Pith. sign in

REVIEW 4 major objections 5 minor 73 references

Is PMBOK Guide the Right Fit for AI? Re-evaluating Project Management in the Face of Artificial Intelligence Projects

T0 review · 4 major / 5 minor · reviewed 2026-08-07 · deepseek-v4-flash

Pith's one-line read The paper argues that the PMBOK Guide can fit AI software projects if its eight performance domains are tailored to five AI-specific features, from data dependency to ethics.

desk verdict A useful practice-oriented PMBOK-AI tailoring that overstates what its small survey can prove; still worth refereeing. read the letter →

arxiv 2506.02214 v1 pith:EJXVFRVH submitted 2025-06-02 cs.SE cs.CV

classification cs.SEcs.CV
keywords PMBOKGuideAIprojectmanagementdatadependencyuncertaintyandexperimentationiterativedevelopmentspecializedexpertiseethicalconsiderationstailoring
verification ladder T0 review T1 audit T2 compute T3 formal

The pith

A machine-rendered reading of the paper's core claim, the machinery that carries it, and where it could break.

The reading

The paper evaluates the widely used PMBOK Guide as a management framework for AI software projects and concludes that its principles often miss what makes AI work different. The asserted problem is concrete: the Guide gives little guidance on data as a managed project asset, treats uncertainty as something to plan away rather than experiment through, offers no lifecycle for iterative model development, says little about multidisciplinary expert teams, and does not embed ethics into project delivery. The proposed fix is also concrete: map the five AI project features onto the Guide's eight performance domains and add tailoring recommendations, including data lifecycle management, hybrid agile-plus-experimental lifecycles, and ethics and fairness measurement. If this is right, organizations using PMBOK do not need to replace it for AI projects; they need to supplement it in a structured way.

What carries the argument

The machinery is a feature-to-domain mapping. The five AI project features (PF1 data dependency, PF2 uncertainty and experimentation, PF3 iterative development, PF4 specialized expertise, PF5 ethical considerations) are paired, one by one, with the PMBOK Guide's eight performance domains—stakeholder, team, development approach and life cycle, planning, project work, delivery, measurement, and uncertainty—producing tables of tailored recommendations. The mapping does the argumentative work because it converts the broad claim that PMBOK does not fit AI into a checkable list of specific gaps, each with a named domain where the Guide is silent and a concrete practice that fills the silence.

What would settle it

A field study that found a substantial class of successful AI projects whose critical success factors fall outside PF1-PF5, or where PMBOK guidance already covered the reported practices, would undercut the gap analysis. More directly, a controlled comparison in which AI teams using the tailored PMBOK guidance show no improvement in schedule adherence, model quality, or stakeholder satisfaction over teams using untailored PMBOK would falsify the claim that the tailoring closes real gaps.

Watch

Extended reading notes

Core claim

The central claim is that the PMBOK Guide's principles often fail to address the unique complexities of AI projects, yet the Guide remains usable as a foundation once tailored. The paper supports this by identifying five AI project features—data dependency, uncertainty and experimentation, iterative development, specialized expertise, and ethical considerations—and analyzing each against the Guide's performance domains. It reports an expert survey of 39 ratings across six AI software projects that strongly supports the hypothesized feature set, with data dependency, uncertainty and experimentation, and specialized expertise ranked highest and iterative development ranked lower because some respondents found Agile workable. The output is five tailoring tables recommending concrete practices within each performance domain, such as privacy impact assessments, MVP-based experimentation, hybrid lifecycles for mixed software-and-model teams, and bias and fairness audits.

Load-bearing premise

The load-bearing premise is that the five AI project features are the essential distinguishing characteristics of AI software projects and that the expert survey of 39 data points from six projects is representative enough to confirm them.

Editorial extensions

If this is right

  • AI project teams that follow the tailoring tables keep PMBOK as their management backbone while adding data lifecycle stages for sourcing, cleaning, validation, and licensing into planning and delivery.
  • Hybrid lifecycles become the default: conventional software parts run on agile sprints, while model development runs on experiment-driven iterations that may span several sprints or change direction.
  • Ethics becomes a measurable project function, with bias audits, fairness metrics, social-impact assessments, and post-deployment monitoring attached to the measurement and delivery domains.
  • Stakeholder and team practices expand to include education about model limitations, shared vocabulary between data scientists and non-technical members, and explicit roles for ethicists and legal experts.
  • The Guide's tailoring guidance is judged sufficient to absorb these additions without rewriting its core standard.

