I have new data to share. Some of it deserves a housewarming. Some of it deserves a call to the landlord.

In Cella’s quarterly Expert Perspectives survey, we asked creatives, marketers, UXers and ops experts across 17 industries how AI shows up in their work. Here’s what their responses tell us: AI made it into their workflows, but the operating model didn’t come with it. 

See what’s inside the latest Expert Perspectives Series report

Move-in day was a while ago 

Raise your hand if you’re among the 82% who are using AI daily or more. Most creatives, marketers, and UXers use it several times a day (author raises hand). What’s ground zero for their AI use? It’s a thinking tool first. Writing, ideation and research were at the top of the list, while design and video were at the bottom. Most agree that the payoff is speed; nearly three quarters named that, and nothing else came close. 

A high-maintenance roommate

AI works fast but never cleans up after itself. Quality control is the biggest new problem AI has created. One survey respondent called it a “free for all.”

Job loss made the list of concerns, but the bigger fear is the work getting worse. And creatives feel it most; they use AI less than everyone else and also worry about quality more than anyone else. 

Nobody’s the landlord

Two-factor falter: 43% of teams have no clear owner for AI. The same share does not measure impact at all, and only 1 in 5 uses real metrics. 

When nobody owns AI, nobody is on the hook to measure it. No landlord. Nobody reads the meter. And on the theme of lawlessness, more than a third of our survey respondents use tools nobody approved, with directors leading the pack (I see you!) 

Stop buying furniture

When we asked what would help most, workflow redesign tied for first (and this is consistent with what we hear from our clients). Tool selection came in dead last. We seem to have enough furniture now. 

Only 1 in 5 feels very prepared to scale AI responsibly, even though most say they have clearance to do so. Just because the rules are established doesn’t mean the team is ready. Most teams don’t have a tool problem, they just don’t know which problems to solve first. 

Nobody reassigned the rooms

When AI moved in, no one reassigned the rooms. It’s like AI started camping out on the living room couch and eating all the food, but nobody knows which room it gets or how it should contribute to the household. Translation: no one figured out how to integrate AI effectively into the everyday work and, more importantly, named what the tangible gain is for teams. They haven't redrawn the floor plan to show who owns what. The skills gaining ground are strategic thinking, AI fluency and judgment. Two out of three of those aren’t tools at all. 

Yet, 80% of teams have no new AI roles, and very few have plans for hiring them. Training is doing a lot of the heavy lifting to help teams evolve. Everyone else is just treading water.

The Philly walkthrough 

This gap helped inform the focus for our recent Signature Event in Philly: The team you’ll need next.  Creative, Marketing, UX and Ops leaders came together to work on closing it. My colleague Nickole Brown nailed the framing: companies are adopting AI faster than they’re redefining the workforce to run it. 

So we went to work on solving that. Groups built role profiles for a creative director, an experienced designer and an account manager. (They effectively wrote the survey responses all over again, but this time on sticky notes.) Here’s how it looked in practice.

  • On roles. Human skills beat tool skills in every profile. Creative directors defined the job as influence: "Empower the team, not just power the work.” Experience designers listed duties no job description has yet, like AI solution gatekeeper. One group even called the “belief you need to hire AI experience” a threat. The survey agrees. 

  • On ownership. The account manager group talked more than they wrote, and the conversation kept circling back to who gets to decide. Their board was mostly questions about clarity, RACIs, hybrid roles and one solid answer: align accountability and decision authority hand-in-hand. Silos showed up in every corner of their SWOT.  

  • On quality. “AI is so fast, we’re not even reading or evaluating the output.” And the line of the day was, “Speed to output is no longer the bottleneck.” Who defines what “good” is? That feels like a role question.

  • And on tools. “Too much focus on tools over process.” Also, “Train [to master one process] and another pops up; [it’s like playing] whack-a-mole.” Every big opportunity the room named was about the operating model. 

Your move-in checklist

AI already has a key to your place. What your team needs now is a floor plan, a landlord, a security system and a list of chores. Here’s what to start. 

  • Measure one or two things AI should change. Measure them now, while you still have a “before” story to tell. 

  • Define the problem. Finish this sentence: We’re using AI to solve ________________ and we’ll know it worked when _____________. If your team can’t agree, congratulations, that’s your first project. 

  • Name an owner. Back to that authority plus accountability. One without the other is just a box on a RACI chart.

  • Install the security system. Governance should work like a good alarm system: always on, barely noticeable, but providing security you absolutely need. Give the folks using unapproved tools a safe lane to pivot, and build your brand’s standards into the tools themselves. Arm the system by adding a quality check before anything ships.

  • Design the roles and train for them, too. Decide who owns what and what defines good. Build judgment in fluency within your team. Your best AI talent is already part of the solution, maybe using a tool IT hasn’t approved yet.

Conclusion

AI has the keys, a drawer full of takeout menus and strong opinions about the thermostat. It’s here to stay and not moving out anytime soon. The survey and Philly both said so. The tools settled in fast and most teams are still living out of boxes. So unpack, decide who gets which room, who defines house rules, and who gets the final say on what good looks like. 

Our experts can help you evaluate where AI is impacting your everyday work, from how to set up your approach with AI ownership to accountability, workforce preparation and governance that sets brand-safe boundaries.

Contact us for a consultation and insights customized to you.