This article was based on the interview with Front’s Kevin Yang on maximizing both speed and coordination by Greg Kihlström, AI and MarTech keynote speaker for The Agile Brand with Greg Kihlström podcast. Listen to the original episode here:
We are in the midst of an enterprise-wide mandate: adopt AI, move faster, become more efficient. Marketing leaders, in particular, are under immense pressure to demonstrate ROI on significant technology investments, and the promise of generative AI feels like the answer we’ve been waiting for. We see the potential everywhere—from content creation and campaign personalization to customer service and analytics. The logic is straightforward: if we can make each individual on our team more productive with powerful AI tools, the entire organization will, by extension, become more effective. But what if that logic is flawed?
This is the central tension that many of us are now beginning to experience. While individual task completion may be accelerating, the overall velocity of our customer-facing operations feels stubbornly stagnant, or in some cases, even more convoluted. The problem isn’t the technology itself, but our application of it. We are layering sophisticated automation onto workflows that were never designed for it, creating new friction points and hidden costs. In a recent conversation, Kevin Yang, the Director of AI at Front, introduced a powerful concept to describe this phenomenon: the “coordination tax.” It’s a term that should be on the desk of every leader planning their next AI initiative, as it perfectly captures the invisible drag on our teams’ performance.
The Real Cost of Customer Operations: The Coordination Tax
For decades, we have measured customer operations through a standard set of metrics: response time, resolution time, and customer satisfaction (CSAT). These are lagging indicators of efficiency. They tell us the outcome, but they reveal nothing about the internal effort, the cross-team handoffs, or the system-hopping required to achieve that outcome. Yang’s research quantifies this invisible work, and the results are, to put it mildly, sobering. He argues that the majority of time isn’t spent solving the customer’s problem, but navigating the internal complexity to get to the solution.
“What we found is that on average, teams are spending three hours coordinating for every hour they’re actually spending solving customer problems. And… the teams that are afflicted with heavy coordination and taxes say that they spend most of their time finding the right person or system, managing handoffs, re-explaining context, and then closing loops between the teams rather than actually fixing the customer problem.”
A 3:1 ratio of coordination to problem-solving is a startling figure. It suggests that for every dollar we spend on a highly-trained customer success manager or marketing specialist to engage with a customer, we are spending three dollars on the internal friction of our own organization. This is the coordination tax. It’s the time spent in a Slack channel trying to find an engineer who can answer a technical question, the effort to summarize a long email thread for a colleague in another department, and the all-too-familiar dance of re-explaining context every time an issue is escalated. For marketing leaders, this tax manifests as delayed campaign launches, inconsistent messaging as information gets lost in translation, and a disjointed customer experience because the journey is fragmented across siloed teams and systems. Recognizing this tax is the first step; understanding why our modern tech stack often makes it worse is the crucial next step.
The Automation Paradox: When More Tech Creates More Friction
The most counterintuitive finding from Yang’s work is what happens inside organizations that are ahead of the curve on AI adoption. One would assume these “AI-first” companies would have the lowest coordination tax. They have the most advanced tools, after all. Yet, the opposite is often true. These organizations report the highest satisfaction with their technology at an individual level, while simultaneously reporting more coordination issues at the team level. This is the automation paradox.
“AI first organizations, the ones with the most adoption of AI, do simultaneously report the highest satisfaction with their technology, but also, reporting more coordination issues as a result. The takeaway isn’t that, hey, AI doesn’t work, it’s that automation alone doesn’t solve coordination. AI actually performs best when it’s embedded in the workflows with access to full customer context, the right systems, and clear ownership across teams.”
The analogy of a highway is useful here. Giving one driver a sports car in the middle of a traffic jam doesn’t get everyone to their destination any faster. It just creates frustration. Similarly, giving a marketing operations specialist an AI agent that can draft 100 email variants in a minute doesn’t help if the legal, brand, and product teams are still operating on a manual, week-long review cycle. The bottleneck simply moves, and the increased output from the AI-powered individual can actually overwhelm the downstream human processes, creating more management overhead and more coordination chaos. The lesson is clear: you cannot simply bolt AI onto an existing, fragmented workflow and expect organizational-level gains. True progress requires a fundamental re-architecting of the work itself.
