Last week, my co-founder and I got into a heated debate at the office over one deceptively simple question: is AI currently sitting in a massive valuation bubble?
His position was yes. His argument: much of today's corporate AI adoption is driven by executive fear of missing out rather than genuine operational necessity. C-suite leaders are rushing to buy enterprise AI licenses largely so they can announce "AI integration" in their next quarterly report.
My position was no. My argument: businesses aren't spending on AI for clout; they're spending because it solves real, bottom-line problems. It cuts operational drag, automates high-volume manual work, and reduces labor overhead in ways previous generations of software simply couldn't.
The funny thing is, once you look at the data, we were both right.
The Case for "Yes, It's a Bubble": The FOMO Trap. Recent enterprise surveys paint a telling picture: while the large majority of companies now use AI in some form, only a minority report a clear, measurable impact on their bottom line. A notable share of executives say they've seen neither revenue growth nor cost reduction from their AI investments over the past year.
My partner has a point. Billions are currently being spent automating workflows that were already broken to begin with , producing expensive software that ends up sitting idle on company servers, technically "adopted" but never actually driving value.
- The Case for "No, It's Real Value" , The Productivity Engine On the other side of the ledger, the companies that implement AI well are seeing genuine, substantial returns. A small tier of top enterprise adopters now attribute a meaningful share of their total bottom-line profit directly to AI-driven efficiency gains.
- Deployed properly, for fraud detection in fintech, supply chain optimization, or automated contract processing , firms are registering real, double-digit cost reductions. The technology itself isn't the problem. The businesses that are willing to actually redesign their workflows around it are the ones winning.
- The Verdict: The Hype Will Burst, the Technology Will Stay
- Just as in the 1990s Dot-Com era, companies relying solely on buzzwords and hype will eventually face a market correction. But the underlying technology isn't going anywhere , it's already too deeply embedded in how modern businesses operate. The shift we're moving into isn't about buying shinier AI tools. It's about building the legal and operational infrastructure around them properly:
- - Data foundation. You cannot automate chaos. A clean, well-structured internal data architecture isn't optional , it's the prerequisite for any AI system to work.
- - IP and contract governance. When a custom AI tool generates outputs based on your inputs, who actually owns them? This question needs a clear contractual answer before deployment, not after a dispute arises.
- - Regulatory proofing. Automated decision-making tools need to comply with local data protection law , including, in Kenya, the requirements set out by the Office of the Data Protection Commissioner (ODPC).
- Settling the Office Debate
- The verdict: the bubble of "AI hype" will pop. Overhyped companies coasting on buzzwords will face a reckoning, the same way Pets.com and its peers did in 2000. But the era of practical, system-driven AI , built on solid data, clear governance, and regulatory compliance , is just getting started.
- Who do you think won the debate? Team "FOMO Bubble" or Team "Bottom-Line Reality"? I'd love to hear your take.