Key Takeaways
- Skepticism is a discipline, not a personality trait—build it into the workflow.
- Undisclosed AI content doesn't just underperform—it costs you trust.
- Speed without judgment when implementing AI doesn't scale; it just moves the bottleneck.
A few weeks ago, Steve Wozniak joined us for an episode of our AI Executive Briefing series. He talked about fifty years of tech cycles, what separates real innovation from noise, and why he thinks the industry needs a lot more skepticism about AI than it currently has.
A few points from our conversation kept surfacing in the days that followed.
We stopped thinking of them as mystical stories from a garage in 1976 and started seeing them as instructions for how a 2026 marketing organization should build its AI strategy.
The discipline underneath what Woz was talking about translates directly to CMOs implementing AI today: deliberate innovation will always outlast momentary hype.
Here’s what leaders can take back to their own organizations and how we use the same principles as we build Validity Engage.
Build for a market of one before you build for a market of everyone
When Woz designed the Apple I and Apple II, there was no business plan, no competitor to out-feature, no committee weighing in. In his words, the market was one person—himself. He built something he personally needed, and everything else followed.
Most AI rollouts in marketing run in the opposite direction. A tool gets purchased because it can do a long list of things, and only afterward does anyone go looking for the problem it’s supposed to solve. That ordering shows up in our own research: nearly half of marketers we surveyed report only a limited or basic understanding of how AI is actually reshaping buyer discovery inside their own funnel. A lot of AI budget is being spent by teams who couldn’t precisely tell you which problem it’s supposed to fix.
Rather than starting a search for a shinier vendor, start from the single most acute pain point in your —the thing that costs your customer real hours every single week—and build or buy toward that. This is the same discipline behind how we prioritized the Validity Engage roadmap internally: one workflow at a time—pre-send QA, deliverability, data quality—instead of bolting on generic AI capabilities on top of existing tools because the market expects to see it.
Skepticism is a skill, not a weakness
Woz spent a good chunk of our conversation on Murphy’s Law, and specifically O’Toole’s Corollary—Woz claimed Murphy was an optimist.
His point wasn’t cynicism for its own sake. It was that we need far more skepticism about what AI produces than most organizations currently apply.
The mistake marketing leaders make is treating skepticism as a personality trait—something a sharp analyst brings to the table, rather than something built into the workflow itself. That distinction matters more than it sounds like it should, because the data underlying many AI projects isn’t reassuring. In our own research, 45 percent of marketers say their CRM data isn’t AI-ready, and the average respondent estimates that less than half of their CRM data is accurate and complete. More concerning: only about 55 percent of leaders say data quality risk actively informs their AI strategy—meaning close to half of leadership is either underestimating that risk or hasn’t thought about it at all.
Skepticism as a discipline means building in QA gates and provenance checks that don’t depend on any one person remembering to be careful.
Don’t choose convenience over trust
One of Woz’s hot takes was about AI regulation: he thinks every AI-generated claim should be traceable back to its source. He framed it the way you’d frame a bill of rights—rules exist to rule out the clearly bad, not to block the good. What frustrated him more than regulation itself was companies that say “we support responsible AI” without ever specifying what, concretely, they commit to.
Our own research speaks to a need for more transparency: 38 percent of consumers say they’re less likely to trust marketing emails they know were written by AI. 27 percent of consumers reported that they’ve turned off AI features in the inbox, a sign that consumer sentiment doesn’t fully line up with the hype or the “wild west” approach to regulation.
The practical takeaway isn’t a policy PDF nobody reads. It’s a specific standard that’s clear and visible to audiences—”we disclose X, we don’t do Y”—that lets you build trust directly instead of scrambling to retrofit compliance and beg for forgiveness later. It takes more time to add that kind of transparency into your process. It’s also the same logic Woz applied to regulation generally: ambiguity is what erodes trust, not the existence of rules. The relationship with your subscribers will be worth the additional effort and care.
The bottleneck doesn’t disappear—it moves to human judgment
Woz brought up AI making mistakes. It’s not perfect, and when you add all the correction time into your process, where does that leave you in terms of efficiency?
That’s a tough question for marketing leaders to sit with, because the same math applies to content production. Generating more marketing assets faster doesn’t generate positive ROI if review, brand-safety, and compliance checks can’t scale at the same rate—the constraint just relocates downstream, from creation to review.
The practical move is measuring against a pre-AI baseline—cycle time and engagement together, not efficiency in isolation—before declaring any AI workflow a win.
Building what lasts means choosing the slow right over the fast wrong
The story that stuck with us longest wasn’t about Apple at all. Woz described writing a chess program to solve the Knight’s Tour—a puzzle a computer would take 10^37 years to work through exhaustively. He solved it in one night using intuition instead of raw computation.
The marketing organizations winning with AI in 2026 aren’t surrendering everything to tech. They’re the ones applying human judgment at the specific points where it matters most—because, like the chess challenge example, tech and efficiency don’t always go hand in hand.
Building what lasts, in marketing or anywhere else, has never really been about how fast you can jump on a bandwagon. It’s about understanding which parts of the process are allowed to move fast, and which ones still need a human paying close attention.
If you want to build a program around that discipline—one where AI runs on data you can actually trust, with human judgment built into the workflow rather than slapped on after the fact—that’s exactly what we built Validity Engage to do. And if you’d like to hear the full conversation with Steve Wozniak, the Building What Lasts episode is available on demand, along with the rest of the AI Executive Briefing series.
Learn more about Validity Engage today.