THE EVIDENCE BEHIND THE PRACTICE
Research connects psychological safety with learning and team effectiveness. Workplace health and readiness research widens the view to the conditions that help people carry change.
The programme combines established research with Trace’s practitioner-development experience. The application to forward-deployed engineers is a new proposition to evaluate in live work.
RAND · 2024
Interviews with 65 experienced AI practitioners identified misunderstood objectives, inadequate data, technology-led priorities and other recurring reasons projects fail.
Qualitative findings inform discovery and sponsor conversations. The interview sample does not establish a population-wide failure percentage.
RAND · AI project interviews, 2024BRYNJOLFSSON, LI & RAYMOND · QJE 2025
The study followed 5,172 customer-support agents at one company. AI assistance increased issues resolved per hour by 15% on average; gains differed by skill and experience, with small quality declines among the most experienced and skilled agents.
This supports examining who benefits and how work changes. It is not evidence that Fieldcraft produces a productivity gain. The linked author manuscript was revised in November 2024.
Brynjolfsson, Li & Raymond · Author manuscript; QJE, 2025GOOGLE · PROJECT ARISTOTLE
Google examined 180 teams and identified five important dynamics, with psychological safety ranked first. Dependability, structure and clarity, meaning, and impact also mattered.
Fieldcraft application: develop candid participation alongside clear responsibilities and follow-through. These findings concern teams at Google; they do not prove that any single condition guarantees transformation.
Read Google’s original accountAMY EDMONDSON · 1999
In a field study of 51 manufacturing teams, psychological safety was associated with learning behaviour. The tested model connected that learning behaviour with team performance.
Fieldcraft application: practise asking for help, discussing mistakes, checking assumptions, and seeking feedback in consequential conversations.
Read the original studyFRAZIER AND COLLEAGUES · 2017
A meta-analysis examined 136 independent samples, representing more than 22,000 individuals and nearly 5,000 groups, to investigate the antecedents and outcomes of psychological safety.
Fieldcraft application: draw on the broader literature while evaluating this programme in its own setting. These samples are not trials of Fieldcraft.
Read the meta-analysis summaryBUVIK & TKALICH · 2021 MANUSCRIPT / HICSS 2022
A survey of 236 members of 43 Norwegian software teams linked autonomy with psychological safety, and psychological safety with team reflection and performance.
Fieldcraft application: include the design of work and opportunities for team reflection. Internal agile teams differ from engineers embedded in client organizations; the survey does not establish causation.
Read the authors’ paperMENTAL HEALTH COMMISSION OF CANADA · WORKPLACE STANDARD
Canada’s voluntary workplace framework addresses the promotion of mental health and prevention of psychological harm through a systematic organizational approach.
Fieldcraft application: connect the quality of conversations with the working arrangements people are being asked to carry. The programme draws on these principles; it is not a standards certification.
Explore the workplace frameworkWORLD HEALTH ORGANIZATION · MENTAL HEALTH AT WORK
WHO identifies risks including excessive workload, low control, unclear roles, and job insecurity. Its guidance addresses organizational conditions alongside training, with meaningful worker participation.
Fieldcraft application: follow where implementation moves demands and authority. Involve affected people and work with responsible leaders on a viable response.
Read WHO’s guidanceBRYAN WEINER · 2009
Weiner’s theory describes shared commitment to a change and confidence in the organization’s collective ability to implement it, informed by demands, resources, and circumstances.
Fieldcraft application: examine willingness and capability in relation to the specific work ahead. Readiness is not a fixed label assigned to an employee.
Read the readiness theoryO’DONOVAN & MCAULIFFE · 2020
A review of 14 healthcare interventions found mixed outcomes for psychological safety, speaking up, and voice. The authors recommended sustained, multifaceted approaches with participant involvement and visible leadership support.
Fieldcraft application: combine rehearsal with field application, sponsor action, and follow-up. Evidence from healthcare informs the design without proving its effects in AI deployment.
Read the intervention reviewCan participants demonstrate the capabilities? Can affected people raise concerns and see an appropriate response? Did the working arrangement change? Is there meaningful delivery value?
We examine those questions separately. An engineer’s assessment score cannot tell us whether another person feels safe, and a quicker deployment does not establish a healthier workplace.
The proposed pilot uses observable practice, voluntary feedback, operational decisions, and agreed delivery measures. Its size allows evidence of contribution, rather than a claim that training alone caused every improvement.
See how the pilot worksThe fictional field encounter illustrates the teaching approach. No existing AI-deployment client results or financial returns are claimed.