THINXSTER
Blog/AI Marketing
AI Marketing7 min readAugust 18, 2026

What Companies Are AI Adopters? The 3 Tiers, Explained

AI adopters split into three tiers: tech firms selling AI, enterprises like JPMorgan and Walmart running it in core ops, and the 44% of US businesses paying f

RK
Ryan Korsz
Founder & CEO, Thinxster

TL;DR

AI adopters split into three tiers: tech firms selling AI, enterprises like JPMorgan and Walmart running it in core ops, and the 44% of US businesses paying f

→ See how this applies to your business (free 30-min call)

AI adopters fall into three tiers. First, the companies where AI *is* the product — Microsoft, Google, Amazon, Meta, Nvidia, Salesforce, Adobe, Palantir. Second, large non-tech enterprises that have pushed AI into core operations: JPMorgan Chase (its LLM Suite rolled out to roughly 200,000 employees), Walmart, Moderna, Verizon, Klarna, Yum! Brands, Lemonade, John Deere, IKEA, and Duolingo. Third — the biggest and least-covered group — the roughly 44% of US businesses now paying for AI tooling as of mid-2025 (Ramp AI Index), including tens of thousands of HVAC companies, law firms, dental groups, and roofers running AI call answering and scheduling. Adoption is wildly uneven: US Census data shows the information sector near 25% adoption against about 4–5% in construction and food service.

The three tiers, and where your company actually sits

The word "adopter" is doing a lot of hiding. A useful split:

  • Tier 1 — AI-native. Revenue depends on AI models. Nvidia's data center revenue went from about $15 billion in FY2023 to over $115 billion in FY2025. These companies are not a template for anyone else.
  • Tier 2 — Enterprise deployers. AI is embedded in a named workflow with a budget owner. Klarna's AI assistant handled the equivalent workload of about 700 full-time agents and drove a reported $40 million profit improvement in 2024. Moderna deployed thousands of internal GPT instances across research and legal. Walmart uses generative AI to enrich millions of product listings.
  • Tier 3 — Tool subscribers. They pay $20–$500/month for ChatGPT, Jasper, an AI receptionist, or a CRM with AI features bolted on. This is where nearly every US service business under $50M in revenue sits. Most Tier 3 "adopters" have no baseline metric, no owner, and no idea whether the tool made money.
  • The gap between Tier 2 and Tier 3 is not budget. It's whether anyone measured anything before the tool showed up.

    What the adoption numbers actually say

    The headline statistics conflict because they measure different things:

  • McKinsey's State of AI puts organizational AI use at roughly 78% in at least one business function, up from about 55% in 2023. That's a survey of large firms, self-reported, and "one function" is a low bar.
  • The US Census Bureau's Business Trends and Outlook Survey — which asks all US firms, including the corner plumbing shop — put AI use around 3.7% in late 2023, climbing past 9% by 2025. That number is far lower because it's the real denominator.
  • Ramp's AI Index, drawn from actual corporate card spend rather than surveys, showed paid AI adoption rising from roughly 5% of businesses in early 2023 to about 44% by mid-2025.
  • Spend data beats survey data. A company that tells a surveyor it "uses AI" may mean one marketing coordinator has a free ChatGPT account. A company with a recurring $2,400/year line item made a decision.

    The honest read: about 9% of US businesses have real AI in an operational workflow, roughly 44% are paying for something, and close to 78% will tell a surveyor they've adopted. Pick which number you're benchmarking against before you panic about falling behind.

    Which industries are adopting fastest — and which are barely moving

    Census BTOS data by sector, roughly as of 2025:

  • Information / software / media: ~25% adoption — the clear leader.
  • Professional, scientific, and technical services: ~18% — law, accounting, consulting, agencies.
  • Finance and insurance: ~12–14%, heavily concentrated in fraud detection and document review.
  • Healthcare: ~8%, dominated by ambient clinical documentation (Abridge, Nuance DAX) rather than anything patient-facing.
  • Construction, transportation, accommodation and food service: ~4–6% — the laggards, and also where labor scarcity makes the ROI case strongest.
  • Firm size matters more than sector. Companies with 250+ employees adopt at roughly 3–5x the rate of firms under 10 employees. That's not because small firms are dumb. It's because a 6-person company has no one whose job is to evaluate software.

    What "AI adopter" looks like at a $2M service business

    This is the part the enterprise-trend articles skip entirely. A genuine Tier 2-style deployment at a local service company usually looks like three specific things, not twenty:

  • Speed-to-lead automation. Inbound lead gets a text and a call attempt in under 60 seconds. Lead response research going back to the Harvard Business Review study on the subject found contact rates drop by roughly 10x between a 5-minute and a 30-minute response. Cost: typically $200–$800/month in tooling.
  • AI voice answering for overflow and after-hours. Missed calls at most home-services companies run 20–35% of inbound volume. If your average job is $600 and you miss 40 calls a month, that's a real number you can compute. Voice agent platforms run roughly $0.07–$0.20 per minute plus a platform fee.
  • Content and review operations. Responding to every Google review within 24 hours and publishing service-area content on a schedule. This is where AI cuts cost, not where it creates advantage — everyone has the same tool.
  • We've broken the underlying benchmark data out in more depth on our AI marketing statistics page, and you can run your own numbers on the ROI calculator before you talk to anyone, including us.

