Every conversation I have about AI and design starts the same way, with someone asking which tool to buy or which workflow to change. It rarely stays there for long.
The more interesting conversation is not about tools at all. It is about judgement, and whether an organisation has enough of it to make speed worth having.
I think that is the right question to be asking, and it is not the one most hiring managers are currently asking about AI.
Before you approve any AI tool for your design team, it is worth checking one thing first: how early design already gets pulled into decisions on a typical project. That single check tells you more about what AI will do for you than any feature comparison.
What AI speeds up in design work
To be fair to the tools, the speed is real. AI can compress hours of research synthesis, exploration and production into a fraction of the time they used to take.
A designer can move from a raw set of interview transcripts to a structured set of themes far faster than before. A team can generate a wider spread of prototype directions in a fraction of the time a manual process used to take.
None of that is an exaggeration of what the tools can now do.
But speed only creates value if the team was solving the right problem before the acceleration started. If the brief was wrong, or the research question was poorly framed, AI does not correct that.
It produces the wrong answer faster, with more polish, and with more apparent confidence than a slower, more manual process would have.
That is the turn in the conversation that matters for a hiring manager. Speed is not the same thing as progress, and it is worth being honest with yourself about which one you are buying when you invest in AI-accelerated design work.
The problems AI did not create
Here is what struck me most from these conversations. The challenges organisations tend to associate with AI were, almost without exception, problems that already existed before AI arrived.
Poor knowledge sharing between research, design and delivery was already a source of wasted effort. Weak handoffs between teams were already causing rework and confusion about who owned a decision.
Design getting brought into a project after the important calls were locked in was already limiting what design could change. Teams already had a habit of moving on to the next project without measuring whether the last thing they shipped worked.
AI did not create any of these problems. What it does is amplify whichever version of them already exists inside an organisation, because it removes the natural pauses in a slower process that used to force a team to stop and check its own work.
Two different outcomes from the same tools
This is why the same AI tools produce such different results in different organisations, and why it has little to do with how skilled the design team is at using them.
An organisation with strong foundations, meaning design is brought in early, handoffs are documented, knowledge moves between teams, and outcomes get measured after launch, gets a real multiplier from AI. Research moves faster, exploration widens, and the team ships more of the right things sooner.
An organisation without those foundations gets a different result from the same tools. It becomes significantly faster at producing the wrong things, with more confidence behind each wrong decision than a slower process would ever have allowed.
For a hiring manager, this means AI fluency in a design team or agency is not the first thing worth checking. It is a multiplier on whatever is already there, for better or worse, so the condition of the foundation matters more than the tool sitting on top of it.
Before you invest in AI tooling, or push an existing design team to move faster with it, take an honest look at a current or recent project against four questions:
- When was design brought into the project, relative to when the important decisions were locked in?
- Was the handoff between research and design, or between design and delivery, written down anywhere a new team member could follow it?
- Does the team have a working habit of sharing what it learns with product and engineering, or does that knowledge stay inside the design team?
- Once the last thing shipped, did anyone go back and measure whether it worked?
If the answer to more than one of those is no, fix that before you layer AI-driven speed on top of it. Otherwise, you are not buying a faster design function; you are buying a faster way to produce the same problems you already have.
Ready to talk about your next design hire? Get in touch.
Liam
Design, Creative & Research Recruitment
Last updated: September 2026
