I get asked sometimes why a brand strategy researcher is borrowing a framework built for corporate strategists studying manufacturing firms and technology companies, and the honest answer is that I didn't find anything closer to home that explained what I was actually seeing across case after case: some brands catch drift early and recover cheaply, and some brands let it run for eighteen months and pay for it publicly, and the difference isn't luck. It's a specific, describable capability some organizations have built and others haven't.
What Teece was actually trying to explain
David Teece's Dynamic Capabilities Theory, developed through the 1990s and refined in his later work, was built to answer a question that has nothing to do with branding on its surface: why do some firms consistently adapt to a changing environment while others with equally good resources fail to. His answer splits organizational capability into three distinct capacities. Sensing is the ability to notice a shift in the environment before it becomes an obvious crisis — spotting a weak signal early, while it's still cheap to respond to. Seizing is the ability to actually act on what's been sensed — mobilizing resources, making a real decision, rather than noticing a problem and letting it sit in a meeting note nobody follows up on. Transforming, sometimes called reconfiguring, is the ability to restructure the organization itself so the same problem doesn't simply recur the next time conditions shift — building the fix into the structure, not just patching the specific incident.
I think this maps onto Brand Drift with an almost uncomfortable precision, because drift is fundamentally a story about which of these three capacities was missing at each stage of every case I've studied.
Where CNET's response actually broke down
CNET's compound-interest-error scandal is a useful test case precisely because the company's response, once the problem became public, wasn't nothing — it's genuinely instructive to look at what it did and didn't include. CNET paused its AI-content program and ran a full internal audit once Futurism's reporting made the errors public. That's a seizing response, and a reasonably fast one once the problem was externally sensed. But notice what's missing from that sequence: the sensing itself never happened internally. It took an outside outlet, not an internal review process, to notice that forty-one of seventy-seven articles needed correction. CNET seized well once someone else sensed for them. That's a meaningfully weaker capability than sensing your own drift before an outsider does it for you, and it shows up in the sequence of events: months of unreviewed output before the correction cycle, rather than a caught error in week one.
The transforming stage is where I think CNET's response gets genuinely interesting to study, because the company didn't simply abandon AI-assisted content after the scandal — it said it intended to keep using the tool, with, according to its own statements, a tighter review process attached going forward. Whether that reconfiguration actually holds is an empirical question I don't have full visibility into from the outside, but the intent, at least, points toward transforming rather than just seizing once and hoping the same gap doesn't reopen.
Where Levi's shows a different pattern entirely
Levi's 2023 AI-model backlash shows almost the opposite sequencing, and it's worth holding the two cases side by side. The sensing here wasn't internal either — the tension between "AI-generated diversity" and actual representation was spotted by the public, immediately, at scale, rather than by an internal review before the announcement went out. But the seizing was dramatically faster: a revised statement within about a week, clarifying the initiative was never meant to replace real photoshoots or substitute for the company's actual diversity commitments. That's rapid seizing once sensing happened, externally, at a speed CNET's slower-burning scandal never had to contend with, because a viral social backlash simply moves faster than a financial journalist's slow-building investigation.
What I don't have public evidence of, in Levi's case, is the transforming stage — whether the company built a new internal check for future initiatives that touch identity-sensitive language, or whether the fix was purely reactive, addressing this one announcement without changing the underlying process that let the mismatch ship in the first place. That's precisely the gap Dynamic Capabilities Theory would predict matters most for whether the same kind of misstep recurs the next time an efficiency initiative brushes up against a stated value.
Why sensing is the capacity nobody builds on purpose
Here's what I think is the most useful, and most uncomfortable, implication of applying this framework to Brand Drift specifically: almost every organization I've studied has some version of seizing capability, because a public problem forces a response one way or another, and almost none of them have built genuine internal sensing capability, because sensing has no external trigger forcing it into existence. Nobody's stock drops because a sensing gap exists quietly. It only drops once the gap produces a visible incident, which is precisely why sensing is the capacity organizations underinvest in relative to how much it actually matters — it's invisible until it's missing, and by the time it's visibly missing, you're already relying on seizing and transforming to clean up after the fact.
This is, not coincidentally, exactly what my own BIG Framework's Intelligent Human Oversight pillar is trying to operationalize into something an organization can actually build on purpose rather than discovering the hard way that they never had it: a scheduled, deliberate sensing mechanism — someone reading a representative sample of output against a specific standard — rather than waiting for an outside journalist, a viral backlash, or a tribunal to do the sensing on the organization's behalf, for free, in public, at maximum cost, and on a timeline nobody inside the organization ever got to choose for themselves.
Why I don't think this is just relabeling common sense
I want to address the objection I'd raise against myself here, because I think it's the honest one: isn't "notice problems early, act on them, and fix the underlying process" just common sense dressed up in academic language? I'd push back on that, gently, because common sense is exactly what's been failing across every case in this piece. CNET, Levi's, and the dealerships and delivery firms I've written about elsewhere are not staffed by careless people who lack common sense. They're staffed by competent professionals operating inside organizational structures that never explicitly assigned sensing to anyone, which is a subtly different problem than a lack of common sense — it's a lack of structure for common sense to actually operate inside of, on a schedule, with a name attached. Dynamic Capabilities Theory earns its keep precisely by making that structural gap visible and specific, rather than leaving it as a vague, unassignable virtue everyone assumes someone else is handling.
What building sensing capability actually costs
I think part of why sensing gets skipped isn't laziness — it's that building seizing capability feels urgent in a way sensing never does, because seizing responds to a fire that's already visibly burning, and organizations are structurally biased toward funding visible fires over invisible ones. Building a genuine sensing capability costs remarkably little compared to what it prevents: a named person, a recurring calendar slot, and a specific standard to check against, which is a fraction of the cost of the correction cycle CNET eventually ran, or the reputational cost Levi's absorbed in a week of public criticism. The businesses that treat sensing as a real line item, not an afterthought squeezed in when someone remembers, are making a genuinely good trade, even though the payoff is invisible right up until the moment it quietly prevents something from ever becoming a story at all.
The capability a brand actually needs, stated plainly
If Dynamic Capabilities Theory is right that these three capacities are genuinely distinct and independently buildable, then the practical takeaway isn't "get better at handling scandals when they happen," which is a seizing-only strategy dressed up as resilience. It's "build the boring, unglamorous sensing habit deliberately," because sensing is the only one of the three capacities that actually prevents the scandal from reaching the public in the first place, rather than simply managing it better once it has.
Curious whether your own organization has a real sensing capability for brand drift, or just a seizing reflex you've never had to test? Get Your Free Brand Audit →
References
Duffy, C. (2023, January 25). Plagued with errors: A news outlet's decision to write stories with AI backfires. CNN Business.
Rangefinder. (2023, April 6). After backlash, Levi's says AI-generated models "to increase diversity" won't replace real shoots.
Teece, D. J. (2007). Explicating dynamic capabilities: The nature and microfoundations of (sustainable) enterprise performance. Strategic Management Journal, 28(13), 1319–1350.
Teece, D. J., Pisano, G., & Shuen, A. (1997). Dynamic capabilities and strategic management. Strategic Management Journal, 18(7), 509–533.