Postcards from the Field: The Orchestration Gap
Postcards from the Field is where I share brief field reports from my work with real leadership teams. Rather than polished case studies, these notes surface the moments that stop me mid-session and point to the everyday habits that shape performance when conditions are uncertain. My aim is to translate what I’m seeing in the field into practical value readers can apply in their own teams.
This week, I spent an evening with a room full of people analytics leaders here in Austin. The group convenes regularly to exchange notes on what they are seeing inside their organizations, and this time they asked me to speak about what changes on teams when AI becomes embedded in the workflow rather than sitting off to the side as a drafting tool.
I should admit at the outset that I ran long. Facepalm! This is a constant growth edge for me…I get geeked about all the things I think the audience will want to know or experience, and I add just one too many slides into the deck. During the Q&A, one audience member said, “This felt like two whole keynotes!” Signal that yours truly did the thing again. GAH…working on it.
One of the areas that got shortchanged was the final segment of the keynote, where we explored what I am seeing in a small subset of AI-heavy teams. This deserved more breathing room than the clock allowed. It is a messy topic, still very much in motion. If I had Hermione Granger’s “Time Turner” from Book 3, I’d have fully explained the following. And, yes, that was a full-on nerdcore Harry Potter reference.
In most organizations, AI remains assistive. It helps generate language, summarize meetings, draft code, or analyze data more quickly. The underlying workflow, however, is largely unchanged. Humans still define the problem, sequence the work, and absorb the accountability.
But in a handful of teams I coach, something more structural is beginning to shift. AI is not simply accelerating tasks. It is altering the shape of the loop itself. Work moves differently. The number of small decisions the team has to make grows quickly. Work can iterate quickly, but endlessly. In those environments, AI starts to feel less like a tool and more like a teammate, not because it possesses judgment, but because it now occupies a defined place inside the team system.
What has surprised me is not the technical sophistication of these teams. It is the ways of working they have had to build in response.
The first adjustment is sharper framing. As output scales, ambiguity becomes more expensive. When a system can generate ten variations in seconds, the absence of a clear definition of “high-quality” quickly turns into churn. In the more stable teams, leaders have become unusually explicit about what matters now, what “good enough” looks like in this context, and when a piece of work is considered done. This is not about adding flowery, aspirational language, but about defining functional guardrails for major workstreams. Without it, the speed AI introduces does not create the execution outcomes that we want.
The second shift involves coordination. Once AI is embedded in the workflow, handoffs multiply. Work now moves from human to AI, from AI to human, and then often across to another human whose expectations may not match the original framing. In teams that are struggling, the sequencing of these handoffs remains implicit. The result is rework, frustration, or simply bad product. In teams that are rethinking their ways of working to match this profound technological shift, the handoffs are clearly mapped. Who initiates the loop, who validates the output, and who carries final accountability are clarified. Iteration is sequenced in service of completion, not endless refinement.
A third adjustment is more human than technical. As certain tasks become automated, even partially, people naturally begin to reassess where their human judgment and contribution matter. On teams that are rapidly adding AI, but not changing their ways of working, trust and belonging erode fast. The teams that are operating closer to real orchestration have not ignored that undercurrent. Leaders are actively reinforcing where human discernment is essential, where context cannot be outsourced, and how individual contributions shape the final outcome. They are also investing in relational teaming habits that stabilize trust and belonging, even as speed increases. People need evidence that they are still seen and valued inside the loop.
Finally, there is calibration. AI introduces abundance. More drafts. More analyses. More options. Left unchecked, this abundance expands scope. Teams find themselves doing more simply because they can. In the AI-heavy environments I am watching closely, leaders are learning to decide where to go deliberately slower, where to prune low-impact work, and how to install feedback rituals that match the new tempo. A quarterly after-action review is rarely sufficient when iteration cycles have shortened to days or even hours.
It is important to say that most organizations are not operating in this version of the future yet. In higher education, healthcare, banking, and many other regulated sectors, bureaucratic and compliance constraints rightly temper how quickly workflows can change. For many teams, AI remains an experiment at the margins or a writing assistant.
But in the rooms where AI has begun to function more like a teammate than a tool, orchestration has become a core leadership skill.
That was the part of the conversation that lingered after the meetup ended. The tech is evolving quickly. Our approach to teaming needs to catch up.
