Published research selected for executive practice.

Editorial summary

Gosline distinguishes friction that obstructs people from friction that protects their choices, understanding, and autonomy. She encourages organizations to examine where AI-supported experiences need deliberate checks and greater transparency.

ClarityPlan application

Review one recurring process. Identify steps that create avoidable effort and steps that protect accuracy, informed choice, privacy, or access to human assistance. Test changes against both efficiency and the experience of the people served.

Editorial summary

Summarizes advice from Robert I. Sutton and Huggy Rao on subtracting unnecessary work, valuing people's time, using clear language, and preserving useful friction. Staff as well as senior leaders can identify obstacles.

ClarityPlan application

Use this as a reading prompt before examining a recurring approval or reporting process. Ask which steps serve a necessary purpose and which deserve investigation.

Editorial summary

Drawing on nearly 40 interviews with human-services nonprofit leaders, this brief examines duplicative contracting, reporting, reimbursement, and audit requirements. It proposes ways to reduce unnecessary administrative work while preserving accountability.

ClarityPlan application

Map a recurring reporting or approval process. Identify duplicated submissions, unclear requirements, and time diverted from service delivery; ask which changes would reduce effort without weakening necessary oversight.

Editorial summary

This article explains a report from MIT’s Working Group on Generative AI and the Work of the Future. It connects deployment choices and changes in work with the question of whether jobs improve.

ClarityPlan application

Before choosing software, ask how responsibilities, learning, review, and work allocation will change. Use the article to frame a discussion with staff about the process around the tool.

Editorial summary

The authors organize AI choices around augmenting organizational capacity, advancing mission, and advocating for responsible development and use. The framework encourages mission-related choices rather than undifferentiated adoption.

ClarityPlan application

Ask which organizational priority the proposed AI use serves and whether the organization can support it. Read alongside cross-sector studies rather than treating nonprofit affiliation as the only selection criterion.

Editorial summary

Summarizes research presented at MIT’s Initiative on the Digital Economy conference on AI design, prompting, workflows, human judgment, and jobs. It highlights ongoing research on asking people to explain why they follow AI recommendations rather than accepting them uncritically.

ClarityPlan application

Before acting on an AI recommendation, document the reasoning and check important claims independently. Examine how prompts frame the advice and where deliberate review belongs in the whole workflow.

Editorial summary

Explains generative AI, model selection, reliability, testing, and data management. The guidance warns that human-like language does not demonstrate understanding and describes evaluating outputs against expected answers and relevant risks.

ClarityPlan application

Define a proposed task and its expected result, test representative cases, and decide which outputs require independent checking. Review privacy and organizational policy before sharing information with an AI service.

Editorial summary

Employees may override effective AI systems when they remain personally accountable for outcomes but lack clear organizational backing to follow the system. Narayandas argues that leaders must align decision authority, accountability, and employee roles as AI changes how work gets done.

ClarityPlan application

Before introducing AI, define which tasks and decisions it may perform, when employees should intervene, who explains outcomes, and how leaders will address errors when approved procedures are followed. Examine whether existing review requirements create unnecessary duplicate work.

Editorial summary

Explains Shakked Noy and Whitney Zhang’s experiment with 453 professionals completing writing tasks. The study measured completion time and quality assessed by evaluators unaware of AI use; the article also explains why factual checking and organizational context matter when applying the findings.

ClarityPlan application

Compare a small assisted workflow with the existing method using agreed quality criteria as well as completion time. Include fact-checking, review effort, and rework rather than treating a faster draft as a finished improvement.

Editorial summary

The article explains research that treats nonprofit capacity as several distinct capacities. It describes an instrument developed and tested across organizations rather than equating capacity with one general impression.

ClarityPlan application

Identify which capacity a proposed change should affect, such as operational capacity or staff management. Do not claim that ClarityPlan’s own reflection tool inherits the instrument’s validation.