Don't Pair an F1 Engine with a Flintstone Frame

Hi!

Over the past few weeks, one thing’s stood out in my conversations, new model releases, and industry reports: the AI tools are finally maturing, but most businesses are still stuck trying to tack them on to on messy systems, brittle data, half-formed strategies, and dated processes.

The gap between potential and readiness is where the real story is right now as you’ll see below.

What they’re saying: IBM’s latest CEO study echoes what we’re seeing at Switchboard: AI has huge promise, but success starts with strong data and systems.

Why it matters:

Only 25% of AI projects are hitting ROI — not because AI is overhyped, but because many teams are trying to scale it on top of tangled tech stacks. The good news? CEOs are catching on. They’re realizing that cleaning up data and tightening infrastructure is step one, not an afterthought.

This isn’t a setback — it’s a signal. The real AI payoff comes when your foundation is ready to support it.

📉 Fast Moves, Shaky Foundations

What’s broken:

• 61% of CEOs are scaling AI agents — but 50% say they’re behind the 8-ball on this because of fragmented systems .

• 68% say they need integrated data across departments, but most haven’t built it yet .

• Only 16% of AI initiatives have scaled company-wide.

TL;DR: Rushing into AI without fixing your data is like flooring it with no gas in the tank.

🚀 Smart CEOs Are Doing This Instead

Here’s what’s working:

• Get serious about your data layer. Think enterprise-wide architecture, not isolated fixes. Clean, governed, connected data is AI rocket fuel.

• Prioritize ROI, not FOMO. 65% of CEOs are shifting to AI use cases with clear returns. Flashy demos don’t pay the bills .

• Build talent or borrow it. 54% are hiring for AI roles that didn’t exist last year; 67% say partnerships are key to filling the gaps .

🧭 Bottom Line

CEOs who win won’t just adopt AI — they’ll rebuild their tech foundations so AI can actually deliver. Want to compete? Start with your data.

Check out the full report here →

Ramp has so much data and they share it in such great, simple ways.

Their latest data reveals how quickly businesses are embracing AI—and who’s pulling ahead in the race.

🏭 Sector Surge: Tech, Finance, and Manufacturing Climb

The data shows: AI adoption rates by sector show tech (69%) and finance (55%) leading, with manufacturing (37%) catching up fast.

Why it matters: Core industries are now operationalizing AI with real spend happening and serious use cases taking root.

🧱 Mid-Market Momentum: Bigger Moves from Medium Biz

The data shows: Mid-sized businesses jumped to 42% adoption, with small businesses (36%) not far behind. Large enterprises still lead at 47%.

Why it matters: AI isn’t just for giants anymore. The mid-market’s rapid rise shows increasing adoption & democratization.

What this means for you: If you're a mid-market leader, the question isn't whether to adopt AI, but how to do it strategically while your competitors are still figuring it out. The window for competitive advantage is still open, but it's closing fast.

⚙️ Model Wars: OpenAI Dominates

The data shows: 40% of U.S. businesses now pay for AI tools. OpenAI leads with 32% of paid usage—far outpacing Anthropic (8%) and Google (0.1%).

Why it matters: OpenAI isn’t just winning, it’s setting the pace. A sharp rise in Q1 2025 suggests AI budgets are shifting decisively in its direction.

Check out the index here →

Why it matters:

The latest release from Anthropic (Claude Opus 4 and Sonnet 4) doesn't just beat benchmarks. It changes expectations for what AI should do inside a business.

📌 The shifts Claude 4 brings

1. AI that actually finishes what it starts. Opus 4 can now run tasks for hours without crashing or losing context. That's a big leap from chatbots that flake out after 2 minutes. It means your back-office automations can finally scale with minimal supervision.

2. Precision meets intelligence. Claude now follows complex instructions 65% more accurately and can reason while pulling from web, APIs, and local files simultaneously. No more manual orchestration layers or vague outputs on business-critical flows.

3. Budget-conscious power. Sonnet 4 offers much of the new capability — deeper thinking, multitool use, persistent memory — at a fraction of Opus's price. That makes real AI operations viable for companies that aren't hyperscalers.

🧭 For those debating when to look at adopting AI

Every frontier model is now rushing toward one use case: agents that run your business logic, not just reply to questions. Claude 4's edge is how stable and coherent it is over long tasks.

If you've hesitated on AI because it felt flaky or "too beta," Claude 4 might be your moment.

Check out the full Claude 4 summary →

The challenges Dan Cremons outlines here are exactly what we’re seeing at both PE and non-PE-backed companies alike.

His perspective is rooted in real operational experience — and resonates with the kinds of AI hurdles we’re seeing businesses look to overcome daily. (I also had the chance to give Dan feedback on his new book on AI in PE-backed firms before release and it’s one worth checking out.)

🚧 Insights from Dan on How AI initiatives go off the rails:

1. No clear strategy

   As Dan puts it: AI can’t support a strategy that doesn’t exist. Directionless = pointless.

2. Scattered experiments

   FOMO isn’t a roadmap. Random pilots waste time, focused initiatives win.

3. Chasing the wrong use cases

   High-shine ≠ high-impact. Effective AI starts with solving real business challenges, not novelty.

4. Fuzzy success metrics

   You can’t improve what you don’t measure. Define clear, quantifiable outcomes.

5. Lack of executive buy-in

   If leadership isn’t aligned and engaged, the rest of the org won’t be either.

6. Internal resistance

   Even great tech flops without team adoption. People-first rollouts matter.

7. Sloppy deployments

   Tight execution beats big ideas. Clear milestones and follow-through make the difference.

8. Overlooking risks

   Risks don’t disappear by ignoring them. Spot them early and plan accordingly.

Dan’s breakdown is a blueprint for getting it right. It all starts at the top: leaders must modernize, define where the business is going, and commit to guiding their teams there with clear, focused AI transformation strategy.

➡️ Give Dan a follow Dan on LinkedIn. His perspectives on aligning strategy, people, and AI is among the best out there. A must-follow if you’re navigating transformation.

💡 PS, we help PE firms figure this stuff out.

Chat soon 👋