Reading between the lines

Editorial extensions of the paper, not claims the author makes directly.

  • A natural extension is that the same gap analysis would likely apply to other general-purpose project-management standards, since the five AI features are not PMBOK-specific.
  • If the five-feature set is accepted, a direct next test is to compare project outcomes under tailored PMBOK against outcomes under AI-native management approaches, measuring schedule accuracy, model quality, and stakeholder satisfaction.
  • The lower expert ranking of iterative development suggests a sharper boundary worth investigating: which AI activities are genuinely agile-compatible, and which require open-ended experimentation.
  • A larger, more diverse sample of AI projects could check whether additional features, such as regulatory compliance or infrastructure cost, deserve their own performance-domain treatment.
Share X Bluesky LinkedIn Reddit HN

Editorial analysis

A structured set of objections, weighed in public.

Desk editor's note, referee report, and a circularity audit.

Referee Report

4 major / 5 minor

Summary. The paper evaluates the applicability of the PMBOK Guide (7th edition) to AI software projects. It hypothesizes five distinguishing features of AI projects—data dependency, uncertainty and experimentation, iterative development, specialized expertise, and ethical considerations—and reports expert ratings from 39 data points across six AI projects (Figure 1). The paper then maps each feature against the PMBOK Guide's eight performance domains and provides tailoring recommendations in Tables 1–5, arguing that the PMBOK Guide's principles "often fail to address the unique complexities of AI projects" and should be supplemented with data lifecycle management, iterative/experimental lifecycles, multidisciplinary team practices, and ethics integration. The paper concludes that such tailoring can keep the PMBOK Guide relevant for AI-driven initiatives.

Significance. If its central claim were established, the paper would offer a useful practical contribution: it addresses a genuine gap in a widely used project management standard and translates identified gaps into actionable, domain-structured recommendations. The mapping to the eight performance domains is transparent, and the paper draws on a broad citation base, including recent empirical studies of AI project management. The tailoring tables are concrete enough for practitioners to adopt or critique. The main weakness is that the empirical validation is too thin to support the strength of the central claim; as it stands, the paper is best read as a hypothesis-generating gap analysis rather than an evidence-based verdict on PMBOK's suitability for AI projects.

major comments (4)
  1. [Section II, Figure 1] The survey asks experts only to rate the importance of PF1–PF5 for AI projects on a 1–10 scale; it never asks them to compare AI projects with traditional software projects. Importance is not distinctiveness, so Figure 1 cannot support the Conclusions' claim that PMBOK's principles "often fail to address the unique complexities of AI projects." To support the uniqueness premise, the authors need a matched comparison—for example, having the same experts rate the same features for a set of non-AI software projects, or explicitly rating how much more pronounced each feature is in AI projects. This is load-bearing because the gap analysis in Sections IV.A–IV.E identifies gaps as AI-specific on the basis of PF1–PF5.
  2. [Section IV, Tables 1–5] The gap analysis asserts, for each performance domain, that PMBOK lacks a practice needed for AI projects, but it does not compare PMBOK's handling of these features for non-AI projects. For instance, data dependency (PF1) and specialized expertise (PF4) are salient in many traditional data-heavy software projects; without a baseline, the recommendations cannot be distinguished from generic advice for complex software projects. The authors should either provide a comparative analysis of how PMBOK supports analogous features in non-AI projects or explicitly weaken the conclusion from "unique to AI" to "particularly pronounced in AI." This issue affects every tailoring table and the paper's overall claim of AI-specificity.
  3. [Section II, AISP1–AISP6] The empirical basis is 39 survey points from six projects, all of which appear to be Intel-associated open-source projects. The paper reports no descriptive statistics beyond what is shown in Figure 1, no error bars, no significance tests, and no details on how experts were selected or how many experts per project were surveyed. The case studies are named as illustrative examples but are not presented through a systematic case-study protocol. This is sufficient to motivate hypotheses but not to "strongly support" them, as the paper claims in Section II. Please report the survey instrument, sampling frame, per-project respondent counts, and treat the results as exploratory rather than confirmatory.
  4. [Section IV, Recommendations] The tailoring recommendations are not validated: no evidence is presented that following Tables 1–5 improves project outcomes, reduces failures, or addresses the identified gaps in practice. A controlled experiment is not required, but at minimum a retrospective comparison of projects with and without the recommended practices, or a structured expert elicitation on the recommendations' expected impact, would support the paper's claim that "the PMBOK Guide can better meet the needs of AI-driven initiatives." Without this, the recommendations remain plausible but unsubstantiated.
minor comments (5)
  1. [Section IV.A] The text contains a typo: "AI project projects" should be "AI projects."
  2. [Section IV.B] The subsection heading reads "B. 4.2. Uncertainty and Experimentation (PF2)" with a stray "4.2" numbering artifact; please clean up the heading numbering.
  3. [Figure 1] Figure 1 is referenced as showing expert rankings, but the figure is not clearly visible in the manuscript; please ensure a high-resolution image and consider providing the numeric summary (e.g., mean and range per feature) in the text or a table.
  4. [Section II] Reference [5] is cited as a source on experimentation in AI projects, but [5] is the PMBOK Guide itself, which is not a natural source for that claim; please re-check the citation or replace it with an appropriate empirical reference.
  5. [Section I] The literature-selection procedure is not reproducible: "Generative AI helpers" is not a defined search method. Please specify the databases, search strings, and inclusion/exclusion criteria used for the literature review.

