As artificial intelligence rapidly reshapes the landscape of digital product design, it presents the industry with a profound dual reality. On one hand, AI tools could grant designers unprecedented autonomy, stripping away bureaucratic hurdles and enabling them to move directly from conception to execution. On the other hand, this same autonomy risks exposing deep structural gaps that traditional workflows and organizational constraints have long helped hide.
Examining both the optimistic bull case and the sobering bear case, digital design expert Andy Budd explores what happens when modern product designers finally need less permission to act.
For years, digital designers have operated under a shared professional frustration: the belief that they could produce significantly better work if only the surrounding organization would get out of the way. While rarely voiced in such blunt terms, this sentiment usually manifests as a litany of operational bottlenecks. Designers frequently report not getting enough engineering time, finding that product teams have already decided on a solution before design even begins, or struggling against overly packed roadmaps where leadership cares exclusively about quarterly financial targets.
Further compounding these challenges are research cuts, poorly run experiments, and known design debt that no one ever wants to allocate a sprint to fixing.
Much of this friction is rooted in objective reality. Most designers have spent their careers navigating the awkward middle space situated squarely between product management and engineering. Product frames the problem—or at least assumes it does—while engineering dictates what is technically feasible or financially affordable. Meanwhile, design is expected to make the resulting artifact clearer, simpler, more coherent, and more usable, all while taking care not to disrupt the established plan too much.
This precarious position has always made the profession uncomfortable. Designers are consistently told to think strategically, yet they routinely lack the organizational power required to act strategically.
A designer can easily spot a broken onboarding flow, a confusing upgrade path, an empty state that makes users feel inadequate, or a feature that looks entirely reasonable in a sterile product review but makes no real sense in actual use. However, seeing a problem is fundamentally different from getting it fixed.
Consequently, design work frequently devolves into an exhausting process of internal advocacy and argumentation. Designers must meticulously make their case, annotate complex user flows, bring forward compelling research clips, point directly to customer support tickets, and present polished Figma prototypes. They must painstakingly explain why what others dismiss as a small edge case actually represents the first-run experience for half of all new users. Yet, the typical outcome involves everyone nodding in agreement, acknowledging that the point matters, and then immediately moving on to whatever initiatives were already locked into the roadmap.
This dynamic explains why the ongoing rise of artificial intelligence is far more interesting for the discipline of design than the tired, repetitive debate over whether AI will eventually replace designers altogether. The true structural shift is not simply that AI allows individual designers to manufacture more screens—after all, virtually no company actually needs more screens. The genuinely transformative change is that designers may soon require significantly less permission to implement their ideas.
The Bull Case: Designers Need Less Permission
Under the emerging bull case for AI in design, a skilled practitioner can now move rapidly from observing that something is broken to directly fixing it and pushing the solution live. A designer can independently prototype an alternative onboarding flow, write and test clearer product copy, build a rough working version of an interaction, clean up lingering design debt without waiting months for a roadmap slot, and make a better alternative visible enough that it becomes remarkably difficult to ignore.
This capability fundamentally alters the internal politics of digital product development. Historically, design has relied heavily on persuasion because practitioners lacked direct means of production. AI weakens this dependency, even if it does not eliminate it universally. Complex software products will always rely on foundational architecture, underlying data models, security protocols, compliance measures, legacy systems, and myriad other unglamorous realities that make software engineering genuinely difficult.
Even so, the operational boundary is shifting. A growing portion of the vast gap between merely having an idea and making that idea a reality can now be bridged independently by a motivated designer equipped with the right technological tools.
In this prospective future, designers evolve into figures who are much less permission-dependent. They are less reliant on product managers to bless a problem statement, less dependent on engineering teams to code every minor improvement, and far less trapped in the traditional role of internal critic, taste-provider, or mere Figma operator. Instead, they become fully empowered to make, test, repair, and ship solutions.
The strongest designers in this ecosystem will begin to resemble hybrid product leaders rather than traditional production designers. While they will still care deeply about interaction design, visual hierarchy, language, user flow, brand integrity, and craft, they will also grasp the broader commercial shape of the problem. They will be capable of making informed trade-offs, prototyping in code—or at least close enough to code—and leveraging AI to explore vast arrays of options quickly before using their own professional judgment to discard the majority of them. They will be able to sit directly with a founder or product manager and move from a vague product concern to a tangible, working concept by the end of the day.
While there may ultimately be fewer designers of this caliber required in the workforce, those who remain will wield significantly more direct influence over the products they build. The traditional design organization model was built, in large part, around scarcity—scarce engineering time, slow production cycles, expensive prototypes, handoffs between rigid specialists, and heavy coordination overhead across disparate teams.
If artificial intelligence reduces some of that underlying scarcity, it will inevitably reduce the need for many of the supporting roles that grew up around it. The optimistic outlook is not that every single designer keeps their job while simply receiving an arbitrary productivity boost, which borders on wishful thinking. A far more believable trajectory is that the total number of design professionals contracts, but the individuals who remain retain much more direct influence over product outcomes.
For the strongest practitioners, this is far from a negative outcome. In many ways, it represents precisely the level of empowerment that many designers have spent years demanding.
The Bear Case: Autonomy Exposes The Gaps
However, true autonomy has its drawbacks. If AI affords designers significantly more room to act, it simultaneously strips away their traditional protective cover. The very organizational constraints that held good designers back have historically also protected weaker practitioners from being tested too directly.
For years, it has been easy for underperforming designers to claim they had a superior idea, but simply never received the necessary engineering time to build it. Sometimes, that account was entirely accurate. On other occasions, however, the supposedly superior idea was never much more than a superficial critique. It had never been made concrete, rigorously tested, or forced to reconcile with awkward real-world trade-offs. It sounded compelling precisely because it lived safely in opposition to whatever product had already been shipped.
