Gen AI and the Generation of New Work Organisations

The Impact of Gen AI

With generative AI (Gen AI), it is possible to renew an organisation extensively and fundamentally. The question ‘What does Gen AI mean for my organisation?’ is occupying many companies and institutions. Three topics in particular call for answers: (1) Can productivity be improved with Gen AI? (2) What is the effect of Gen AI on the quality of work? (3) How should the work process be adapted in order to optimise productivity and job quality together?

Artificial intelligence (AI) is the simulation of human abilities by a computer system, such as learning, reasoning, anticipating and planning, enabling it to adjust itself automatically without the assistance of human intelligence. A relatively new form of AI is generative AI (GenAI). GenAI systems can, with a considerable degree of autonomy, create new content such as text, programming code, images, music and videos. Well‑known examples include ChatGPT and Microsoft Copilot. GenAI holds significant promise for productivity improvement, but many organisations do not yet know how, or in what form, to realise that potential.

GenAI affects employment, occupations, job roles and tasks, although precisely how remains uncertain. Work that relies heavily on data, text, programming code and images may experience substantial influence from GenAI, while contact‑intensive occupations and manual skills may be less affected—particularly where automation offers fewer opportunities, such as hairdressing. Van der Torre and colleagues (2025), based on an analysis of several work processes in financial, legal and creative services, found positive effects on task variety, autonomy and functional support. The level of difficulty rises—creating challenge and development opportunities—but work intensification may also increase; a higher mental workload is to be expected. Social support tends to decline as technology takes over part of the communication, though this has only a limited impact on overall job quality. Crucially, the choices surrounding the introduction and use of AI technology, and the degree of influence employees have over these choices, determine job quality. The authors are cautiously positive, while warning against generalisation due to the small number of cases. Studies of algorithmic management—the steering, monitoring and assessment of workers using data—likewise argue that AI can have both positive and negative effects on job quality and labour relations, depending on managerial vision and organisational choices in applying the technology (Das et al., 2025; HRM‑collective Future of Work, 2025).

Beyond job quality, GenAI also has consequences for employment. It can perform tasks in programming, text editing, image processing and data analysis. It can prepare plans, administer and record. GenAI is increasingly capable of interactive tasks with customers, pupils and patients. Estimates differ on how many jobs are threatened and how many tasks and jobs will be created. Overall employment is expected to change only modestly, but most workers will certainly encounter GenAI in their work (OECD, 2023).

How to Assess What Gen AI Means for the Organisation

How can an organisation determine what GenAI—and digitalisation more broadly—means for its own context? First, (1) strategic choices must be made. Next, (2) possibilities for raising productivity with GenAI must be explored. At the same time, (3) the effects on employment and the quality of work should be assessed. Finally, (4) implementation should take account of redesigning organisational processes and job roles.

On strategy, Ben Allouch (2025) highlights four focal points when introducing AI. Because AI implementation effectively implies a transformation process, leaders should first consider how AI contributes to organisational goals, such as improving productivity. Second, leaders must ensure that the right people possess the necessary knowledge, qualifications and competences. Third, advance consideration is needed for the human–machine interaction and the task division between humans and AI, because those choices can either diminish or enlarge the human role, steer and monitor people or support and empower them. Organisations that rely on high‑skilled staff for future innovation capacity will view this differently from organisations that, facing labour shortages or strong cost pressure, seek opportunities to automate processes. Fourth, leadership should strengthen the organisational culture around technology adoption, overcoming resistance and enabling learning.

An exploration of possibilities to increase productivity using GenAI can be conducted with the Technology Impact Method (TIM). TIM is a TNO framework for determining the impact of technological possibilities (here, AI) on products and services, work processes and employees, and—additionally—where automation may be feasible if there are too few staff (Oeij et al., 2021; for an example see Hulsegge et al., 2022). A precondition is to involve stakeholders—the employees and, where possible, their representatives—from the outset (Van der Torre et al., 2021). This enlarges the chance of choosing technology solutions that suit both organisational and workforce needs. Management and employees jointly analyse issues where AI technology could help improve productivity. Possible technological solutions are inventoried, consulting both internal and external informants and experts (suppliers, trade fairs, etc.). Working together enables simultaneous assessment of the technology’s effects on productivity (for products, services and processes) and on job quality.

