Epistemic Design — Bridging Communal Design & Institutional Design

A collaboration is itself a designed instrument. A designer–designer pair has certain perceptual capacities. A designer–developer pair has different ones. A designer–founder pair has another. Each pairing creates its own language, tempo, tolerance for uncertainty, decision style, resolution, blind spots, and expressive range.

The question no longer becomes: “who is on this project?” It swiftly turns into: “What form of intelligence emerges between these particular people under these particular conditions?”

Agent-config works in parallel patterns: a model cannot be properly understood in isolation from prompt, task, memory, tools, and user. Likewise, a designer cannot be properly understood in isolation from collaborator, institution, incentives, constraints, and trust. Pair Design, within Communal Design, reveals the importance of the relationships between capacities.

As these theories are tested and hypothesized, Who* Math— defined in 2017, where the use of the asterisk attached to "Who" shows relativity of identity, may not be a (purely) additive equation, in 2026. Once a team is formed, each step below produces a categorical and felt change. “Me” and “You” are individual identities. “We” is a new relational identity. “Us” introduces a boundary, shared context, and likely shared stakes. “Community” introduces continuity, memory, norms, and belonging.

Perhaps Who* Math is no longer conventional arithmetic — rather a mathematics of state transition through relationship. Transformation over (purely) quantity:

The intentional act of pairing within a team, versus not, reveals that Who* Math may hold multiple pathways based on “for whom and what.” A more static, additive approach may be the first step to collaboration before a team forms up. Then, the transformation of Pair Design reveals its true self.

A statement of work often defines a design problem as an artifact deficit:

➔ No brand = create brand ➔ No website = create website ➔ No campaign = create campaign

Once we start peeking through the cracks of any given SOW, the line items rarely add up to the artifact we're rearranging with the client and our teams. Things like unclear authority, weak communication channels, missing roles, unarticulated knowledge, low trust, broken decision pipelines, absent institutional memory, and incentives misaligned with intent often show up long before final designs or ongoing development.

Within standard SOWs and projects, these secondary problems are often not identified before the work begins — then these invisible forces become the design challenge itself. A design problem is a condition in which the relationships necessary for an intended outcome are missing, misaligned, invisible, or unable to persist. That definition is broad enough to include interfaces, teams, organizations, AI systems, governance, and institutions. It also explains why traditional deliverables often feel insufficient: they modify the visible object without necessarily changing the relational conditions that produced the problem in the first place.

The situated claim is epistemically stronger because it describes the context that generated the observation.

We cannot point to trust, or things like intent, the same way we can point to a button or logo. Nevertheless, even though invisible forces like trust and intent are invisible, they contain observable effects: changes in state, and individual & collective wins and losses.

In physics, a force is often inferred through its effects on motion. In organizations, a team member or client might infer trust through its effects on behavior. These invisible forces shape the organization long before artifacts are delivered.

Institutional Design (ID) then becomes an important figure within Epistemic Design. ID could be understood as the practice of identifying and shaping invisible forces so that desired behaviors can emerge and persist.

A client or team member's “Who*” is not only a named identity. It is a stable pattern of behavior that remains recognizable across time. The “How” becomes the relational system that makes that pattern remain recognizable and reproducible.

Both humans and AI systems are composite systems whose identity is inferred from recurring behavior. When working with composite systems, the perceived Who* is produced by a network of relationships. Inside a project we say things like “this is how the company behaves” or “this is how this team works.” This approach creates a stable pattern of behavior recognizable across time, budget, expectation, artifact, and project.

The epistemic shift moves from asking about isolated properties to asking about situated performance. The old framing looked like: Is this designer good? Is this model good? Is this tool good? Is this organization effective? The relational framing looks like: Effective with whom? Under what constraints? Toward what intent? Through which interface? With what degree of trust? Over what period of time? Inside which governance structure?

Providing people with this epistemology, we start to carve out trust through truth-building — replacing universal claims with contextual and conditional claims: an assertion that something is true or a demand for something believed to be due.

For a Design Director role, the problem space becomes something like: “Design the relational conditions through which people, models, tools, products, and institutions can produce coherent behavior at scale.” This includes both visible systems and invisible forces:

➢ How designers and engineers pair · ➢ How AI participates in decision-making · ➢ How intent survives handoffs · ➢ How local teams interpret global systems · ➢ How trust is established between humans and automated systems · ➢ How knowledge moves · ➢ How decisions become memory · ➢ How identity persists as products change.

The Design Director is no longer only directing artifacts: they are directing the conditions of emergence itself. A Design Directors capacity changes depending on who they work with. A model's capacity changes depending on how it is prompted and tooled. An organization's capacity changes depending on its incentives, trust, governance, and memory. A community's capacity changes depending on how participation is structured.

The meaningful unit of design may no longer be the arithmetic 1+1 = 2 artifact, the individual, or even the system component itself. In 2026, the relationship that changes what each component is capable of becoming and how we quantify that bridge seems to be the new object emerging.