Last November, a few days before I ran the Athens Marathon, I went looking for a hole in the ground.

Not just any hole in the ground. The Stoa Poikile—the Painted Stoa—was one of the few places in Athens I had decided beforehand that I absolutely had to see. The Acropolis was magnificent, obviously. The Panathenaic Stadium mattered for reasons that probably require no explanation from someone who had come to Athens to run a marathon. But the Stoa was different. Somewhere amid the traffic, storefronts, apartment buildings and ordinary clutter of modern Athens were the remains of the place where Zeno once taught and where Stoicism got its name.

There is not much left to admire architecturally. If you didn’t know what you were looking at, you might reasonably wonder why someone had traveled several thousand miles to stand beside what looks a little like an archaeological trash pit, complete with corrugated steel sheeting covering excavated holes in the landscape.

I wasn’t there for the architecture. I was interested in the practice the place represented: people with different experiences and forms of knowledge making a habit of coming together, arguing, questioning assumptions, testing ideas and trying to arrive at some kind of wisdom none of them possessed alone. I had been thinking about spaces like that for years. Looking back now, I wonder whether some part of the current version of Workflow Studio was planted there.

When the Model Doesn’t Know Tuesday Afternoon

Months later, I was back at Vizcaya Museum and Gardens, where I work as technology director. A necessary disclaimer before I get myself into trouble: this article reflects my own professional inquiry and does not represent an official position of Vizcaya Museum and Gardens.

We had begun experimenting with something I was calling Workflow Studio. It is not supposed to be another software training. It isn’t really an AI program either, although AI has developed an uncanny ability to wander into nearly every technology conversation whether invited or not. The idea is simpler, and maybe stranger, than that: bring people together around something they are actually trying to solve, add people who look at the problem from somewhere else, and give them enough structure that the conversation has somewhere to start but enough room that it does not have to end where anybody expected.

At one of the first sessions, my colleague in visitor services, John Settles, brought in a forecasting model he had built using admissions and ticketing data. On the surface, we were talking about visitation, pricing, online purchases and revenue. Our guest, neuroscientist and AI educator Michael Mannino, almost immediately began asking questions more difficult than anything the model could calculate. What were the assumptions? What was known and what was projected? How would we verify it? What would someone else need to know before trusting the result?

Somewhere in that exchange, Michael landed on a sentence I haven’t been able to shake: We can outsource parts of the task. We can’t outsource understanding.

That was about the point when Workflow Studio stopped looking to me like a technology session, because John knew things the model didn’t. Vizcaya right now is less a static historic site than a moving target. Windows and doors are being replaced for hurricane resistance. Collections have temporarily disappeared from some rooms. Construction changes where visitors can go and what they encounter when they get there. Somewhere inside all of that is a person who bought a ticket and is quietly deciding whether the experience they received bears much resemblance to the experience they thought they were buying. It is, at the moment, a considerable distance from the uninterrupted Gilded Age opulence for which we are generally known.

You can feed a model years of admissions data without giving it the faintest idea what that visitor saw Tuesday afternoon. And that isn’t because the context is unavailable. In its current state of glorious disruption, the place is practically screaming context at us. The harder problem is noticing which context matters and, perhaps more importantly, having enough imagination to stand somewhere other than where you normally stand in order to see it.

John knows one part. Visitor-facing staff know another. Collections staff know what disappeared from a room and why. Finance sees another piece of the picture. Somebody managing a construction project is thinking about schedules, dependencies and contractors while a visitor is thinking, quite reasonably, I paid how much to see this? None of those perspectives is false, and none is sufficient. Putting everybody into a meeting does not magically solve that either.

I have been in enough collaborative projects to know that you can have all the collaboration in the world and still watch important things sail directly over everyone’s heads because everybody’s gaze is fixed on roughly the same plane. That is why I have started thinking that context is not primarily an information problem. It is a perspective problem.

AI makes this especially obvious. It can help us organize, calculate, summarize and find patterns. It can reveal relationships that would take us much longer to notice on our own. But it does not arrive knowing which human vantage point matters. Somebody still has to observe keenly enough to recognize what is important, gather pieces that live in different heads and different corners of an institution, and decide what they mean together. At some point that stops sounding like a technology problem and starts sounding like a question about how institutions think.