Circularity Check

0 steps flagged · score 0.0 of 10

No significant circularity: the paper's argument is a literature- and survey-based gap analysis with no derivation that reduces to its own inputs.

full rationale

The paper contains no formal derivation, no fitted parameters, and no prediction that is defined in terms of its own output. Its chain is empirical and taxonomic: Section II hypothesizes five AI-project features (PF1-PF5) based on the authors' experience and external literature, then tests them with an expert survey that rates the importance of each feature. That survey is independent evidence, not an assumption of the conclusion. Section IV compares PF1-PF5 to the PMBOK performance domains and maps recommendations onto those domains; this is a classification/expert-judgment exercise, not a self-referential reduction. The authors' employer's Intel open-source projects are used as illustrative case studies (e.g., [15], [16]), but they are not the sole or load-bearing authority: the same claims are supported by external citations, and the case studies serve as examples rather than as premises that already contain the conclusion. The skeptical concern that importance ratings do not establish that PF1-PF5 are unique to AI is a real internal-validity threat, but it is not circular reasoning: the conclusion could be false or under-supported without the argument being self-referential. Therefore no circular step can be quoted or exhibited, and the appropriate finding is no significant circularity.

Assumptions & free parameters 0 free parameters · 4 assumptions · 0 invented entities

No free parameters or invented entities. The paper's burden is carried by domain assumptions about the distinctiveness and representativeness of AI projects and the choice of PMBOK as the lens; these are plausible but not independently validated.

assumptions (4)
  • domain assumption AI software projects are fundamentally different from traditional software projects along the five features PF1 to PF5.
    Section II states this as a hypothesis and the entire gap analysis in Section IV is organized around it. If the features are not essential or exhaustive, the recommendations are misdirected.
  • domain assumption The expert survey results are representative of AI project management priorities.
    Section II uses Figure 1 to support the feature hypothesis, but the sample is 39 data points from 6 projects with no disclosed respondent selection or diversity, so representativeness is an assumption.
  • domain assumption The PMBOK Guide 7th edition is the appropriate foundational framework for evaluating AI project management.
    Section III introduces PMBOK as the standard to be evaluated; the paper assumes its eight performance domains are the right lens for comparison.
  • domain assumption The selected Intel open-source projects provide generalizable insights into AI project management.
    Section I lists Geti, OpenVINO, Training Extensions, Datumaro, Anomalib and XAI Toolkit; all are tied to the authors' employer, and no evidence is given that they represent the broader AI project landscape.

