A map out of the messy middle: 9 ways to make AI actually stick

Most teams aren’t asking whether they should use AI anymore. They’re asking something much harder:

“Why does this still feel chaotic?”

Between constant new tools, fragmented experimentation and rising pressure to keep up, many organizations are stuck in what Head of Brand at AirOps, Jessica Rosenberg, calls the “messy middle.”

In her session during Superside’s recent SHIFT summit, Rosenberg broke down exactly how to move from overwhelm to real adoption. Not with theory, but with practical shifts, repeatable frameworks and real examples from her team.

1. Start with the right conditions (not the tools)

Before AI workflows, prompts or tooling, Rosenberg emphasizes one critical foundation: environment.

What to do:

Audit your team’s “creative state” (chaotic vs calm)
Reduce unnecessary pressure where possible
Treat psychological safety as a strategic priority, not a perk

Fear, and stress are creativity's actual enemies. The environments we work in are not neutral. They either prime us for creativity or they literally shut it down.

Jessica RosenbergHead of Brand, AirOps

Why it matters:

According to Rosenberg, “when you build a team culture that's prioritizing psychological safety and calm… people get curious and curious people experiment.”

And experimentation is what actually drives AI adoption.

2. Create space for joy and play

AI adoption doesn’t thrive in high-pressure environments. It happens when people feel safe to explore. Rosenberg’s team operationalized this with a simple but powerful ritual called “Design Cooldowns”.

Every week or biweekly, her team gets together and works on a prompt. Each week, a different team member chooses a prompt. The goal is to provide team members with opportunities to get creativity flowing without the pressure of performance expectations.

The side perk? It’s actually fun and a great team-building activity.

In this Design Cooldown we had to draw a self-portrait or portraits of one another on the team in whatever medium we chose. And most of us actually chose pen to paper, which was refreshing given all the AI tools. But there are no goals, no KPIs, no feedback, just pure joy and play. And we laughed and my face and my cheeks hurt after I left the meeting.

Jessica RosenbergHead of Brand, AirOps

Rosenberg’s Design Cooldowns:

Weekly or biweekly creative sessions
Prompt-based (e.g. drawing teammates)
No performance expectations

Why it works:

Activities like the Rosenberg’s Design Cooldown sessions enable AI adoption and creativity by:

Lowering pressure
Building creative confidence
Encouraging experimentation with (and beyond) AI

3. Lower the stakes to unlock experimentation

One of the most practical shifts Rosenberg shared: don’t start with mission-critical work.

Lowering the stakes is so important to consider when you're thinking of how you can integrate AI into your team. Are there projects or opportunities in your day-to-day that are low stakes? Projects where it’s safe to experiment and safe to fail on the other end. An idea that won't break the team or the business or your KPIs or whatever if it goes wrong.

Jessica RosenbergHead of Brand, AirOps

Building an AI-powered sales game

Rosenberg’s first Claude Code experiment was built as a way to help her sales team remember the story about a product launch they did back in January. Instead of overhauling core workflows, she built:

A retro-style educational game for sales enablement
“Oregon Trail”-inspired experience
Links to product content and training materials

This was low stakes. And people loved it. I was like, okay, Claude Code is amazing. I shipped something that people like, and they're finding value in it. And it was also really fun to make?

Jessica RosenbergHead of Brand, AirOps

Result:

Immediate team engagement
Proof of value
Momentum for bigger idea

That kind of opened up the floodgates on the team and instead of asking ourselves, how do we use this? It started to become what else can we build? So we moved beyond templates into actual tools, starting with something called a brand scale.

Jessica RosenbergHead of Brand, AirOps

4. Solve real friction (not hypothetical use cases)

Instead of asking “How should we use AI?”, Rosenberg focuses on:

“Where in my day-to-day is the work that I'm doing, manual, repetitive, time consuming and just flat out annoying?”

According to Superside’s Breakpoint report, 4 in 5 creative professionals say they want to create bolder work, but they’re always racing against the clock, and 77% say they feel bogged down with too many lower-priority tasks. Finding ways to use AI to remove those obstacles and make way for more creative work has never been more important.

