Designing for complementary strengths

Begin with the work, then choose the roles

The collaboration report asks how people and computational tools can contribute different capabilities to a shared task. Its practical recommendation is to make that division of work intentional: state the purpose, choose the material, define useful contributions, and reserve time for interpretation.

The report uses complementarity in a specific research sense: a combined arrangement performing better than either component alone. That is something to demonstrate for a particular task and comparison, not a property to assume for every interaction. Here, the four patterns are reusable proposals for structuring work.

The Reflective Coach

Use this pattern for practicing an explanation, a language exchange, or a presentation. You define the setting and the skill you want to practice. An AI can take a role in a rehearsal and offer feedback against criteria you have chosen. You decide what to apply and repeat the exercise.

For an everyday demonstration, explain a hobby to an imagined newcomer. Ask the AI to pose three questions a beginner might ask. Answer those questions in your own words and revise the explanation. Compare the first and second versions for clarity, not for resemblance to the AI's style.

The Constraint Explorer

Use this pattern when a creative problem has several workable answers. Name the audience, materials, time, format, and qualities you want. Ask for genuinely different approaches within those boundaries. Choose an option and refine it.

The report illustrates constraint-led exploration through Autodesk and JPL's interplanetary lander concept. The supplier's project account describes engineers setting design requirements and using generative-design tools to explore alternatives. We use it here as a documented design project, not as a claim that the concept flew a mission. Project account.

An accessible application is a workshop poster: one page, two type sizes, three essential details, and a clear invitation. Compare arrangements, then make the final layout yourself.

The Judicial Challenger

This is the report's name for a pattern that begins with your independent draft. Ask for a constructive alternative reading, a question about your assumptions, and a suggestion for strengthening the evidence. Decide what deserves a revision.

In this edition, the pattern is used for ordinary writing, project design, and research interpretation. For example, offer a proposal for a community zine and ask which audience it already serves well and how another audience might use it. The result is a clearer proposal shaped by deliberate consideration.

The Transactive Synthesizer

The report uses transactive memory to ask who, or what, is responsible for which part of a group's knowledge. A synthesis workflow starts with a defined question and a set of materials. The AI helps organize or compare them. People check the material, supply context, and construct the final account.

Try it with three short notes you wrote yourself. Ask for a map of shared themes and distinct observations. Inspect the map beside the originals, select a connection worth developing, and write the interpretation in your own voice.

The Collaboration Canvas

This canvas comes from the report's proposed workflow. The interactive version builds a copyable prompt from the roles and task you enter. It runs locally as a template builder; it does not contact an AI service.

Close with a contribution record

Write down the human decisions, the AI contribution, and the result. “I chose the audience and examples; the AI proposed three structures; I combined and edited two of them” is more informative than “made with AI.” It helps a future collaborator understand where they can contribute to the next turn.