Assessing the impact of each technology option on employment and job quality requires attention. Productivity gains associated with GenAI may affect employment: will employees move to different tasks, will some leave, or will the workforce grow as production increases? GenAI may also change the task package per employee and entail qualification and training needs. How are tasks divided between people and AI technology? This can be assessed using the WEBA method (Well‑being at Work). WEBA is a framework for evaluating job quality (Vaas et al., 1995) before and after changes to a task package due to AI technology. It considers, among other aspects, job completeness, autonomy, the balance between regulatory discretion and health risks, and learning opportunities. When conducted jointly by management and employees, productivity goals and job quality can be balanced—creating a win–win with new AI technology.

Once choices on AI technology have been made, some redesign of organisational processes and job roles may be necessary. This arises where chosen human–machine interactions and task divisions between people and AI create new or altered roles. The WEBA method also provides guidance for redesign based on principles from sociotechnical organisational design (Oeij et al., 2024; Vaas et al., 1995). Using these principles, jobs can be designed to ensure good job quality within a productive work organisation. The following set of questions helps in practice:

In this job…

1. Is AI used in preparatory tasks via tools/equipment (e.g., working with software‑driven devices)?
2. In supporting tasks, is AI used in materials/sources (e.g., working with data)?
3. In executive and other tasks, is AI used through office software (such as Copilot, ChatGPT)?
4. In regulatory and executive tasks, is AI used in job/assignment information (such as provided work       orders and instructions)?

If so (per subtopic)…

1. To what extent does the operator influence the application of AI?
2. To what extent is the impact of AI on job autonomy known—regarding pace, method, order and workload in work execution?
3. To what extent is it known how the AI works (its algorithm)?
4. Is AI necessary to perform the work (autonomously)?

Redesigning processes and roles yields insights not only on job‑quality requirements but also on training and education needs, and on the supply and demand of (new) employees—i.e., employment.

How to Shape Organisational Choices about Gen AI

The work organisation is the arrangement of operational employees and managers to realise organisational goals. This entails a division of labour in which tasks must be grouped and aligned within a chosen organisational structure—such as bureaucratic, flow‑based, functional, project‑based organisations or task‑oriented platforms (operating online or otherwise). The ultimate form of a work organisation depends on how production or service processes are configured and how labour is distributed among people, and between people and technology such as GenAI.

To ‘generate’ new work organisations, designers require instruments:

• For strategic choices on applying AI: various canvas methods (e.g., the AI Strategy Canvas, Introducing the AI Strategy Canvas – Amsterdam Data Academy).
• To explore effects of AI on job quality and employment: TIM.
• To determine job quality: the WEBA method.
• For redesign solutions of organisation and roles: the WEBA method, supplemented with a module on AI aspects of work.

AI Strategy Canvas

Following the success of the ‘canvas’ approach to business models (Osterwalder, 2004), various AI‑canvas approaches have appeared, such as the ‘AI Strategy Canvas’ by the Amsterdam Data Academy (Figure 1). These rightly assert that AI is not only a technological challenge but also a strategic one. The AI Strategy Canvas supports strategic thinking and action on AI and comprises nine blocks across three zones: left column (inputs, enablers), middle column (core strategy, risks), and right column (stakeholders, adoption, impact). Answering the questions (cf. the right‑hand prompts) sharpens what AI should deliver and what is needed to realise it.

Figure 1. AI Strategy Canvas

Exploring the Effects of AI on Job Quality and Employment

The central idea of TIM (Figure 2) is that technology has consequences—or offers possibilities—for new or improved products and services; and that this in turn affects work organisation, workflow design and human capital (Oeij et al., 2021). In terms of human capital, think of impacts on task packages, required competences and working conditions. Technology is not deterministic: the selection of GenAI and other technologies, and the manner of application in the organisation, is a human decision. Current policy trends emphasise that technology should support and empower people in their work, rather than steer and control them through algorithmic management. This ‘human‑centricity’ improves job quality and aligns with the EU AI Act and privacy legislation (GDPR) (EC, 2025; Oeij et al., 2025). From TNO we therefore recommend human‑centric innovation (Tjin‑A‑Tsoi, 2025a, 2025b).