Re-Architecting Work: Delineating Human and AI Responsibilities
If simply automating existing tasks is a trap, how do we move forward? The solution lies in deliberately redesigning workflows by playing to the distinct strengths of humans and machines. Too many leaders fall into the trap of trying to automate an entire human-led process from end-to-end. This is a fragile and often frustrating approach. A more robust strategy is to deconstruct the workflow into its component tasks and assign them to the right resource—human or AI.
“What companies should be doing is saying, ‘All right, here are the research and analysis tasks that AI is just excellent at… And then over here we have another bucket that is empathy and context and especially relationships, that at this point, AI is not quite able to grok.’ …By kind of restructuring the workflows, they’re actually going to be able to reap the rewards and the gains of what AI is capable of without being frustrated that AI is not good at the relationship stuff.”
This is where the rubber meets the road for leadership. It requires mapping the customer journey not just from the customer’s perspective, but from an internal, operational one. Consider an account-based marketing (ABM) motion. The AI’s job is research and analysis: monitoring buying signals, analyzing past communications for sentiment, summarizing product usage data, and drafting initial outreach based on firmographic data. These are tasks that AI can perform more thoroughly and faster than any human. The human account manager’s job, however, is relationships and judgment: interpreting the AI’s analysis, understanding the nuanced political landscape within the target account, building rapport with a key champion, and making the final strategic decision on when and how to engage. By structuring the workflow this way, we empower our teams to operate at a higher strategic level, using AI as an intelligent assistant rather than a flawed replacement.
Governing the Future: Managing a Multi-Agent World
As it becomes easier for teams to build or buy specialized AI agents, we are heading toward a new and even more complex coordination challenge. If we are not careful, every department—marketing, sales, support, product—will have its own set of agents operating with different data sets, different instructions, and different goals. The risk of creating a chaotic, disjointed customer experience is immense. The next frontier of operational excellence will be about governance and creating a shared source of truth in a multi-agent environment.
“Just having multiple humans involved in a workflow creates coordination challenges, and this can get much worse when you introduce one or more AI agents into the mix, because they can propagate mistakes system-wide, instantaneously. There are two big governance gaps between a prototype of an AI agent and one that’s production ready, and it’s two types of trust: trust that the AI agent’s behaving as expected… and trust that it’s providing high-quality interactions.”
Yang’s concept of a “bring your own agent” initiative, supported by a central governance layer, points toward the future. Platforms must evolve to become orchestrators, not just containers, of work. We need systems that can provide guardrails, unit tests for agent behavior, and continuous quality monitoring (for both AI and human interactions) across the entire customer lifecycle. For marketing leaders, this means demanding interoperability and establishing clear rules of engagement for any AI agent that touches a customer. Before our tech stacks begin to resemble a chaotic digital bazaar of single-purpose agents, we must build the infrastructure to ensure they work in concert, not at cross-purposes.
The rush to implement AI is understandable, but speed without direction is just a faster way to get lost. The concept of the “coordination tax” provides a critical lens through which to view our technology strategy. It forces us to look beyond the surface-level metrics of individual productivity and ask harder questions about the underlying health of our operational workflows. Are we simply making individuals faster, or are we making the entire team more effective? Are we automating tasks, or are we truly orchestrating better customer outcomes?
The path forward requires a shift in mindset. We must evolve from being buyers of technology to being architects of human-AI systems. This involves meticulously mapping our processes, ruthlessly eliminating sources of friction, and thoughtfully delineating the roles of our people and our technology. As Kevin Yang’s insights suggest, the greatest returns from AI won’t come from the companies that adopt it the fastest, but from those that integrate it the most intelligently into the core of their customer operations. By focusing on reducing the coordination tax, we can ensure our investments in AI pay the dividends of genuine, sustainable agility.
