Human-Centered AI: Lessons in Adoption and Empowerment

Two Audiences, One Challenge

Within minutes of kicking off, hands shot up across the room. One of us stepped off stage to hand the microphone to an attendee. That small gesture set the tone: this wasn’t going to be a one-way presentation. It would be a collaborative dialogue.

The curiosity was undeniable. Attendees wanted to understand Agentic AI; not just in theory, but in practice. As questions flowed, most leaned technical. So, we paused and asked:

“Show of hands — who here has a non-technical role?”

Half the room raised their hands.

That was the turning point. We realized we weren’t speaking to a single audience but two: technical professionals seeking solutions for complex systems and business leaders looking for practical applications, all within the same AI context.

We pivoted — and with clear intent. Half the room represented non-technical roles, and if the discussion was to remain inclusive and productive, it needed a foundation that everyone could engage with. Agentic AI is not solely a technical conversation; it is fundamentally about value, impact, and ensuring that organizations are solving the right problems in the right way. We reframed the dialogue from technical details to strategic purpose, allowing us to explore the “why” before the “how.” That shift transformed the energy in the room and deepened engagement across both audiences.

Bridging the Technical and Business Divide

The moment was eye-opening. The promise of AI is that it’s not solely technical. Tools like MS Copilot Studio and paradigms like “vibe-coding” embody a future where AI is democratized across the workforce, regardless of whether our role is technical or not.

As we explored this divide, we found alignment, seeing AI systems as essentially pattern recognition machines. Like any good engineer or product manager in that room, the attendees already understood the fundamental truth: machines need consistency to function effectively. When we talked about the necessity of templatizing information and preparing clean data, we could see the recognition in their faces. They’d all dealt with messy documentation and inconsistent processes before. This wasn’t foreign territory, just a new application of familiar principles.

What struck us most was how the attendees’ experiences reinforced something we’ve learned at One North: the highest AI adoption failures happen when organizations focus solely on the technology and skip the human-centered design work. It was rewarding to guide the discussion toward this principle in the room, bridging both audiences — technical and business — around a shared understanding that effective AI strategies begin with clarity of purpose and user needs. This shift allowed us to move beyond technical mechanics and into conversations about value, impact, and solving the right problems in the right way, with technology as an enabler rather than the sole focus.

We shared our perspective on why orchestration agents are critical for effective AI implementation. Rather than relying on a single, all-encompassing system to manage diverse tasks, we design networks of specialized agents with clear roles and purposes. For example, instead of one agent attempting to handle all aspects of sales support, we create focused agents — one dedicated to SOWs, another for proposals, and another for pitch decks. Each agent is optimized for its specific function, and together they operate through orchestration to deliver a more reliable, scalable, and efficient solution.

The Challenges of Our Current Environment

The volume and variety of questions we received highlight the complexity of today’s AI landscape. Change is happening fast. We still have work to do, but we also need to adapt and find new ways to work, lead, and build.

There is no shortage of voices promising AI’s superpowers and instant results. In reality, success is far more nuanced. As we explored the challenges of current technology and the legacy systems and technical debt that came before, we focused on building knowledge, capabilities, and momentum with AI and agents.

It would be nice if success came with a simple formula: do X, then Y, and voilà. But the truth is more complicated. Through stories of our own journey — the wins, the setbacks, and the lessons — we returned to a fundamental idea: progress does not come from textbooks or lectures. It comes from engagement. From building, testing, and exploring.

Learning AI is like learning to ride a bike. You have to feel the wobble, find your balance, and, sometimes, take a fall. That is how real growth happens.

Permission to Play

The key insight we kept returning to? We all need permission to play. Play has been the fastest, and most fun, way to learn. We seem to forget that truth. After all, we’re adults. And this is work. And technology. So, it needs to be serious. It needs to be formal. It needs to follow THE plan.

But with permission to play comes responsibility. We have boundaries and constraints. We need to be thoughtful about where, when, and how we play. About the data we use, where our creations live, and the policies we need to adhere to. But even within these boundaries, there are numerous opportunities to build our understanding. To discover the nuances. To see what’s real, and what’s future potential. And there are opportunities to imagine, create, and engage. To see the wonder in new capabilities and how we might use them, engage with others, and bring experiences to our colleagues, partners, and customers.

What This Means Going Forward

The Adobe PgM Summit reminded us that every successful AI initiative begins with people. Before exploring technical capabilities, we need to understand the realities of work — the workflows, pain points, gaps, and opportunities that shape how things get done. When we start with people, we build AI solutions that are truly useful. Without that foundation, even the most advanced technology can miss the mark.

This shift calls for a more collaborative mindset. People working closest to the business are encouraged to get involved, join conversations, ask questions, and experiment with the tools available. Learning through hands-on experience often leads to the most meaningful insights. And that experience, that authenticity, then helps ideas to spread and flourish.

At the same time, technical teams have a chance to lead by making AI more approachable. By inviting others into the process, sharing knowledge, and supporting exploration, they help build confidence and foster innovation across the organization.

AI is not just a technology initiative. It is a transformation of how we work. Everyone has a role to play, and the best way to begin is by staying curious and engaged. We’re grateful to Adobe for inviting us to the PgM Summit and for fostering such an open, passionate dialogue. Their commitment to innovation and to delivering success for their customers made the discussion both energizing and insightful.

Ready to explore what AI could do for your organization? We’d love to help you figure out where to start, what to build, and how to give your team permission to play.

Photo Credit: Sylwia Bartyzel | Unsplash