Understand, measure and develop the human side of AI at work

We work on the human and organisational side of AI at work through four areas: Assess, Validate, Design and Develop. Assess examines workforce readiness, skills and organisational context. Validate reviews the evidence behind AI-enabled assessments and other systems that make claims about people. Design addresses how roles, workflows and human capabilities need to change as AI becomes part of the work. Develop builds the skills, behaviours and leadership practices needed to work effectively with AI. Each area can be delivered independently, or several can be combined within one framework.

A typical survey may produce a single self-rated confidence score. Our approach separates the factors that matter to the defined question, such as skills, attitudes, behaviour and workplace context. Measures are linked to defined constructs and evaluated for their measurement quality. Where appropriate, results can be compared across teams, functions or relevant benchmarks. The outcome is not only a profile of current readiness. We translate the findings into priorities, actions, ownership and timescales.

We may combine psychometric data, organisational and workforce data, role analysis, stakeholder evidence, behavioural science and relevant professional frameworks. We distinguish between measured evidence, stakeholder input and professional judgement, and make the basis for recommendations clear. Where evidence is limited, we identify the gap rather than treating an assumption as a finding.

Yes. We can work with existing competency frameworks, job architectures, employee surveys, psychometric assessments, workforce data and AI-readiness models. We first assess what is already available, identify gaps and determine what can be retained or adapted. This allows the work to build on existing evidence and organisational practice rather than introducing a new framework without need.

The work is led by an organisational psychologist and qualified test user with professional training in psychological assessment, organisational psychology, behavioural science and people development. Relevant qualifications and professional training include BPS Test User qualifications in Occupational Ability and Personality (Levels A and B), completed through the University of Cambridge Psychometrics Centre; an MSc in the Psychology of Human Resources and Organisational Health; EQ-i 2.0 and EQ 360 certification; training in Rational Emotive Behaviour Therapy through the Albert Ellis Institute; and the Level 2 Certificate in Counselling Skills and Level 3 Certificate in Counselling Studies awarded by CPCAB. The work combines psychological measurement with organisational and workforce analysis rather than treating AI adoption as a technology-only issue.

Yes. We can independently review an existing AI-enabled assessment and evaluate the evidence supporting its intended use. Depending on the assessment and available data, the review can cover the construct definition, content, response process, internal structure, reliability, relationships with other measures, group differences and fairness, scoring processes, and the evidence supporting the intended decisions. Where AI is involved in scoring or interpretation, we can also examine the relationship between human and automated scoring, model behaviour, transparency and relevant sources of measurement error. The output can include a validation report, supporting evidence and recommendations for further analysis or remediation. The appropriate review cycle depends on the assessment, its use, the evidence available and any material changes to the model or underlying data.

A single engagement typically takes 3–6 weeks, depending on scope, data availability and the level of analysis required. Projects combining several areas will usually take longer. Where measurement is part of the work, we can establish a follow-up cycle, typically including remeasurement after 3–6 months. The timeline and milestones are agreed before the engagement begins.

The outputs depend on the scope. Readiness work can produce a workforce profile, priority areas and action plan. Validation work can produce an independent review, evidence summary and recommendations. Role and competency work can produce competency requirements and role pathways. Work redesign can produce redesigned workflows, responsibilities and capability requirements. Development programmes include facilitated sessions, learning materials and outcome measurement. Where appropriate, every engagement finishes with clear priorities, ownership and next steps.

This depends on the project. We may need role and workforce information, existing competency frameworks, assessment materials, survey or assessment data, organisational documents and access to relevant stakeholders. We agree the required inputs at the start and work with the information available.

Both. Assessments and data collection are normally completed online. Consultancy, interviews and review meetings can be delivered remotely or in person. Training and workshops can also be delivered in either format. The delivery model is agreed around the project, participants and location.

Our approach is that AI outputs concerning people should have clearly defined decision boundaries and appropriate human oversight. For people-related decisions, we identify where AI can support analysis, where human review is required and who remains responsible for the final decision. The appropriate level of oversight depends on the purpose, consequences of the decision, quality of the evidence and the way the system is being used. Our work can therefore define decision boundaries, review points, escalation routes and accountability rather than treating AI output as a decision in itself.

AI changes how work is performed, not simply which tools employees use. We analyse tasks, workflows, decisions and responsibilities to determine where AI can automate, support or change the work, and where human judgement remains important. We then translate those changes into role requirements, workflows, capability needs and workforce implications. This helps organisations consider the work and workforce together rather than adding AI to an unchanged job design.

Yes. We work with employers that need to understand and prepare their workforce for AI, and with organisations developing AI-enabled products used in hiring, assessment, profiling, coaching or people analytics. For vendors, our work can examine the psychological constructs, measurement evidence, scoring approach, fairness and intended use of a product. For employers, we focus on workforce readiness, role change, adoption and human capability.

We establish the relevant measures at the beginning and, where appropriate, repeat them after implementation. Readiness and capability programmes can be remeasured after 3–6 months. Development programmes can include measurement before and after delivery, with a further follow-up around 90 days later. The measures depend on the objective of the engagement. We aim to assess changes in skills, behaviour, readiness or organisational conditions rather than relying on satisfaction alone.

Before work begins, we agree what information is required, why it is being collected, who will have access to it and how it will be used in reporting. Data is handled in accordance with applicable UK data protection requirements. Access is limited according to the agreed purpose, and reporting is structured to protect confidentiality where individual-level information is not required. The specific data handling arrangements for each engagement are confirmed before data is collected. Further information is available on our data privacy policies page.

Pricing depends on the scope of the engagement, the areas involved, the number of people or roles covered, the data and assessments required, and the level of consultancy support. We define the scope with you first and provide a customised proposal following an initial discussion.