Respect Before Collaboration

We talk a great deal about collaboration as though getting enough smart people into the same room is the magic formula. It isn’t. The more basic requirement is respect—not politeness, collegiality or agreeing with everybody, but respect in the deeper sense that you genuinely believe another person may be capable of seeing something you do not see and contributing beyond the narrow lane the institution has assigned to them.

Without that, collaboration is mostly theater.

Institutions are organized around expertise for perfectly good reasons. Somebody should know who is responsible for the collection, the garden, the budget, the technology, the visitor experience. But expertise has a way of hardening into territory. “Stay in your lane” starts sounding like good organizational discipline at precisely the moment many of the problems worth solving have stopped respecting the lanes. AI doesn’t respect them. Climate adaptation doesn’t. Public trust doesn’t. Digital preservation, institutional memory, audience behavior and the increasingly complicated ways people encounter cultural organizations certainly don’t.

The answer is not that everybody should suddenly become an expert in everything. That sounds exhausting, among other problems. The point is that people are frequently capable of contributing things their job descriptions never anticipated.

I think about people who became central to earlier projects I worked on, including HistChem and Beyond Vizcaya: a historian, a retired chemist, a learning-program facilitator. People who brought curiosity, judgment and an authentic voice to work that lived well outside what might have appeared to be their normal comfort zones. I did not invite them because an org chart told me they belonged there. I heard something in the way they thought and invited them into the development of the idea.

That distinction matters. You can ask somebody to participate in a thing you have already designed, or you can recognize that they have something you did not anticipate and allow that contribution to change what the thing becomes. The second is considerably less predictable. It is also where some of the best work I have ever been part of has come from.

When institutions fail to hear those voices, they lose more than individual contributions. They lose possible versions of themselves: different ways of thinking and working, unexpected routes toward goals that seem impossibly far away, connections nobody knew were there because nobody had thought to put those particular people, problems and ideas in the same room. Too often those voices pass through unheard because everybody is listening for the same tune and rhythm they already know.

When a Space Becomes Their Wings

I have seen what can happen when you change the conditions. Years ago, one version of this was something we called GLAM Café. The name was fairly literal. It had the feeling of a cybercafé: intimate breakout spaces, a little food—which never hurts—and enough informality that people could decide from month to month whether they wanted to socialize, disappear down some intellectual rabbit hole, or find somebody else who cared about the same strange problem they did.

People came from different institutions, but they shared enough purpose to recognize one another. Mary Ockerbloom was working with me as a Wikipedian in Residence and brought a network of Wikipedians who had largely known one another virtually. Philadelphia’s digital-humanities community was gaining traction. THATCamp had created another gathering point. At GLAM Café, some of those connections became more granular.

One group might be working on a Wikipedia edit-a-thon while another was talking through a big-data project or an interactive. Nobody had to work on the same thing. That was part of the point. It gave scattered ideas somewhere to land, but more importantly it gave people who sometimes felt isolated or unsupported inside their own institutions a place where their curiosity did not have to justify itself before they were allowed to act on it.

It became their wings.

That remains one of the tests I keep in the back of my mind for Workflow Studio. Not attendance, not whether people tell me afterward that they enjoyed the session, and not even whether we build something useful. The interesting test will be whether, eventually, I do not have to be there to know something meaningful is happening.

People find one another. Small groups form around questions nobody assigned them. Someone recognizes a capacity in a colleague that their title does not advertise and pulls them into a problem. The practice escapes the program. At that point, it starts becoming culture.

The Point Was Never the Porch

This is also where experiments like these become fragile. A champion can light something up quickly. The harder question is what happens when day-to-day work returns, priorities shift, appreciation is uneven, budgets tighten, attention wanders or the person carrying the thing eventually leaves.

If the habit disappears with the champion, the institution never really learned it.

It borrowed it.

Which may be why I keep returning to that unimpressive hole in the ground in Athens. The point was never the porch. The point was the practice.

A modern stoa does not need marble columns. It needs repeatable conditions. Bring a real problem instead of a polished presentation. Keep the pressure low enough that people can expose uncertainty without immediately regretting it. Bring together people who would not ordinarily be expected to have a stake in the same question. Let an idea become interesting before immediately asking who owns it, what it costs, whether it appears in the strategic plan and who has just accidentally volunteered themselves for six months of additional work. And leave enough room for the frame around the problem to change.

At that same Workflow Studio session, Vizcaya horticulture manager Quinn Metoyer started with an extremely practical question. Could someone working in the field simply dictate observations rather than stopping whatever they were doing to populate a database? Could those observations follow a crop from planting through harvest and then connect outward to school programs, volunteers, public events, donor reporting and larger measures of impact?

