It’s time.
It’s time to call it. June 30, 2025 marks the time we need to wrap up the era of stalling, of waiting to see in AI. Because we’ve waited, and seen, and learned. We’ve officially crossed the threshold where “AI-first” has moved from competitive advantage to table stakes.
If your engineering team, your product team, or your leadership team hasn’t shipped meaningful AI capabilities into production by now, that have led to a material increase in revenue — you don’t have a strategy problem. You have a talent (and vision) problem.
The 18-Month “Let’s See” Window Has Closed
Let’s be brutally honest about the timeline here. ChatGPT launched in November 2022. Claude, GPT-4, and the foundation model explosion happened through 2023. By early 2024, every serious B2B company had access to the same foundational AI capabilities through APIs that cost pennies on the dollar.
That gave everyone roughly 18 months to figure it out. Eighteen months to experiment, to build, to learn, and to ship something that moves the needle.
The companies that used those 18 months wisely aren’t just winning—they’re lapping the competition. Look at the hypergrowth AI B2B companies that are redefining what’s possible:
- Anthropic (Claude): Hit $3 billion in annualized revenue by May 2025, up from $1 billion in December 2024 — that’s 200% growth in 5 months
- Sierra: Reached $50M ARR in 2024 and $4.5 billion valuation — 300%+ growth rate for an AI customer service platform
- Cursor: The fastest-growing SaaS company ever, hitting $100M ARR in just 12 months and now at $200M ARR — 10x faster than traditional SaaS
- Loveable: Reached $1M ARR in 8 days, $10M in 2 months, and $60M ARR in 6 months — Europe’s fastest-growing AI startup ever
- Harvey AI: Surpassed $50M ARR and targeting $100M within 8 months at a $3B valuation — revolutionizing legal tech
- Perplexity: Crossed $100 million in annualized revenue just 20 months after launching premium subscriptions — transforming AI search •
- Mercor: Hit $75M ARR in 2 years with 51% month-over-month growth and $2B valuation — AI-powered recruiting that’s already profitable
- Owner.com: Hit $50M ARR growing 150% YoY at $1B valuation with AI-powered restaurant tech — helping local businesses compete with giants like Domino’s
- RevenueCat: AI has turbocharged it powering $1B+ in mobile subscriptions for leaders including … ChatGPT.
- ElevenLabs rocketing to $100m+ ARR powering voice for AI leaders.
- Gong: Reaccelerated to $300M+ ARR driven by AI features seeing 400%+ YoY growth and 50% usage increases — revenue intelligence powered by AI
- Dialpad: Surpassed $300M ARR with 50%+ YoY growth, fueled by DialpadGPT and 250M+ AI Recaps in 6 months — communications intelligence at scale
- Palantir: Revenue acceleration from 13% in 2023 to 36% in Q4 2024, driven by AIP launch — proving even established companies can reboot with AI
Meanwhile, if you’re still “exploring AI initiatives” or “evaluating use cases,” you’re not behind. You’re irrelevant. Even if your existing customers aren’t leaving. Yet.
What “Truly Great AI” Actually Means
I’m not talking about slapping a chatbot on your landing page or adding “AI-powered” to your marketing copy. I’m talking about AI that genuinely transforms your core product experience in ways that create measurable business impact.
Great AI in production looks like:
Superhuman-level assistance that users can’t live without. Linear’s AI issue triaging doesn’t just categorize tickets—it predicts resolution time, suggests optimal assignees, and auto-generates technical context that saves engineering teams 2-3 hours per sprint. Users report they “feel helpless” working in other project management tools now.
Invisible intelligence that makes your product fundamentally better. Figma’s AI design suggestions don’t feel like a separate feature—they feel like the product got smarter. Conversion rates on design handoffs increased 60% because the AI anticipates developer needs during the design process.
Workflow transformation that creates new value. Loom’s AI meeting summaries didn’t just transcribe—they created entirely new workflows around asynchronous collaboration. Teams that adopted it saw 30% reduction in follow-up meetings and 50% faster project velocity.
The common thread? These aren’t AI features bolted onto existing products. They’re AI-native experiences that reimagine what the product can do.
