. The AI Explosion: Data-Driven Perspectives

• The “AI Bubble” Debate (03:00 – 06:00)

  • Michael frames the question dominating markets: "Is AI a bubble, will it crash the stock market, or is it the start of a new industrial revolution?"
  • Ara emphasizes the lack of reliable public data, making Ramp’s spend data distinct: “A lot of companies are trying out AI, but are they using AI in a long term, concerted way? ... That’s the real question about whether or not we’re in a bubble.” (03:49, Ara)
  • Contrasts perception (stock valuations) with real-world business spend and adoption.
  • "Retention [of] AI products is higher, 80% now in 2024, about 50% in 2022. And AI contract sizes themselves are growing. We estimate that they'll hit about $1 million next year on average." (05:17, Ara)

• Enterprise vs. Broader Market Adoption (05:29 – 06:40)

  • Enterprise adoption is slower due to risk and liability concerns, especially in customer-facing roles.
  • "For enterprise, I think growth has been a little bit slower...They're hesitant to implement AI if it causes...liability at the enterprise level." (05:41, Ara)
  • But Ramp data reveals about 40% of companies on their platform have significant AI contracts, often tailored for specific teams.

• Discrepancies in AI Adoption Estimates (06:40 – 09:42)

  • Michael notes Ramp’s estimate for AI adoption (44%) versus the government’s much lower (9%).
  • Ara explains government data is based on outdated, confusing survey questions: "Does that include customer service automation by AI?...Because that's most of the way. The AI product ecosystem has developed since 2023." (07:50, Ara)
  • There’s underreporting both in official stats (due to bad survey design) and Ramp’s paid-use–only dataset (which misses free tools).

• Why Spend Data is a Better Indicator (09:42 – 12:47)

  • Ramp’s dataset tracks line-item receipts, revealing granular insights into which AI tools and services companies are actually paying for—and renewing.
  • Ara’s research helps business decision-makers benchmark their own adoption and look for productivity gains, rather than rely on traditional expensive consulting.

Memorable Quote:
"My entire job is to be public facing...the whole company's ethos is around allowing you to have access to better data about where your business spends money." (11:20, Ara)

2. Trajectory of AI Adoption & Bubble Risks

• Is AI on Track for “Internet Penetration” Levels? (12:47 – 14:57)

  • Michael asks if AI adoption will ever reach 90%+ like internet did.
  • Ara: “Short answer is yes... There are businesses that will never be fully using AI...but they will probably be using AI for back office tasks and marketing...Growth has slowed in adoption. The actual size of the contracts and...deepness of adoption is increasing.” (13:13, Ara)
  • Many decision-makers—even in large companies—aren’t yet familiar with key players, proof that “we’re still very early.”

• What Would Signal Trouble? (14:57 – 16:37)

  • Ara: If contract sizes or retention start to decline, then there's trouble: "If the size of contracts start to go down and if retention rates start to go down, that means that companies are trying these AI products and services and they're just not working for them." (15:11, Ara)
  • Stresses difference from dot-com bubble: companies building AI now are “producing technology people are buying...and they're revenue generating." (16:11, Ara)

• Tech Spend Trends and Potential “Losers” (16:37 – 18:46)

  • Michael questions whether non-AI software companies (e.g. Salesforce, Adobe) are seeing spending drop.
  • Ara: On the contrary, spending on legacy software is still growing—"It's way too soon to say who [the losers] are going to be." (18:09, Ara)

• Size of Companies: Who Benefits Most? (18:46 – 20:13)

  • Ben asks: Are smaller firms biggest beneficiaries?
  • Ara: Large companies with engineering teams can deeply integrate AI, especially in coding and customer service; small companies use AI mostly for lightweight tasks (marketing, social media), which may not be major GDP drivers but still provide real value.

3. AI & the Labor Market — “The Biggest Question”

• Impact on Jobs and Society (20:13 – 22:36)
  • Hosts discuss the "learn to code" industry’s pain, and the broader question of AI’s impact on labor.
  • Ben: “That might be the biggest question we see this century.” (20:37, Ben)
  • Ara offers a nuanced, optimistic take: Automation is affecting white-collar jobs, but “most jobs in the US economy today are not exposed at all. Most jobs require some kind of physical component...Nurse, waiter...None of those jobs are going to be automated anytime soon.” (21:15, Ara)
  • Early data suggests displacement is limited to specific sectors, with the overall unemployment rate stable.

