The squad model

A squad for a company's hardest, most ambiguous problems: a forward deployed designer, a technical counterpart, and a researcher. Three people, deployed against one problem, working like a startup inside the company. They produce working prototypes and validated direction. Three people need a shared problem and a whiteboard, and little else.

There is no PM. The shared problem is the roadmap, and the researcher holds the ground a product manager would. AI tools give the designer enough leverage to prototype at engineering speed, so the person who finds the problem can build the first version of the answer. Three people with these tools cover what used to take a ten-person cross-functional team, because the cost of exploring has collapsed.

Rotation matters as much as composition. Four to eight weeks, then disband, so the squad never hardens into a feature factory and designers keep moving through the hardest problems. Depth is the counterweight: people who stay in one problem space build context that rotation gives up. Handle the tradeoff deliberately.

The embedded posture

This is how most civic technology teams already work, in their own flavors, some more decentralized than others.

Design gets left out of this work because of how organizations frame the role. Hire designers for handoffs and no one invests in the headcount. Hire for what predicts forward-deployed performance: tolerance for ambiguity, collaboration, low ego, and a willingness to work in the open. The portfolio gauntlet that dominates design hiring rewards polished artifacts made in isolation, which is the wrong signal.

Hire designers who do more than one thing. Headcount is tight, and design has to justify itself more than most practices, so it cannot carry people too specialized to staff across engagements. In practice: designers who code, researchers with content and information-architecture chops, service designers who lead and PM, and every designer a strategist to some degree. Breadth is a structural requirement here. A specialist dropped into a six-week scoping sprint is useful for about one week of it.

Put designers where problems begin, while scope, framing, and constraints are still open. Most design organizations sit downstream, invited in once product and engineering have decided what to build and handed over a brief that already answers the questions design should ask.

Stand the practice up through adjacent skills. Hire UX people who can flex into service design; the work proves the value, demand grows, and the dedicated service designers follow. UX engineering grows the same way. Rotate people often, shaped by the project, the timeline, and what is in the pipeline that fits them, with the designer's own read in the mix.