What to do:

Identify repetitive creative work (e.g. resizing ads, decks, one-pagers)
Prioritize automating or templating those first
Free up time for higher-value creative thinking

Mindset shift:

Rosenberg put emphasis on using creative brain power on the work that actually matters and leveraging AI to do everything else. “If you're a creative or a designer and you're doing decks and one sheeters and resizing ads all day, when you can really be using your beautiful brain to design campaigns and think of, apps and immersive experiences, like the creative juices that we wanted to get into this career for, that's what we should be doing in our day to day.”

5. Build systems, not one-off outputs

One of the most powerful examples Rosenberg shared is the idea of “skills”—structured AI guardrails.

A skill is essentially like a style guide for machines. I like to call it a contract between the human and the AI agent. And the more detailed that contract is, the better the output will be via that AI agent or machine.

Jessica RosenbergHead of Brand, AirOps

A brand skill turns your static brand guidelines into something AI can actually use. Instead of sitting in a PDF, your rules for color, typography and tone become embedded directly into the outputs your team generates. That means anyone, not just designers, can create work that stays consistently on brand.

Benefits of building a Brand Skill (AI Style Guide)

Encodes brand rules (colors, typography, usage)
Installed across the organization
Applied automatically in AI-generated outputs

Impact:

Non-designers can create on-brand assets
Designers scale their influence
Consistency improves across teams

Maintaining brand consistency with AI isn’t easy, especially as more teams start experimenting. This is why solutions like Superside’s Brand Brain are emerging, giving teams a way to remember past projects and embed brand guidelines directly into their AI workflows.

6. Turn individual wins into team-wide adoption

Once early experiments work, scale them intentionally.

Rosenberg did this through collaborative upskilling:

She incentivized adoption by creating a series of one-pagers on how to set up Claude Code and then hosted a vibe marketing hackathon. Team members also used the Brand Skill keeping everything on brand.

We gave the team two weeks to build certain tools and things within certain parameters and we were blown away. Seeing what everyone created was honestly one of the coolest moments in my entire career.

Jessica RosenbergHead of Brand, AirOps

Tactics that worked:

Internal one-pagers and onboarding guides
Company-wide hackathons
Weekly vibe coding office hours
Dedicated Slack channels

It really has become like a company-wide cultural movement that started from a few experiments.

Jessica RosenbergHead of Brand, AirOps

7. Use AI to eliminate work (then reinvest the time)

AI is all about making space for work that matters. For Rosenberg, this means building tools that remove recurring work entirely and save time.

“We do a lot of webinars at AirOps. There's four a month at least and they need artwork and assets for every single one. And we do not want to spend our time making these assets over and over again,” Rosenberg explained.

For their hackathon they built a templating tool that allows marketing team members to now create webinar assets. Team members can create a whole slew of assets and have different color options or patterns and still keep it perfectly on brand.

That has removed an entire workload of work that we do not want to be doing as human creatives off of our plate.

Jessica RosenbergHead of Brand, AirOps

Using this AI templating tool freed-up time that was then reinvested in:

Campaign strategy
Experiential work
High-impact creative

8. Make learning social (not solo)

AI can feel overwhelming—but shared exploration changes that. Looking back at her AI beginnings Rosenberg reflected on how her newfound mission changed her perspective. “I literally thought, this is going to change everything. And I am so pumped. And I knew a lot of people didn't feel the same way as me. They felt scared and anxious and confused,” she explained.

Her takeaway? Find ways to share her excitement so others could also experience the benefits of AI.

I promised myself at that moment that I would make it my personal mission to help everyone feel as excited as me by building in public, sharing information about the teams that I'm leading, helping them upskill and hopefully get excited too.

Jessica RosenbergHead of Brand, AirOps

Rosenberg socialized AI by implementing:

Weekly AI sharing sessions
Show-and-tell demos
Open discussions about confusion and fear

Finding an hour together every week to just talk about experiments, really helps because you're not alone.

Jessica RosenbergHead of Brand, AirOps

9. Embrace the beginner mindset

Even experienced creatives feel behind.

“I still feel like imposter syndrome around it too… I'm a beginner again”

But Rosenberg reframes that as a strength:

“I also love feeling like a beginner again.”

Normalize not knowing
Reward curiosity over expertise
Build momentum through learning, not perfection

The big shift: From Adoption to Integration

Ultimately, Rosenberg’s framework isn’t about tools. It’s a behavioral change.

Her process in practice: Defining the conditions, creating a culture of joy and play, learning with low stakes, sparking team-wide exploration and getting people together to build.