LAS VEGAS — Artificial intelligence provides speed and efficiency but lacks taste and aesthetic judgment, making human oversight essential to prevent automation that eliminates human input entirely.
“What we’re missing is this notion of taste and judgment,” Paul Woolmington, CEO of Canvas Worldwide, told Beet.TV contributor David Kaplan at CES. “Taste is very aesthetic and I think it’s going to be a long time before AI will truly nail taste. So that is an aspect of human judgment that I think is important and sometimes lost.”
While Canvas embraces AI for expanded options and operational improvements, believing technology can automate everything without human involvement represents a destructive approach.
Hero, hygiene, hub framework
Canvas addresses AI complexity through three operational buckets: hero applications that solve major business problems, hygiene processes that automate existing workflows, and hub partnerships with technology platforms.
“Hero applications are really solving big business problems,” Woolmington said. “Hygiene is this race to cheaper, faster, more efficient — how can we automate processes? A lot of it is hidden. A lot of it is stuff that we’ve been doing for years in media.”
Strategic AI partnerships include operating system integrations like Ella, a confederated platform of frontier LLMs designed to marry robust intelligence while safeguarding client IP and enterprise data privacy, Woolmington added.
“We also have ongoing execution partnerships with scaled players such as Google, Amazon, and Meta but we are also increasingly developing innovative programs with nimble growing partners, such as Chalice and Gigi,” he said.
Training extends beyond internal teams
Canvas creates supportive, trusted environments before introducing AI rather than layering technology onto unstable foundations, particularly when holding companies face cost-cutting pressures from mergers.
“We’ve created an environment that is supportive, trusted, safe. Therefore the introduction of AI doesn’t feel like it’s layering on top,” Woolmington said.
Training extends into client environments through universal programs and leadership labs that address collaboration with AI agents as team members.
Leadership lab addresses future work
Canvas’s leadership development focuses on collaboration in AI environments, bringing agents “into a seat at the table” alongside human workers, and defining success metrics for future work structures.
“What we’re trying to do is talk about what does it mean to collaborate in this new environment? What does it mean to bring AI agents into a seat at the table next to you?” Woolmington said. “What will define the success of the work environment in the future is determined today and the future, not the past.”
The framework enables operational work streams, training programs, and partnerships with measurable KPIs rather than random testing without clear objectives or cadence.
“Rather than randomly doing a thousand tests, we can train against them, we can invest against them,” Woolmington said. “We can actually have KPIs internally and externally about them.”
You’re watching Beet.TV coverage from CES 2026. For more videos from this series, please visit this page.