The Talent Reckoning
Here’s what I’ve learned from talking to 200+ SaaS founders over the past six months: The companies shipping great AI aren’t necessarily the ones with the biggest AI budgets or the fanciest ML infrastructure.
They’re the companies that made hard decisions about their teams early.
The CTO who said “we don’t need AI people, our engineers can figure it out“ is now 12 months behind companies that hired AI-native talent in early 2024. Traditional software engineering skills and AI engineering skills have meaningful overlap, but they’re not the same thing. Prompt engineering, model fine-tuning, vector databases, retrieval-augmented generation—these aren’t concepts you pick up over a weekend.
The product leader who treated AI as “just another feature” missed that AI requires fundamentally different product thinking. Great AI products aren’t built by adding smart components to dumb workflows. They’re built by reimagining the workflow around what AI makes possible. This requires product leaders who viscerally understand the technology, not just the market opportunity.
The CEO who delegated AI strategy to someone who “really gets AI” discovered too late that AI strategy is business strategy. The companies winning with AI aren’t optimizing existing processes—they’re creating entirely new business models. That requires leadership that understands both the technology possibilities and the market implications.
The Reboot Playbook
If you’re reading this and realizing you’re in the “reboot” camp, here’s the harsh but actionable truth:
Stop trying to retrain your existing team. I’ve watched dozens of companies spend 6-12 months trying to upskill engineers who fundamentally don’t believe in AI-first development. Meanwhile, their competitors hired AI-native talent and shipped three product iterations. You can’t afford the learning curve anymore.
Hire for AI-first thinking, not AI expertise. The best AI hires I’ve seen aren’t necessarily the ones with the deepest ML backgrounds. They’re the ones who instinctively think about problems through an AI lens—who see a manual process and immediately envision how LLMs could transform it, who understand that great AI products feel magical because they anticipate user needs.
Give your new AI team real authority. Half-measures don’t work in AI. The companies succeeding are the ones where AI engineering sits at the leadership table, where AI considerations drive product roadmap decisions, where AI capabilities influence go-to-market strategy. If your AI team reports to someone who doesn’t fundamentally believe in AI-first product development, you’re setting them up to fail.
Accept that you’re starting over. Your existing product roadmap, your engineering processes, your QA workflows—much of it was designed for deterministic software in a pre-AI world. AI products require different testing methodologies, different deployment strategies, different success metrics. Fighting this reality will slow you down more than embracing it.
The Competitive Reality
The uncomfortable truth is that we’re already seeing market separation. Companies with great AI in production aren’t just growing faster—they’re fundamentally changing customer expectations in their categories.
Customers who experience Notion’s AI writing don’t just prefer it to other wiki tools—they find other wiki tools frustratingly primitive. Users who work with GitHub Copilot don’t just code faster—they find traditional IDEs limiting and outdated.
This isn’t about features anymore. It’s about raising the bar for what software can do.
Your competitors with great AI aren’t just winning deals. They’re redefining what winning looks like in your category. And if you’re still debating whether AI is hype or reality, you’ve already lost the positioning battle.
Big Companies Don’t Really Have More Time
Yes, bigger companies have a bigger base to protect, and move more slowly in general. But just becausae your installed base isn’t leaving, doesn’t mean you new customer acquisition hasn’t already materially slowed due to AI competitors. And that there isn’t “stealth churn” in your base.
The Path Forward
The excuses have now run out.
- “We’re a small company” doesn’t work when two-person startups are shipping AI features that feel more sophisticated than your enterprise product.
- “Our customers aren’t asking for AI” doesn’t work when your customers are using AI tools to work around your product’s limitations.
- “We need to see more ROI data” doesn’t work when your competitors are creating entirely new value propositions that make ROI comparisons irrelevant.
The companies that will dominate the next decade of SaaS aren’t the ones with the best AI today. They’re the ones building AI-native cultures, AI-first product experiences, and AI-enabled business models.
If your team hasn’t shipped great AI into production yet, the question isn’t whether you need to make changes.
The window for gradual transformation has closed. The window for dramatic transformation is still open.
But not for much longer.