4. Tariffs: Reality vs. Theory

• Announced vs. Actual Tariffs — What the Data Shows (23:01 – 27:02)
  • Ben raises the confusion that tariffs should drive up inflation, but effects seem muted.
  • Ara unpacks why: "We can see in ramp data tariffs, because we see that on invoices...And what we found was that tariffs are increasing...but it's much more slower and more granular than it than you'd think." (23:44, Ara)
  • Despite high-profile announcements, the actual share of invoices with tariffs is still just about 3% (up from 1.4% last year).
  • Implementation lags and legal frictions, plus companies restructuring supply chains, mean the adjustment is slower and less dramatic than headlines suggest.
  • Companies can reclassify or shift imports; full pass-through to costs is rare, and overall price increases are thus much less than economic theory would predict.

Memorable Quote:
“There are a lot of frictions to implementing these kinds of ideas. You need to actually figure out, okay, wait, what is actually tariffs? Because we have trade agreements that are still in place…” (24:41, Ara)

5. Quick Hits & Closing Topics

• Ramp Data in Practice & Ara’s Work (27:02 – 30:47)
  • Ara publishes weekly at Ramp Economics Lab and is open to ideas—"My business spend data set really allows us to...identify what the latest trends are. What's just a vibe that people are talking about in the news versus what is an actual thing that's going on in our economy right now." (27:06, Ara)
  • Hosts joke about tracking Chipotle prices/shrinkflation, and Ara details the quirks of spending data (28:04).
  • Briefly touch on the 996 working culture and increased Saturday work/eating out among San Francisco workers (29:07).
  • Ara pitches Ramp’s finance automation benefits—eliminating manual, annoying expense and receipt management through AI (30:00).

Timestamps for Key Segments

• AI “Bubble” & Spend Data Reality: 03:00 – 06:00
• Enterprise vs. Overall AI Adoption: 05:29 – 06:40
• Government vs. Ramp AI Data Discrepancy: 06:40 – 09:42
• Ramp Data & How It’s Used: 09:42 – 12:47
• Will AI Adoption Be Ubiquitous? 12:47 – 14:57
• What Would Make This an AI Bubble? 14:57 – 16:37
• Software Spend Trends: 16:37 – 18:46
• Who Benefits—Big vs. Small Firms: 18:46 – 20:13
• Labor Market Impacts: 20:13 – 22:36
• Tariffs: Why Theory ≠ Reality: 23:01 – 27:02
• Fun Data Applications/996 Culture: 27:02 – end

Notable Quotes

• On AI Adoption, Retention & Growth:
  "Retention AI products is higher, 80% now in 2024, about 50% in 2022. And AI contract sizes themselves are growing. We estimate that they'll hit about $1 million next year on average." (05:17, Ara)

• On Why Traditional Adoption Surveys Miss the Mark:
  "Goods and services is kind of like an econ speak term, and it was written a couple years ago...The way that most organizations are using AI...is not producing widgets in your factory." (07:50, Ara)

• On Tariffs Filter Through:
  "Despite the announcements, the average tariff rate was something like 1.4%. Now we're only about 3% as far as like share of invoices with a tariff transaction on them." (24:22, Ara)

• On Automation & Society:
  “Most jobs require some kind of physical component…None of those jobs are going to be automated anytime soon, if ever.” (21:15, Ara)

Tone and Style

The discussion is candid, nuanced, and data-driven, with Ara emphasizing clarity over hype and the hosts threading in humor and market skepticism. The episode is accessible but advanced, making data-centric arguments without technical jargon overload.

Summary Takeaways

• AI adoption is robust and accelerating among businesses, measured best with direct spend data, not outdated surveys.
• Large companies, due to resources and tech teams, benefit most from deep AI integration, but adoption is growing everywhere.
• We are early in the AI adoption curve; risk of bubble seems lower than in prior tech manias due to clear, ongoing business spend and revenue.
• Tariffs, despite headline coverage, have had muted inflationary pass-through effects due to many practical and structural frictions in implementation.
• Business spend data, like Ramp’s, brings essential ground-truth clarity to economic debates too often dominated by lagging, incomplete surveys.

For more insights and Ara’s latest research:

• Check out Ramp Economics Lab
• Explore Ramp’s public data at ramp.com/data