Shaken, Not Stirred

There's a reason James Bond orders his martini shaken, not stirred.

It's not about flair. It's about intent.

Shaking breaks things apart. Stirring preserves what already exists. When you're trying to understand what something is really made of—and whether it still works—you don't gently mix. You disrupt.

That framing became unexpectedly useful when I started reexamining how our design team actually worked.

Not how we said we worked.
Not how our process diagrams claimed we worked.
But how decisions, communication, and execution truly moved through the system.

Shaken: Breaking the Workflow Apart

When we talk about "improving design workflow," the instinct is often additive: new tools, new rituals, new check-ins. That assumes the underlying structure is sound.

I wasn't convinced it was.

So instead of optimizing what we had, I chose to shake it.

I took our existing workflows—design-to-design collaboration, design-to-engineering handoffs, cross-functional alignment, async reviews, decision-making loops—and disassembled them into their smallest units:

• every touch point
• every handoff
• every moment where context moved—or decayed
• every place where effort repeated itself quietly

Once everything was laid out as individual nodes instead of a linear process, patterns emerged quickly.

Repetitive work masquerading as rigor.
Communication pipelines optimized for availability, not clarity.
Designers compensating manually for gaps the system refused to own.

This wasn't a tooling problem. It was a structural one.

Envisioning the Immediate Future

Rather than imagining a distant, idealized future, I focused on something much more constrained:

What does the immediate future of design work look like if AI is treated as infrastructure, not novelty?

Not "AI replaces designers."
Not "AI makes everything faster."

But:

• Where can cognitive load be removed safely?
• Where does context reliably break today?
• Which decisions deserve human judgment, and which ones don't?

That framing led to a set of concrete workflow inventions—not speculative concepts, but things we could actually test with real work.

Adding Ingredients: Agents, People, Reality

I didn't do this in isolation.

Alongside my direct design partners and managers, I ran team-wide and cross-functional interviews to pressure-test assumptions. I wanted to understand not just what was painful, but why people had normalized that pain.

From that research, I introduced and deployed four AI agents—purpose-built for our workflows—and shared them with the design team and close cross-functional partners.

The goal wasn't adoption.
It was exposure.

I wanted people to feel what it was like to work with systems that carried context, enforced constraints, and reduced repetitive effort—even imperfectly.

The reaction was… bittersweet.

People loved the agents. They used them. They asked for more.

But many also got stuck.

Not because the tools were hard to use—but because the ecosystem was already noisy. Too many AI tools. Too many overlapping promises. Too many choices for the same outcome.

More power, less clarity.

Not Stirred: The Discipline of Restraint

That's when I started giving a piece of advice that felt almost counterintuitive in an AI-saturated moment:

Don't stir.

Don't mix tools casually.
Don't stack AI systems for the same purpose.
Don't chase novelty mid-stream.

Instead:

• Define the outcome.
• Clarify the deliverable.
• Work backward to choose the one system that best supports that goal.

If two tools solve the same problem, pick one.
If a tool doesn't have a clear role, it doesn't belong in the workflow.

This wasn't about control. It was about coherence.

Systems break when responsibility is blurred. AI workflows are no different.

Closing the Loop

Advice alone doesn't hold.

So alongside this guidance, I established a feedback loop: tracking usage patterns, collecting friction points, and encouraging people to surface breakdowns as design input—not user error.

Overrides weren't failures.
Confusion wasn't resistance.
They were signals.

That loop allowed the agents—and the surrounding workflows—to evolve deliberately instead of chaotically. Not faster. Not flashier. Just more understandable.

What This Changed for Me

This work reinforced something I've come to believe deeply:

Senior design impact isn't about introducing more tools.
It's about deciding what not to mix.

Shaking a system reveals its truth.
Stirring too early hides it.

AI makes it tempting to keep blending—layering capabilities, chaining tools, adding complexity under the guise of power. But durable systems don't come from accumulation. They come from intent, boundaries, and restraint.

So when I think about how design teams should evolve next, I don't think about speed.

I think about clarity.

Shaken, not stirred.