A vast number of designers excel at noticing what is wrong with an existing experience. Significantly fewer possess the skill required to decide what should happen instead. Fewer still can translate that alternative vision into something real enough for other stakeholders to evaluate objectively. The widespread adoption of artificial intelligence will inevitably expose this capability gap.
If a designer can easily prototype a recommendation, that recommendation must inherently improve. If they can build an alternative user flow, that flow must survive direct contact with operational details. If they can independently test product copy, they must actually care what happens when users read it. If they can clean up a small piece of design debt, they must justify whether the effort was genuinely worth making.
Some designers are not nearly as strategic as they often claim to be. They have successfully memorized the specialized language of strategy without ever enduring the discomfort of owning actual business outcomes. They can talk fluently about user needs, high-level business goals, systems thinking, and product quality, but they freeze or struggle when explicitly asked to make a definitive call. They desperately want institutional influence, but they reject the professional exposure that must accompany it.
The design profession has spent a long time arguing that it deserves a greater seat of power. If that power arrives, it brings fewer excuses. The work will be judged less by the rhetorical elegance of an argument and more by the tangible quality of what was actually made, tested, or changed. While that represents a healthier and more rigorous standard, it will prove punishing for many in the field.
A second major concern centers on the broader balance of institutional power. In most modern corporations, product management and engineering already hold significantly more structural power than design. They control product roadmaps, technical architecture, sprint machinery, performance metrics, and the specific vocabulary that executive leadership understands. Design has historically been forced to translate its concerns into someone else’s terminology before they are taken seriously.
Rather than rebalancing that dynamic, AI may simply hand product and engineering enough native design capability to make the design function easy to bypass altogether. A product manager who can generate a decent user flow, acceptable copy, and a plausible prototype using AI may feel little incentive to involve a dedicated designer early in the process. An engineer who can use AI tools to spin up a reasonable interface may decide that an existing design system covers enough of the decision-making process. A founder who can pull together a polished product demo in a single afternoon might easily mistake superficial polish for genuine product thinking.
The danger is not that these non-designers will suddenly become elite designers. The real risk is that many corporations fundamentally cannot tell the difference between great design and plausible design. Plausible design is dangerous because it looks coherent during a product review. It utilizes the correct UI components, the spacing looks fine, the copy is unobjectionable, and the overall flow mostly works. Because no one in the meeting feels strongly enough to object, the work simply ships.
A vast number of mediocre product decisions already survive because they manage to look plausible. Artificial intelligence will inevitably produce more of them. This is precisely where the design profession risks losing ground rapidly—not because core concepts like taste, judgment, research, and interaction thinking stop mattering, but because the visible outputs of design become trivial for other corporate functions to imitate.
If a company already operates under the assumption that design is primarily about screens, prototypes, and surface polish, AI simply provides a cheaper and faster way to acquire those deliverables. In such an environment, design does not gain true agency; instead, its scope is narrowed. The remaining designers are relegated to managing the design system, policing component usage, reviewing flows that have already been finalized, tidying up interfaces, maintaining brand consistency, and getting pulled in exclusively for high-stakes launches, executive demos, or messy cross-platform emergencies. While this remains useful work, it represents a drastically smaller surface area—shifting the focus away from shaping products and toward maintaining organizational furniture.
This reality highlights why the comforting refrain that "AI will simply automate the boring 20 percent of the job" may prove dangerously naive. In certain environments, that may well be the case. But within large technology organizations where design teams expanded primarily around coordination, production, and process management, the potential contraction could be far deeper. In places where leadership never truly understood why the design team grew so large in the first place, the reduction could easily approach half the workforce or more.
Where the Industry Might Land
The most uncomfortable aspect of this transition is that both futures can easily coexist. Artificial intelligence can simultaneously make the absolute best designers far more capable while rendering many average designers much less necessary. It can afford the discipline greater strategic agency while systematically reducing overall headcount. It can successfully propel a select group of talented designers closer to true product leadership while pushing others entirely into governance and maintenance roles.
Ultimately, AI will free designers from waiting around for institutional permission, only to reveal that some practitioners were far more comfortable waiting than they ever were taking action.
The designers who thrive in this environment will not be those who merely utilize AI to churn out larger quantities of design options, as options are effectively free now. Instead, success will belong to those who possess the discernment to know which option is actually worth pursuing, understand why it matters, know how to rigorously test it, recognize what needs to be cut, identify where the product is lying to itself, and realize when a "good enough" solution is quietly damaging the underlying business.
These professionals will require refined taste, but taste alone will not suffice. They will need genuine product judgment, technical curiosity, commercial awareness, and the professional nerve to make decisive calls before every variable is neatly settled. They will need to navigate fluidly between a customer conversation, a technical prototype, a pricing strategy, a brand question, and a messy implementation detail without insisting that all of those responsibilities belong to someone else.
Whether the industry ultimately moves closer to the liberating promise of the bull case—featuring fewer permission structures, greater agency, and elite designers finally free to demonstrate their value—or slides into the defensive posture of the bear case remains an open question. In all likelihood, the future will manifest as an uncomfortable hybrid of the two.
Some practitioners will leverage AI to capture unprecedented agency, while some corporations will use the technology to justify employing fewer designers overall. Some teams will produce exceptional work because the distance between critical judgment and execution has shrunk, while others will routinely ship plausible mediocrity because no one in the room possesses the discernment to tell the difference.
For years, digital designers have argued that they could generate transformative value if only they were less constrained by the organizational machinery surrounding them. Artificial intelligence is about to put that central claim to the test. Some designers will finally get the chance to prove it, while others will discover, too late, that their old constraints were doing them a favor.