2 weeks Ago
Eoin Hinchy is cofounder & CEO of Tines , the no-code automation platform for security teams, enabling you to achieve more with less. Amidst all the hype around artificial intelligence, I’ve seen lots of technology companies rush out AI tools that feel distinctly bolted on. In most cases, these tools perform very well during demonstrations.
This is no surprise—in demos, it’s possible to tailor AI to make it look and feel impressive. Vendors can carefully control the AI model using clean data and specific use cases that it can easily understand. As a result, AI always gives satisfactory results.
It almost seems like magic. But the reality is usually very different. When you see AI being demoed or tested in a proof of concept, you’re essentially looking at a sanitized version of it.
In the real world, data is often messy, crucial integrations may not be available, and there are typically a lot of noisy alerts that diminish visibility. This means when AI-driven technologies are deployed in different environments, their actual performance or usability often falls short of what users expected after the initial sales pitch, leaving customers feeling disillusioned and misled. The Danger Of "Pew Pew" Tools This isn't a new phenomenon—we've seen the consequences of "AI washing" before, as in the case of the infamous "pew pew map.
" A few years ago, a well-known cybersecurity company demoed a real-time threat map of cyberattacks happening worldwide. It was impressive, with flashy graphics and animations that gave the impression that a dramatic and intense situation was unfolding. As it turned out, the map was make-believe and not based on real data.
This led to widespread ridicule from the security community, earning it the nickname "the pew pew map" and giving rise to the term "pew pew technologies" to refer to demoware technologies that look good in sales meetings but have limited capabilities in reality. For security teams, the consequences of overhyped AI tools are very real. Security practitioners already face long hours and a rapidly growing number of data-related threats to analyze and remediate.
When solutions fail to deliver, they only add more complexity to an already overwhelming environment. With 63% of security practitioners reporting burnout, these teams need solutions that reduce stress and make their lives easier—not add to their already huge workloads. Avoiding The Demo Trap The incentive to bring AI features to market quickly stems from the pressure companies feel to fuel growth.
In response, many succumb to rolling out their own, usually half-baked, AI capabilities, leading to a cycle of disillusionment throughout the industry. Security practitioners are often tasked with implementing such technologies, only to find that they don’t live up to their initial portrayal. It’s easy to see, then, why they are among the hardest to convince about the supposed benefits of AI—they’ve been burned too many times before.
This post-demo disappointment can be avoided by looking beyond the promises of AI marketing and instead focusing on solutions that can be easily integrated and deliver real value. For example, does the tool offer the flexibility and scalability to grow alongside your business, or will it struggle to integrate with your technology stack or support large volumes of data? There’s no point in investing in something that doesn’t solve problems and help you increase your impact. Security and privacy are also critical.
Make sure the solution you're considering keeps data secure within your infrastructure, with strong guarantees like no logging or training on your data. Usability and accuracy are key considerations, too. If the technology is difficult to implement, requires extensive training or generates a lot of false positives, it may cause unnecessary bottlenecks.
Finally, consider the tool’s ability to maintain performance when handling complex tasks. Remember, things are moving fast—ensure the vendor provides ongoing updates and improvements to keep the solution relevant as your business evolves. AI That Works When it comes down to it, the most effective tools are those that slot seamlessly into your existing environment and enhance what you already do well.
For security teams, this means keeping the business secure and ensuring threats are detected and mitigated swiftly. We still have a lot to learn about AI, and we're really only seeing the beginnings of what it can offer. In the rush to embrace it, let's be careful not to choose buzzwords over substance.
Forbes Technology Council is an invitation-only community for world-class CIOs, CTOs and technology executives. Do I qualify?.
