Close to a third of organizations say executives and day-to-day teams are misaligned on AI strategy, per Adobe's 2026 AI and Digital Trends report, conducted with Oxford Economics.
Another 47% say alignment exists, but it’s partial at best. Executive misunderstanding of AI ranks as the top driver of that gap, ahead of resistance to change.
At the same time, 42% of organizations already plan to design distinct AI agent personalities for different audiences.
Brands are racing toward AI-driven experiences faster than their own leadership and teams agree on what they're building.
AI Infrastructure Lags Agentic Ambition
The 42% figure looks bold until it runs into a second number from the same report. Moreover, 72% of organizations have tools to connect systems for generative AI use cases.
But only 37% have that same setup for agentic AI, the layer that actually drives a distinct agent personality.
When a personality transitions from a chat window to a voice, mobile app, or call center script, it can completely collapse.
Even before a tone meeting is scheduled, there is a disconnect between what a brand wants and what its systems can provide.
Alignment Workshops Cut Rebuild Costs
Misalignment between leadership and a build team rarely announces itself early.
It usually appears in the middle of the project, when the budget has been used up, and the deadline has passed. It is more expensive to fix it later than to get it right from the beginning.
Doug Hughmanick, founder and head of creative at ANML, starts every project with a workshop for leadership and the build team.
ANML is a design and development agency that creates digital experiences across websites, marketing, products, and AI, bringing together strategy, design, brand, content, and engineering.
"We get leadership and the build team in the same room early and work through a few concrete questions together, what the experience should do, who owns which call, what 'done' means," he says.
"Settling that upfront is what prevents the expensive rework later."
Executives Miss the AI Mechanics
Adobe's finding that executive misunderstanding drives most AI misalignment matches what Hughmanick sees on ANML's own projects.
"The blocker is rarely that teams don't want to move. It's that leadership and the people building it are picturing two different things," Hughmanick says.
Executives tend to grasp the ambition behind an AI initiative without grasping what it actually takes to build it.
So "add AI" means something different at the top than it does to the team shipping it. And that gap is the disconnect the agency runs into most often.
AI Personality Is an Engineering Call
Designing how an AI agent sounds seems like it should be a branding task. But Hughmanick treats it as a product and engineering decision too.
How an assistant sounds when it fails, how fast it responds, and what it refuses to do carry as much engineering weight as brand voice.
ANML encodes those choices as design tokens and machine-readable components. They flow directly from Figma into the build.
That keeps the system compatible with AI tools that need to work against it directly.
"The brand travels into the product as live rules instead of a static guideline someone reinterprets later," Hughmanick says.
"That's how the personality survives contact with what's technically possible."
Launch Timing Outpaces Team Agreement
Brands already designing distinct AI agent personalities are moving faster than most teams have resolved what those personalities should really do.
That sequence creates a specific risk. A brand announces an AI assistant with a certain tone, capabilities, and limits, before the team building it has agreed on any of that internally.
Customers then meet a product that acts differently than promised. An assistant might refuse tasks it was marketed to handle. It might behave inconsistently across channels.
"The gap shows up after launch as an experience that doesn't match the pitch, and walking that back costs far more than the hard conversation would have beforehand," Hughmanick says.
For that reason, ANML settles what the experience actually does before the client announces anything, not after the promise is already public.
Cross-Discipline Teams Cut Rebuilds
A feasibility problem caught in the first workshop is a design constraint. The same problem caught after strategy and design have already locked in becomes a rebuild.
"With strategists, designers, and engineers in the room from the start, the plan is grounded in what can actually ship," Hughmanick says.
Keeping all three disciplines on a project end-to-end is what changes what a team catches, and when.
