Venture capitalists are becoming increasingly cautious about 'AI-washing' and are focused on supporting genuine innovation.

      Venture capital investment reached a 10-quarter peak of €108.3bn in Q1 2025, driven by artificial intelligence, which accounted for more than €44.6bn of that total. In recent years, AI has seemed like a machine for generating money. Investors, eager to capitalize on the next big opportunity, were quick to support nearly any startup that included AI in its pitch. The concept didn’t need to be particularly well-executed or practical, and in some instances, even a façade of innovation was sufficient for a unicorn valuation. However, investors are now becoming more discerning regarding AI-washing, which is the practice of overstating a company’s AI involvement or capabilities.

      As the CEO of Gradient Labs, an AI customer service platform catering to highly-regulated sectors, I’ve observed investors becoming increasingly cautious of AI-washing. This caution is justified, as, despite AI’s potential, it carries significant risks. Gartner anticipates that 40% of active AI projects will be abandoned by 2027, while MIT's research indicates that 95% of pilot initiatives fail. Even Sam Altman, considered one of the sector's major beneficiaries, has highlighted that we are currently experiencing an AI bubble.

      History has shown that these highs do not last indefinitely. While AI continues to be a hot area, total VC investment fell by 21% from Q1 to Q2, indicating that the era of easy funding is coming to an end — and startups can no longer rely on trendy terminology to survive.

      Despite this slowdown, I recently guided Gradient Labs through a week-long €11.1mn Series A raise. What did I learn? Instead of worrying about missing out on the gold rush, investors are increasingly interested in whether companies can genuinely deliver results. They prefer proof over promises: functional demos, sellable products, and customers who can confirm bold assertions.

      Claiming to be an “AI-native startup” is no longer enough to stand out. In the past, weaving industry jargon into a pitch deck may have secured funding, but it no longer differentiates a company.

      This doesn’t mean opportunities have disappeared. Many AI companies are vying for attention in the same market without any unique product or innovative vision — just founders attempting to ride the hype wave. However, investors are getting better at identifying AI-washing.

      The silver lining is that authentic innovation — products designed for a specific, clear purpose — is becoming more noticeable. This is particularly true when the founder and team have a genuine understanding of the market they aim to serve.

      For my co-founders and me, our goal was never simply to create an AI startup due to its profitability. We aimed to resolve a problem we encountered while working at Monzo, a leading fintech company in the UK: highly-regulated sectors have been shut out of automation due to stringent compliance demands. At Gradient Labs, we developed a solution to address this issue.

      Our focus was not on AI for its own sake but on AI with a meaningful objective — and that made a significant difference in discussions with investors.

      In a rapidly advancing AI landscape, what feels innovative today may soon become commonplace. It’s crucial to reflect on what distinguishes you and whether that uniqueness will persist in your product pitches. How likely is it that OpenAI will address your problem with the next GPT model release? If the odds are high, you may be heading down the wrong path.

      Our strategy concentrated on hiring individuals with substantial expertise, creating something truly distinct, and validating its effectiveness. We didn’t aim to develop an agent that provided accurate information only 95% of the time. In highly-regulated sectors, a single error can result in damaging and irreparable reputational harm.

      We dedicated 14 months to refining the product, rather than just the pitch. Every aspect had to be meticulously crafted before launch, and it resulted in consistent performance that exceeded human customer service agents — and clients were genuinely impressed.

      As a consequence, we didn't have to rely on flashy marketing or exaggerated promises to attract VC interest. They could perceive the quality, evaluate the metrics, and recognize the potential to redefine the category.

      Building relationships is essential, especially in an environment where skepticism reigns. We established connections and shared updates with investors for months leading up to our funding round.

      By the time we were prepared to pitch, we were not just another email in an inbox; we were continuing discussions with individuals who were already familiar with us and our narrative. For investors, this meant they had the chance to assess our credentials, validate our claims, and speak directly with our customers. They recognized our legitimacy, and when it was time to invest, they were ready to proceed.

      Not every investor will agree, but even rejections can provide valuable insights. VCs thrive on networking, and information circulates quickly. The relationships we fostered and the trust we built meant that many were inclined to assist us, even if they ultimately chose not to invest. This network effect generated its own