Authored by Simon Hodgkins, CMO at Vistatec
As CEOs and CMOs look for signs that AI is shifting from hype to real results, the leaders who are already ahead show that closer human and AI collaboration, supported by stronger governance, is what actually drives progress.
We keep hearing that artificial intelligence is set to transform how we market, sell, and build relationships with customers. Yet if we are being honest, most leaders still feel the gap between big expectations and the impact they can point to. A recent McKinsey piece on agentic AI highlights something important. The real gains come not from new tools but from rethinking how work actually happens.
As someone who lives at the intersection of marketing, technology, and growth, I found the ideas both practical and a bit of a wake-up call. They made me think about where teams get stuck and where real change begins.
Why agentic AI matters
Agentic AI refers to systems that do more than assist. They act, coordinate, and manage multi-step workflows. That is a meaningful shift. It matters for three reasons.
First, real value only shows up when AI links to the biggest growth problems. Incremental productivity is useful, but it is not what moves revenue or customer engagement in a measurable way. If a company improves targeting but ignores how leads flow to sales, the outcome will always be limited.
Second, the workflow matters more than the tool. It is tempting to believe a new AI platform will solve everything. It rarely does. What creates impact is understanding how decisions move through a process, how information is handed off, and where an agent can influence an outcome rather than simply provide an answer. I have seen teams fall into the trap of adding AI to a workflow instead of reworking the workflow itself.
Third, scale requires a new operating model. This is the part most organizations underestimate. You cannot scale by simply deploying more bots. You need cross-functional ownership, shared data products, and a structure that treats AI agents almost like managed contributors. Without this, things get messy quickly. Duplicated builds, inconsistent quality, rising risk. I have watched this happen in more than one company.
What leading organizations are doing
A few examples illustrate what happens when teams redesign work instead of patching it.
A European insurer created personalized campaigns across hundreds of microsegments. Their agents adapted sales scripts in real time and coached frontline teams based on customer cues. They saw significant gains in conversion and reduced service call times.
An airline used predictive models and agent-led workflows to change how it supported travelers during disruptions. Frequent fliers received one type of response, occasional travelers another. Customer satisfaction rose, and churn fell in a measurable way.
A North American manufacturer rebuilt its service organization so that AI handled diagnosis, data retrieval, and summarization. Humans focused on the interpersonal resolution work they were best equipped to handle. The shift improved both service quality and employee morale.
These stories have something in common. None of these companies simply plugged in a tool. They redesigned a sequence of actions, identified the moments that mattered, and placed AI in a role that shaped the outcome.
How to lead this inside your organization
If you oversee marketing, growth, or customer engagement, here is how you might think about the shift.
Identify where decisions drive outcomes. Look for the moments where human judgment affects revenue, loyalty, or efficiency. These are usually the best places to start. Map them. Understand what slows them down. Then ask where an agent might act instead of simply advising.
Map the full workflow, not the isolated task. I have learned that teams often assume they know their own processes. They rarely do. When you map every step, you find delays, duplication, and gaps in ownership. You also see where an agent can support resolution without interrupting the human flow.
Define agent roles, metrics, and governance. If an agent will take action, treat it like a team member. Define its responsibilities. Set KPIs. Decide how you will monitor its learning curve and its accuracy. Without shared standards, things drift.
Strengthen cross-functional capability. Marketing cannot do this alone. Neither can sales, service, data science, or operations. The organizations that do this well create hybrid teams that share accountability for the outcome, not just the tool.
Evolve human skills. This is the most underestimated piece. As agents mature, human skills become more important, not less. Teams need to learn prompt design, escalation management, and continuous optimization. They also need to become comfortable orchestrating humans and agents together.
The risks you need to manage
Every transformation carries risk. Here are the big ones I see most often.
Siloed pilots with limited impact. Companies test AI in small pockets, then wonder why nothing scales. If a pilot is not tied to a core business problem, it will not move the needle.
Uncontrolled agent proliferation. When every team builds its own agents without oversight, the organization becomes flooded with overlapping and inconsistent systems. You need a shared approach and clear ownership to avoid confusion.
Weak data foundations. Agents rely on accurate and connected data. If the data layer is fragmented, the value collapses. This usually requires work earlier in the process than people expect.
Human misalignment. Some teams fear AI will replace them. Others misunderstand how it fits into their workflow. Clear role definitions help. Humans should focus on strategy, judgment, empathy, and creative decisions. Agents take the repetitive work, the complex coordination, and the tasks that slow people down.
My take on where this is heading
The temptation right now is to move fast. Everyone wants to show progress, and the pressure can be intense. But in my experience, speed without structure creates noise rather than value.
Real impact comes from discipline. It comes from asking where work should be redesigned, not just improved. It comes from thoughtfully aligning humans and agents. And it comes from leadership that is willing to question old assumptions about how work gets done.
The companies that will stand out are not the ones that deploy the highest number of agents. They are the ones who choose the right use cases, rethink the workflow, and create an environment where humans and AI complement each other naturally.
The promise of AI turns into real value when we rethink the work, not just the tools. I would love to hear where you see the most meaningful opportunity for agent-based systems in your own marketing or growth function.
