Login to Continue Learning
Venture capital investment surged to a 10-quarter high of €108.3 billion in Q1 2025, fueled by artificial intelligence (AI), which accounted for over €44.6 billion raised.
For years, AI felt like a money-printing machine. Investors eager to avoid missing out on the next big thing were quick to back almost any startup that mentioned AI in their pitch deck. The idea didn’t need to be particularly well-implemented or useful. In some cases, even the illusion of innovation was enough to earn a unicorn valuation. But investors are now wising up to AI-washing—the practice of exaggerating a company’s AI use or capabilities.
As the CEO of Gradient Labs—an AI customer service platform for highly-regulated industries—I’ve seen investors grow wary of this trend. And understandably so, because despite its promise, AI comes with plenty of risk. Gartner predicts 40% of agentic AI projects will be cancelled by 2027, while MIT research shows 95% of pilot projects fail. Even Sam Altman has stressed we’re in the middle of an AI bubble.
History shows these spikes don’t last forever. While AI remains a hot sector, overall VC investment fell by 21% between Q1 and Q2, suggesting the days of easy capital are ending—and startups can no longer rely on buzzwords to stay afloat.
Related Post
Despite the slowdown, I recently led Gradient Labs through a one-week €11.1 million Series A. What did I learn? Investors aren’t focused on missing out on the gold rush; they’re now focused on whether companies can actually deliver. They don’t want promises but proof: demos that work, products that sell, and customers who back up lofty claims.
Being an “AI-native startup” won’t cut it anymore. Sprinkling industry lingo across a pitch deck may have secured term sheets in the past. But being an AI-native startup is no longer a differentiator.
That doesn’t mean the opportunity has passed. Most AI companies are pitching in the same space with no standout product or innovative vision—just founders trying to capitalize on hype. However, investors are becoming increasingly good at spotting AI-washing.
The upside is that genuine innovation—products built for a clear, specific use case—shines through. This is especially true when the founder and team genuinely understand the market they’re attempting to serve.
For my co-founders and me, it was never about building an AI startup just because it was lucrative. We wanted to solve a problem we had experienced while working at Monzo, a leading UK fintech: highly-regulated industries have been locked out of automation by strict compliance requirements. At Gradient Labs, we built a solution to tackle this challenge.
It wasn’t AI for the sake of it; it was AI with a purpose—and that made all the difference in the boardroom.
Products should be irreplaceable. AI is advancing rapidly, and what feels novel today could be normal tomorrow. You have to consider what makes you stand out and whether that will still be the case when pitching your product. How likely is OpenAI to solve the issue with its next GPT model release? If the odds are high, you’re going down the wrong path.
Our approach was to focus on hiring people with deep expertise, designing something truly different, and proving it works. We didn’t want to create an agent that gave good information 95% of the time. In highly-regulated industries, even a single mistake can cause reputational damage impossible to recover from.
We spent 14 months obsessing over the product, not the pitch. Every detail had to be perfect before we went live, and it paid off: the platform consistently outperformed human customer service agents—and customers were genuinely impressed.
As a result, we didn’t need to rely on flashy marketing or inflated promises to get VC attention. They could sense the quality, see the metrics, and recognize the category-defining potential.
Product is primary, but who you know makes a difference—especially when distrust is high. We laid the groundwork for months before our funding round, meeting investors and sharing updates.
By the time we were ready to pitch, we weren’t just another email; we were continuing conversations with people who already knew us and our story. For investors, this meant they already had the opportunity to gauge our credentials, verify claims, and speak to customers. They knew we were legitimate, and when it came to investing, they were ready to move.
Not everyone will say yes, but even noes are valuable. VCs are in the business of networking, and word gets around. The relationships we had built and trust we gathered meant many were willing to open doors for us, even if they ultimately passed up the opportunity. That network effect created momentum—lending credibility, sparking urgency, and signaling that what we’re building is worth serious consideration.
The AI boom may be cooling, but founders have nothing to fear. There’s still plenty of capital available—as long as you intend to solve, rather than deceive.