Figure 2. Technologie Impact Methode (TIM)

In the ‘technology’ block (Figure 2), the focus is on the GenAI variants inventoried from step 1 on strategy. The ‘product/service’ block concerns desired improvements to products, services and production means in the work process using GenAI. The ‘organisation of work’ block focuses—especially with GenAI—on the division of tasks between people and AI, and on grouping tasks within roles and, where applicable, teams and departments. Finally, the ‘human capital’ block covers the quantitative and qualitative aspects of staffing, such as the number of employees and the demand for qualifications and competences arising from GenAI. The outcome of this joint dialogue between management and involved employees is a preference for a GenAI technology applied to products, services or processes, with insights into work organisation and personnel consequences.

Determining Job Quality (WEBA)

Using the WEBA method, job quality can be assessed for roles by evaluating seven characteristic conditions of work: job completeness (coherent task package); absence of short‑cycle tasks (repetitive and monotonous work); level of difficulty (task variety); autonomy (discretion); opportunities for contact (social support); organising tasks (functional support); and information provision (unambiguous instructions). Be aware that this Lowlands Sociotechnical approach of job quality originated from designing the operational organisation layout instead of a psychological job design approach and looks at some different dimensions (Warhurst et al., 2017). A qualitative expert profile for the example role of ‘order picker’ (Figure 3) is based on interviews and workplace observations that examine whether disruptions occur in the work and the extent to which the role‑holder can resolve them (Vaas, 2023); it also considers workload risks and learning opportunities. Where disruptions prove ‘unresolvable’, various issues may exist: too little discretion, high work‑intensity/stress risks, too few learning opportunities.

Figure 3: Profile of a role using the WEBA method

Redesign Solutions for Organisation and Roles

For identifying and formulating redesign solutions, the WEBA method is also available, supplemented with the module on AI aspects of work (see the question block above). Redesign of organisational processes and roles builds on sociotechnical thinking (Kuipers et al., 2018). The existing configuration of the production or service process is analysed to locate ‘unresolvable’ disruptions in the work of a role. One then examines whether the disruption can be resolved by re‑allocating tasks—between organisational levels, between job packages, and between roles performed by people and by systems—here, in applications involving GenAI. As illustrated by the example in Figure 3, this can involve adding tasks that increase autonomy (decision authority), organising tasks (functional support) and job completeness (more variety). A recurrent regulation problem for order pickers is that GenAI prescribes work instructions from which the worker cannot deviate, potentially causing high work intensification and a lack of autonomy. The regulation loop (Figure 4) helps reveal how work execution proceeds within a role. The bold dashed line separates the management structure (above) from the production structure (below) and indicates the regulatory options in the role: regulation sits above the dashed line. A handy rule of thumb is that adding tasks above the dashed line will generally enlarge discretion (autonomy). In this example, being able to create work instructions oneself using the planning software could help compensate for intensification and lack of autonomy.

Figure 4. Regulation loop (Control circle)

The approach is to perform a role analysis of the existing situation and of the future situation with GenAI, by drafting role descriptions for both and jointly assessing them, then comparing the assessments for (possible) disruptions (regulation problems). The role of GenAI in particular must be evaluated. This includes forming estimates regarding intended productivity improvements, potential effects on task execution (shorter time, higher output) and effects on role fulfilment (disruptions, workload, learning opportunities).

Concluding Remarks

The four steps outlined provide insights into:

• strategic choices concerning GenAI technology;
• the consequences of GenAI for the organisation and the work;
• the consequences for job quality and potential regulation problems;
• solution directions for redesigning organisation and roles.

This exercise yields insight into whether the GenAI application offers more advantages than disadvantages in generating a work organisation with GenAI. That insight will likely be an estimation rather than a precise calculation to the decimal place—that is acceptable, because taking a reasoned risk is part and parcel of innovative entrepreneurship, including the feasibility and generation of new work organisations.

Sources

Ben Allouch, S. (2025). Leidinggeven aan AI-implementatie: de onontkoombare transformatie. Interview in: J. Gaspersz & P.R.A. Oeij (red.), Leidinggeven aan innovatie. Kompas voor vernieuwingskracht (pp. 260-271). Zeist: VMN media.

Das, D., Van der Torre, W., Van Dam, L., Zoomer, T., & Preenen, P. (2025). Wat is de impact van algoritmisch management op de kwaliteit van arbeid in reguliere organisaties? Tijdschrift voor HRM, 28(1), 84-107. https://doi.org/10.5117/THRM2025.1.004.DAS

EC (2025). Quality Jobs Roadmap. Communication From The Commission 4.12.2025 COM(2025) 944 final. Brussels: European Commission.