    The uncomfortable part: most AI adoption produces nothing

    This is the section our competitors won't write, and it's the most important one here.

    MIT's NANDA initiative published a 2025 study finding that roughly 95% of enterprise generative AI pilots delivered zero measurable P&L impact. Not negative — zero. Gartner projected that about 30% of generative AI projects would be abandoned after proof-of-concept by the end of 2025. S&P Global found the share of companies abandoning most of their AI initiatives jumped from about 17% to 42% in a single year.

    The failure modes are consistent and boring:

  • No baseline. If you can't state your current close rate, cost per lead, and missed-call percentage, you cannot detect improvement. You'll renew the subscription on vibes.
  • Buying a tool instead of changing a workflow. An AI SDR that books appointments into a calendar nobody checks produces $0.
  • Quality regression that shows up late. Klarna publicly walked back part of its AI-first customer service push in 2025 and began rehiring human agents after service quality slipped. The company that produced the best AI customer-service case study in the industry also produced the best cautionary tale, about 18 months later.
  • Content commoditization. When every competitor in your metro publishes AI-written service pages, the marginal value of AI-written service pages goes to roughly zero. Google's helpful-content and spam updates have repeatedly demoted sites that scaled thin pages.
  • Integration debt. Most small service businesses have their data split across a field-service platform, QuickBooks, a phone system, and a spreadsheet. AI on top of fragmented data produces confident nonsense.
  • Who should not buy AI marketing right now

    Plainly, and against our own interest:

  • If you're doing under roughly $500K in annual revenue, a $3,000–$8,000/month agency retainer is a bad allocation. Fix your Google Business Profile, get to 100+ reviews, and answer your phone. That costs almost nothing and beats most AI deployments.
  • If your close rate on existing leads is under 20%, more leads will not help you. You have a sales problem. Sending 300 AI-generated leads into a broken intake process wastes both.
  • If you can't staff the work you already have, marketing of any kind is the wrong purchase this quarter.
  • If you want AI content at volume with no editorial review, don't hire us — hire the cheapest vendor, because that's a commodity and you'll be disappointed at any price.
  • If leadership won't commit 6–9 months, skip it. Organic and lifecycle programs rarely show clean attribution before month 4, and anyone promising results in 30 days is selling paid ads with an AI label.
  • Our pricing page is public specifically so people in those categories can disqualify themselves without a sales call.

    How to tell if you're a real adopter or a subscriber

    Three questions, answerable in ten minutes:

    1.

    Name the metric. What number should move, what is it today, and who checks it monthly?

    2.

    Name the owner. Not a vendor — an employee whose review mentions it.

    3.

    Name the kill criteria. At what result, by what date, do you cancel? Companies without a kill date are the ones in the 42% abandonment statistic, they just find out 14 months late.

    If you can answer all three, you're a Tier 2 adopter regardless of your size, and you'll likely beat competitors ten times larger who bought more tools and measured none of them. If you can't, you're paying $20–$500 a month for the feeling of having adopted AI — which, based on the spend data, describes most of that 44%.

    Frequently Asked Questions

    What companies are considered AI adopters?

    Three groups. Tech companies where AI is the product — Microsoft, Google, Amazon, Meta, Nvidia, Salesforce, Adobe, Palantir. Large enterprises running AI in core operations, including JPMorgan Chase, Walmart, Moderna, Verizon, Klarna, John Deere, IKEA and Duolingo. And the roughly 44% of US businesses simply paying for AI tools.

    What percentage of US businesses use AI?

    About 44% of US businesses were paying for AI tooling as of mid-2025, according to the Ramp AI Index, which tracks corporate card spending. US Census Bureau survey data reports far lower numbers because it measures AI used to produce goods and services rather than any paid subscription.

    Which industries have the highest AI adoption rates?

    The information sector leads at roughly 25% adoption in US Census data, followed by professional and technical services. Construction and food service sit near 4–5%. The gap reflects how much of each industry's work is digital text and data versus physical labor performed on site.

    How is JPMorgan Chase using AI?

    JPMorgan Chase built an internal generative AI platform called LLM Suite and rolled it out to roughly 200,000 employees. Staff use it for drafting documents, summarizing long materials, and analysis. It is one of the largest deployments of internal AI tooling at a non-technology company.

    Free Weekly Briefing

    One AI Marketing Tactic.
    Every Tuesday. Free.

    What's actually working across our client accounts right now — ROAS moves, follow-up sequences, creative angles. The stuff that isn't in any blog post yet.

    No spam. Unsubscribe anytime. 1,200+ business owners already in.

    Ready to Deploy

    SEE THIS IN
    YOUR BUSINESS.

    30 minutes. We scope the exact systems that apply to your situation and give you a plan.

    ★★★★★ Trusted by 47+ local service businesses

    BOOK A STRATEGY CALL →