Adobe spent its annual MAX conference showing off tools that can generate images, edit video, build prototypes and run multi-step creative tasks from a sentence. Then two of its creative leaders spent a fireside Q&A explaining why none of that replaces the person deciding what is good.

“Speed is a capability, it’s not a creative strategy,” Resh Sidhu, Adobe’s vice president of creative, said during the conversation in Miami Beach. “You can create a thousand options of something – you still need someone to decide that’s the thing that’s worth making.”

That line is easy to nod along to and harder to act on. Adobe is selling speed. Its customers are under pressure to produce more. And the people doing the work are caught between the two.

What Adobe actually announced

This year’s MAX was dominated by agentic AI. Adobe expanded Firefly into what it calls an AI creative studio, added a Creative Agent that can plan and carry out tasks, and showed new video, design and image tools. It also previewed local MCP support, which would let outside agents, including ones from OpenAI, Google and Anthropic, work directly in Photoshop, Illustrator, Premiere and After Effects files.

In the keynote, Adobe framed the shift around choice. Creators can use the agent, ignore it, or step in and edit the result by hand. “You are the director,” the company said. The pitch is that AI handles production chores while the human keeps control of taste, style and direction.

That is a reasonable pitch. It is also a difficult one to prove, because the same tools that shorten a task can raise the number of tasks a team is expected to finish.

The AI paradox at Adobe MAX

Adobe’s executives were not arguing that AI is bad. They were arguing that AI is incomplete. Deepa Subramaniam, Adobe’s vice president of product marketing for creative professionals, described a Photoshop expert who once needed a specialist to animate a project in After Effects. With AI, that person can describe the motion they want and get a usable result without spending months learning a second application.

That is a real expansion of creative range. It means a writer can sketch a visual idea, a photographer can produce a short motion piece, and a small team can attempt work that used to require a larger one. Adobe’s case is that more people contributing ideas leads to more good ideas, not just more clutter.

But it also changes the job. If the production step gets cheaper, the value moves to the brief, the concept, the critique and the final call. Those are the parts that decide whether the output is any good. They are also the parts that are hardest to measure and easiest to rush when a client wants another ten versions by Friday.

Why more content is not automatically better

The strongest argument for AI in creative work is not that it makes one piece of work better. It is that it makes exploration cheaper. A designer can test five directions before lunch. A marketer can see a campaign in ten formats. A filmmaker can storyboard an idea without booking a studio.

Exploration is valuable. So is variety. The problem starts when volume becomes the goal. A thousand options do not help if nobody has the time or authority to choose. They create what teams might call review debt: a pile of plausible work that still needs a decision.

Adobe’s own framing acknowledges this. The company says creators remain in the director’s chair. But a director with no time to review the footage is not really directing. The tool can generate the options; the organisation still has to create the conditions for someone to choose well.

The bottleneck is moving to judgment

For the past few years, the constraint in creative work was often production capacity. How many assets can the team make? How fast can the video be cut? How many versions can the agency deliver? AI attacks that constraint directly.

Once production is cheap, the constraint moves. It becomes attention, taste, brand knowledge and the ability to say no. A model can imitate a style. It cannot tell you whether the style fits the strategy, whether the joke will land with this audience, or whether the campaign should exist at all.

Sidhu’s point is that speed is a capability, not a strategy. The same is true of volume. A team that produces twice as much work in half the time has not automatically become twice as creative. It has simply changed the ratio of output to decisions. If the decisions do not improve, the extra output is noise.

What this means for creative teams

The practical lesson is not to ban AI or to worship it. It is to decide where the saved time goes. A few rules help:

  • Protect the idea stage. Use AI to clear production work, then spend the recovered hours on concepts, critique and revision. If the calendar fills with more assignments instead, the tool has not bought you anything.
  • Measure quality, not just volume. Track approval rates, revision rounds and performance, not only the number of assets shipped.
  • Keep a human owner for every deliverable. Someone must be accountable for the final choice, even if a model made the first draft.
  • Write better briefs. The clearer the direction, the less time the team wastes generating options that were never viable.
  • Build review capacity. If ten people can generate work but one person approves it, the approval step becomes the bottleneck. Train more editors, not just more prompters.
  • Use AI for adjacent skills, not as a substitute for judgment. A writer using AI to visualise an idea is expanding range. A team using AI to avoid thinking about the audience is just automating the wrong part.

The part Adobe can’t automate

Adobe’s business depends on creative professionals continuing to care about craft. Its tools are premium because they offer control, precision and a professional workflow. If AI made taste irrelevant, Adobe would be selling a commodity. So the company has a commercial reason to say that humans still matter.

That does not make the claim wrong. It does mean readers should treat it as both a philosophy and a positioning statement. Adobe is arguing that its AI is a tool for professionals, not a replacement for them. The subtext is that the professionals should keep paying for the tools.

The more useful question is what “human judgment” actually looks like in daily work. It is not a mystical quality. It is knowing the audience, understanding the brand, recognising a bad idea early, giving a clear note, and taking responsibility for the result. Those skills can be taught, practised and measured. They are also the skills AI makes more visible, because when everyone can produce a polished draft, the difference is in what gets approved.

What to watch

Three things will show whether Adobe’s message matches its products. First, whether the agentic tools actually reduce busywork or simply raise the expected output per person. Second, whether Adobe prices AI in a way that lets small teams use it without blowing their budgets. Third, whether the company builds guardrails that make provenance and credit clear, because trust is part of quality too.

For creatives, the safest bet is to become the person who can direct the tools, not the person who competes with them on speed. The ability to generate a thousand options will be common. The ability to choose the one worth making will not.

The bottom line

Adobe’s executives said the quiet part out loud: AI can make creative work faster, but speed is not a strategy. The companies that win with these tools will be the ones that use the saved time to think more clearly, not just to ship more. The tools can produce the options. A person still has to decide what deserves to exist.

Have a tip? Email [email protected] or use the contact form and choose Editorial. We do not promise anonymity.

Sources: TechRadar; Adobe newsroom; Adobe MAX speakers; Fstoppers; Gadgets Now.