At Adobe Summit, AI dominated the conversation, but not in the abstract. The most practical discussions were about what AI is revealing inside modern marketing organizations: fragmented data, heavy processes, too many handoffs and stacks that have grown more complex than the outcomes they are supposed to deliver.

That was the focus of a fireside chat at Adobe Summit between Tribhuwan Kumar, Client Partner at HCLTech, Chris Reeder, Sr. Director, Marketing Technology at Microsoft and Joe Mele, GM – eCommerce, Digital Direct Sales at Microsoft. Their discussion made one thing clear: fixing the MarTech stack in the AI era is not only about adopting new tools. It is about rebuilding the conditions for better customer experience, faster execution and more empowered teams.

Why MarTech complexity is catching up with marketing teams

For Reeder, the starting point is straightforward. “It is complex,” he said. “There’s so many different tools to realize what we’re trying to accomplish.”

But the answer is not to begin with the tools themselves. “It all starts with the customer and working back to understand the capabilities that we need to reach the customer to sell our product.”

That sounds simple, but it is exactly where many organizations struggle. As MarTech estates have expanded, platforms have often been layered on top of one another to serve specific functions rather than orchestrated around a coherent customer and operating model. The result is more capability on paper, but also more friction in practice.

Reeder’s view was that the foundation still matters most. “It really starts with the data, making sure that we have the right data as a foundation, and then bringing together the capabilities for our marketers, for our merchandisers and for our studio teams to deliver that best experience for our customer.”

AI is exposing what was already broken

The discussion highlighted that AI is not just accelerating work. It is exposing structural weaknesses that were already there.

As Kumar noted, if something used to take three days and now a parallel AI-enabled solution can do it in half a day, shortcomings in the platform and process become much harder to ignore.

Reeder’s response was to bring the conversation back to the core issue: “We have to really get our data right foundationally and understand our data to get the best outcomes from AI.”

He added that the challenge is not simply the presence of more AI tools, but the need for consistency across them. “As we take a look at our end and stack and how we’re structuring it, we just need to make sure that the orchestration of the agents is using consistent information from the data.”

That is a useful framing for any organization trying to modernize its MarTech estate. AI does not remove the need for discipline in the stack. It raises the cost of not having it.

Speed is useful, but quality matters more

For Mele, the promise of AI goes beyond efficiency. “The promise of AI is not just to make things faster,” he said. “It is to be more efficient, speed things up, but it’s also to improve the quality.”

That point came up several times in the conversation. The real benefit of faster execution is not simply that work gets done sooner. It is that teams can spend more time on the thinking that matters.

“We talk a lot about shortening the distance between idea and execution, but part of that shortening is that you get to spend more time on the idea and more time on what really good looks like versus all the time that is spent on just trying to get to execution. And that should be the goal, not just speed,” said Mele.

Reeder reinforced that same shift in emphasis from tools to customer outcomes. “We really want our merchandising teams to not focus on the tools. We want them to focus on the customers and how we meet those customers where they’re at across the MarTech stack.”

That is an important distinction. A better MarTech stack is not one that simply pushes more work through. It is one that frees teams to focus on relevance, creativity and customer value.

Do not automate a bad process

“We don’t want to automate bad process. We want to automate good process,” said Mele.

That is the real test of AI-led transformation. If organizations simply layer AI over inefficient workflows, rigid handoffs or siloed teams, they may move faster, but they will not necessarily move better.

Mele described how AI is helping Microsoft step back and re-examine those structures.

“How do you use speed as a catalyst?” he asked. The acceleration created by AI is making it easier to see “where are the inefficiencies in our processes? Where have we built structures or processes or silos between teams that are not useful? Where have we created processes that are too heavy, or created handoffs, too many handoffs?”

This is what makes AI strategically useful for MarTech leaders. It is not only a productivity layer. It is a diagnostic force. It helps reveal what is holding the organization back.

Reeder took that further by connecting it to organization design. “AI is just enabling us to think differently in how we operate,” he said. “If we’re just automating the existing processes, we’re not really transforming.”

For him, the real opportunity is to move “the capabilities to the originator of the idea, closer to that frontier market experience,” so teams can be more agile and closer to the customer.

Upskilling is a culture shift rather than a training program

The conversation also moved beyond systems and into talent. Kumar referenced HCLTech’s Blueprint for AI Leadership research, which found that AI Leaders are far more likely than followers to give employees room to experiment.

In response, Mele said: “Upskilling is not training. Upskilling is a culture shift. It’s a way of thinking shift. It’s a mental model shift.”

He explained that Microsoft’s approach started with the recognition that the future team could not simply be hired in from outside. It had to be built from within. That meant moving beyond chat-based experimentation into broader AI tool usage, including coding environments, while also recognizing that some people will emerge as “super users” who can think in more advanced ways.

More importantly, it meant redefining work itself. “Your job is not the activities you do. Your job is an output...and to deliver great outcomes,” he continued.

Reeder added that this gives people “agency” over their careers. If routine work can be reduced, the time gained can be redirected into “driving the business forward, thinking about the customer [and] thinking about new ways to operate.”

Change management is the harder problem

Both Microsoft leaders were clear that the hardest part is not the technology itself. It is the speed of change and the human response to it.

For Mele, the concern is not whether AI can keep advancing. It is whether people can absorb the pace of change without being overwhelmed. “We move at speed, and we also move with empathy,” he said, stressing the importance of supporting people as they adapt.

Reeder agreed. “Technology is easy. That’s easy part. It’s the change management.” He described the real leadership challenge as helping teams navigate change in a way that is structured, empathetic and sustainable.

What fixing the stack really means

In the AI era, fixing the MarTech stack is not just about rationalizing tools or consolidating platforms. It is about improving the data foundation, simplifying processes, reducing unnecessary handoffs and giving teams the confidence and freedom to work differently.

That means AI becomes valuable in two ways at once. It accelerates work and it exposes what still needs to change. The organizations that benefit most will be the ones willing to treat that second insight seriously.