Bubble, Bubble, Toil and Trouble

AI’s transformation potential is undeniable, bubble or not

If you keep an eye on the AI space, you’ve probably noticed a shift in tone. Earlier this summer there was unbridled optimism – the thrill of new tools and bold predictions about what AI would mean for work and life.

Now the headlines are cooler: There was the messy launch of ChatGPT 5, the MIT study finding that 95% of AI initiatives fail, Sam Altman and The Atlantic suggesting that we’re in an AI bubble, insane paychecks for AI talent, and even an IBM commercial showing overwhelmed corporate leaders expressing frustration as they try to implement AI across the enterprise.

So, are we in a bubble? Possibly. But does that mean AI is smoke and mirrors? Not at all. What it means is that leaders need to look past the hype and focus on what really matters: practical wins, starting with time saved.

Is AI a Flash in the Pan?

I often ask people how long they think AI has been around. Most say 5-10 years, so are surprised when I tell them that it’s actually been around for more than 75 years, going all the way back to 1950 when Alan Turing asked, "Can machines think?"

Over that three quarters of a century, AI has gone through two other hype cycles, with the first one coming in the 1950s and 60s. Each previous peak was followed by an “AI winter” or “trough of disillusionment” and it seems that we may be inching toward another one now. But I would argue that things will be different this time simply because AI has become ubiquitous in the daily lives of hundreds of millions of people – embedding itself in everything from planning our work days to helping us find enlightenment.

The Reality of AI Adoption

The IBM commercial I mentioned above is called “Irony” because “adding AI that makes everyone else’s job easier to manage gives us more to manage.” The reality is that AI adoption isn’t magic; to get real value from it requires clarity, focus, effort and stick-to-it-iveness.

For example, I’m working with a large training and education company on a multi-pronged AI transformation project, and we’ve spent the last couple months evaluating AI-driven design technologies that can help the marketers keep up with ad creation demand.

We’ve mapped the process, established evaluation criteria, demoed a half dozen tools, established baseline performance for post-pilot comparison purposes, and are just about to start a pilot project to decide which tool will best meet the company’s needs.

The potential to save the company time and money is real and significant but getting here has taken much more than a snap of the AI fingers.

The Automation Potential is Real

For another client – a media buying agency – we’re evaluating ways that AI can make the media buying and reporting process more efficient and effective. A recent analysis of 809 insights across seven client campaigns showed that 77.6% of insights had high automation potential.​

• Month-over-Month Analysis​: 117 instances of "X% increase/decrease from last month"​
• Percentage-Based Metrics​: 216 instances of structured performance data (CTR, CPC, VCR, etc.)​
• Standardized Headers​: All reports use identical column structure: Date, Campaign, Tactic, KPI Performance, Benchmark, Creative Performance, Recommendations

Stop and think about that for a second. Reporting is a critical deliverable for a media-buying agency, and more than three-quarters of that process has the potential to be automated by AI. The time savings that represents is significant, and as I’ve argued before, time savings is a realistic first success metric for AI implementations.

What You Do with Time Matters

If time savings is the first measure of AI impact, then to quote Gandalf, “All we have to decide is what to do with the time that is given us.”

Agency teams can take that time saved and spend it in all sorts of powerful ways:

• Strategist: Mine client data for non-obvious growth drivers.
• Creative lead: Prototype bolder concepts, faster.
• Analyst: Design cleaner tests, retire dead metrics, sharpen benchmarks.
• Account lead: Spend more time with stakeholders, less chasing “where is it?” emails.

The possibilities are endless. What matters is that the time saved is used on those activities that a human is uniquely suited for – strategy, creativity, empathy, relationship-building, etc. Otherwise it’s just automation for automation’s sake and the promise of AI truly will burst like so many magical bubbles.

What to Do This Quarter

If you’re wondering where to start, here’s the playbook I use with clients:

1. Pick one workflow where volume and standardization are high (reporting, QA, asset adaptation).
2. Set a baseline: track hours, turnaround, and error rate for a couple weeks.
3. Pilot one tool with a narrow scope and a simple weekly scorecard.
4. Celebrate hours saved and reinvest them intentionally in strategy, creativity, or client engagement.

I’ve written more about this approach:

• What We Do in the Shadows
• It’s Time to Rethink Your Agency
• One Small Step

Beyond the Bubble: What Endures

Bubbles pop. Durable advantages don’t. If you measure minutes, not magic, AI becomes a lever – one that buys your team the time to think, create, and lead. That’s the transformation worth betting on.

Want to chat about your organization’s AI adoption journey? Get in touch.