A conversation that began with cilantro seeds wandered into workflow design, storytelling, data, institutional memory and stewardship. That wandering was not the failure of the conversation. It was the useful part.

Too loose, and you have pleasant conversation. Too structured, and everyone remains trapped inside the problem exactly as it was originally defined. Somewhere between those two is a space where the practical problem stays on the table while people are still permitted to redraw the boundaries around it.

That space interests me enormously.

Stepping Outside the Plan—Strategically

The deeper reason for cultivating this habit is not collaboration for collaboration’s sake. It is adaptation.

Strategic plans are useful. They force choices. They establish priorities, budgets and commitments. But once approved, they acquire gravity. An institution increasingly organizes itself around executing the plan, and the questions naturally become whether we are on schedule, on budget and hitting our targets. Those are all necessary questions.

But institutions also need mechanisms for periodically stepping outside the strategic plan strategically. Not every week, and certainly not every time somebody has a bright idea. Often enough to clear the head, turn down the familiar soundtrack, and ask whether the assumptions beneath our decisions still fit the times we are actually living in.

That is difficult because institutions are rewarded for certainty. A genuinely unfamiliar idea can look like one more threat to it: more work, more expense, disrupted authority, changed expectations, another moving part. If every new idea arrives carrying the possibility of workload, cost or some form of microreprisal, people become remarkably efficient at learning not to have ideas in public.

The irony is that preserving certainty too aggressively may be one of the least stable things an institution can do. Real stability requires confidence that the institution can adapt on something closer to an evolutionary scale—not simply make this month’s revenue number, balance this year’s budget or execute the current strategic plan, but notice when the world has changed and remain capable of changing with it.

Maybe a modern stoa is one mechanism for practicing that.

From Flow to Collective Adaptation

As Quinn and I walked Michael Mannino toward the parking lot after the Workflow Studio session, the conversation was still moving faster than our feet. We would be finishing one thought while already launching into another. At some point I mentioned flow state and told Michael that flow had been baked into the idea of Workflow Studio from the beginning.

That was when he told me this was his field.

Michael studies flow as a neuroscientist. What interested me was not simply the familiar idea of being “in the zone.” It was the possibility that flow is not entirely accidental. Under the right conditions, people can become more deeply absorbed, attentive and responsive to the thing in front of them.

I am wary of claiming that an institution can enter flow in the same sense that an individual can. That is exactly the sort of metaphor that becomes irresistible and then gets repeated until everyone forgets it was a metaphor. But the conditions are interesting. Attention matters. Trust matters. Challenge matters. Distraction matters. And so does whether people feel safe enough to stop performing certainty and start actually thinking.

For a little while in that room, some normal boundaries had loosened. A forecasting model became a discussion about trust. Cilantro became a discussion about institutional memory. Technology became a discussion about judgment. One person’s question changed the frame through which everyone else was looking.

Maybe the institutional outcome is not “collective flow” in the same way an individual experiences flow. But if flow describes a state in which attention, trust and challenge align well enough for a person to think and act at their best, it seems worth asking what the institutional equivalent might be. Maybe flow is one of the conditions that makes collective adaptation possible.

Not adaptation as crisis response—the frantic pivot after circumstances have already made the decision for you. Something earlier and more deliberate: a shared capacity to notice signals, surface different perspectives, question inherited assumptions, experiment and change course while there is still room to choose.

Workflow Studio is much too young for me to claim it has accomplished that. It may work. It may not. I have been around enough ambitious projects to distrust anybody who can already tell you how the experiment ends.

But I am increasingly convinced that cultural institutions need spaces where something like this can happen. Which is what keeps bringing me back to the Painted Stoa.

Not the ruins. Not even Stoicism … The Practice.

People returning to a place where they could think together seriously enough, and often enough, that eventually the habit mattered more than the structure that held it.

If Workflow Studio ever works the way I hope it can, perhaps the best evidence will be that nobody needs to call it Workflow Studio anymore. The practice will have escaped the room. People will have learned to recognize unexpected capacity in one another, step outside the familiar rhythm long enough to hear a different voice, and adapt together before the future makes the decision for them.

That sounds like a modern stoa to me.

And I would very much like to see what happens when we start building them across the cultural landscape.

If any of this resonates, here’s a place to start.