HRM-collectief Future of work (2025). Human-in-control. Tijdschrift voor Arbeidsvraagstukken, 41(4), 432-444.

Hulsegge, G., Van der Torre, W., Verbiest, S., & Oeij, P. (2022). De impact van technologie op de taken, skills en kwaliteit van de arbeid: Hoe verandert het werk van een elektro- en onderhoudsmonteur? Tijdschrift voor Arbeidsvraagstukken, 38(2), 169-191. https://doi.org/10.5117/TVA2022.2.004.HULS

Kuipers, H., Amelsvoort, P. van, & Kramer, E.-H. (2018). Het nieuwe organiseren. Alternatieven voor de bureaucratie. Leuven: Acco. (2e ed.; 1e ed. 2010).

OECD (2023). OECD Employment Outlook 2023: Artificial intelligence and the labour market. Paris: Organisation for Economic Co-operation and Development. https://doi.org/10.1787/08785bba-en

Oeij, P.R.A., Hulsegge, G., van der Torre, W. (2021). Technology Impact Methode 3.0: Een kwalitatieve aanpak over de impact van technologie op werk. Leiden: TNO.

Oeij, P., Pot, F. & Hulsegge, G. (2025). Om de kwaliteit van het werk. S&D Socialisme en Democratie, 82(5), 54-63, Themanummer ‘Arbeid in de nieuwe rood-groene partij’, https://wbs.nl/publicaties/om-de-kwaliteit-van-het-werk

Oeij, P., Vaas, F. & Dhondt, S. (2024). Kun je de HR-professional in gesprek brengen met de organisatieontwerper? Het verbinden van de SMART-werkontwerpbenadering met sociotechnische ontwerpprincipes. Tijdschrift voor HRM, 27(4), 66-93. Doi: 10.5117/THRM2024.4.014.OEIJ

Osterwalder, A. (2004). The Business Model Ontology: A Proposition In A Design Science Approach (Ph.D. thesis). Lausanne: University of Lausanne.

Tjin-A-Tsoi, T. (2025a). Voorwoord ‘Leiders moeten scherp keuzes maken’. In: J. Gaspersz en P.R.A. Oeij (red.), Leidinggeven aan innovatie: Kompas voor vernieuwingskracht (p. 7). Zeist: VMN media.

Tjin-A-Tsoi, T. (2025b). Versnellen van innovatie in het mkb is urgent. In: Job Teurlinx, Anna van der Meij, & Chris Sol (red.), Jaarbericht Staat van het mkb 2025 (pp. 45-50). Den Haag: Nederlands Comité voor Ondernemerschap.

Vaas, F. (2023). Kwalitatieve benadering van KCW: functie-analyse. In: P. Oeij (red.), Kortcyclisch werk. Quo vadis? (pp. 38-43; 95-102). TNO en RIVM i.o.v. Instituut GAK. Leiden: TNO Innovation for Life.

Vaas, S., Dhondt, S., Peeters, M. H. H., & Middendorp, J. (1995) De WEBA-methode: deel 1 WEBA-analyse handleiding. Alphen a/d Rijn: Samsom.

Van der Torre, W., de Geit, E., van Oosteren, C., Yagafarova, A. & Mathijssen, S. (2025), De impact van generatieve AI op de kwaliteit van de arbeid. Tijdschrift voor Arbeidsvraagstukken, 41 (4), 466-488; DOI https://doi.org/10.5117/TVA2025.4.004.TORR

Van der Torre, W., Oeij, P., & Davits, R. (2021). Sociale en technologische innovatie bij gemeenten: Effectieve en gedragen innovaties door betrokkenheid van medewerkers. Den Haag: Stichting A&O fonds Gemeenten.

Warhurst, C., Wright, S., Lyonette, C. (2017). Understanding and measuring job quality. Research report Part 1-Thematic Literature Review. London: CIPD, Chartered Institute of Personnel and Development.

Peter Oeij Senior Research Scientist and Consultant

I am a professional in the field of Research & Consultancy on work, organisation & management issues. My present challenges are innovation management, social innovation and workplace innovation. I am exploring: Effects of Digital Transformation and new technology on organisation, required skills and team dynamics. I am driven by the question why innovations too often fail, especially, what are the antecedents and determining factors of failure, risk avoidance and defensive behaviour between people, and in teams. My focus of interest thus is on behavioural and organisational aspects. I am a professional in the field of Research & Consultancy on work, organisation & management issues.

See Full Bio