how reviews work

0 comments
Cite this review

Pith. "Pith review of Is PMBOK Guide the Right Fit for AI? Re-evaluating Project Management in the Face of Artificial Intelligence Projects." pith.science (2026). https://pith.science/paper/EJXVFRVH

@misc{pith2026250602214,
  author       = {Pith},
  title        = {Pith review of: Is PMBOK Guide the Right Fit for AI? Re-evaluating Project Management in the Face of Artificial Intelligence Projects},
  year         = {2026},
  howpublished = {\url{https://pith.science/paper/EJXVFRVH}},
  note         = {Machine review of arXiv:2506.02214}
}
read the original abstract

This paper critically evaluates the applicability of the Project Management Body of Knowledge (PMBOK) Guide framework to Artificial Intelligence (AI) software projects, highlighting key limitations and proposing tailored adaptations. Unlike traditional projects, AI initiatives rely heavily on complex data, iterative experimentation, and specialized expertise while navigating significant ethical considerations. Our analysis identifies gaps in the PMBOK Guide, including its limited focus on data management, insufficient support for iterative development, and lack of guidance on ethical and multidisciplinary challenges. To address these deficiencies, we recommend integrating data lifecycle management, adopting iterative and AI project management frameworks, and embedding ethical considerations within project planning and execution. Additionally, we explore alternative approaches that better align with AI's dynamic and exploratory nature. We aim to enhance project management practices for AI software projects by bridging these gaps.

Discussion (0). Sign in to comment.

Reference graph

Works this paper leans on

73 extracted references · 73 canonical work pages

  1. [1]

    A dataset is a collection of structured or unstructured data used to train or evaluate an AI model's performance

    Analysis Data is a critical component of AI projects and significantly influences the success of development efforts. A dataset is a collection of structured or unstructured data used to train or evaluate an AI model's performance. The datasets can include various data types, such as images, alphanumeric text, audio, telemetry, and video, and may or may n...

  2. [2]

    Ensure transparent communication with stakeholders regarding data privacy, compliance, and ethical concerns

    Recommendations We give the following tailoring guidance by performance domains to better address the Data Dependency aspect of AI project projects: Table 1 Data Dependency Tailoring Guidance Domain Recommendations Stakeholder Facilitate collaboration between data scientists, legal advisors, subject matter experts (SME), and end -users to align data deliv...

  3. [3]

    Artificial Intelligence Market Size, Share & COVID -19 Impact Analysis, By Technology, By End -User Industry, By Application, and Regional Forecast, 2025–2032,

    Fortune Business Insights, "Artificial Intelligence Market Size, Share & COVID -19 Impact Analysis, By Technology, By End -User Industry, By Application, and Regional Forecast, 2025–2032," 2025

  4. [4]

    Project Management Evolution: From Traditional IT Implementations to AI-Driven Projects,

    M. R. Martins, "Project Management Evolution: From Traditional IT Implementations to AI-Driven Projects," 2023

  5. [5]

    Black Box problem

    Analysis While the PMBOK Guide emphasizes the importance of managing uncertainty for project success and presents various management techniques such as uncertainty identification and assessment, risk management, adaptive planning, team capability enhancement, and stakeholde r management, it falls short in providing specific guidelines for the unique chall...

  6. [6]

    unknown unknowns

    Recommendations Given the prevalence of "unknown unknowns " in AI projects, where business needs and AI models are subject to continual change, an iterative, experimental approach to AI software development is needed in the development process

  7. [7]

    Playbook for Project Management in Data Science and Artificial Intelligence Projects,

    PMI South Asia & NASSCOM, "Playbook for Project Management in Data Science and Artificial Intelligence Projects," 2020

  8. [8]

    Managing artificial intelligence projects: Key insights from an AI consulting firm,

    G. Vial, A. -F. Cameron and T. Giannelia, "Managing artificial intelligence projects: Key insights from an AI consulting firm," Information Systems in and for Practice, vol. 33, no. 3, pp. 669 -691, 2023

Show all 73 references
  1. [9]

    Metrics: Establish AI quality evaluation criteria, continuously measure accuracy, performance, reliability, explainability, bias, and fairness, and align with business performance

    [43]. Metrics: Establish AI quality evaluation criteria, continuously measure accuracy, performance, reliability, explainability, bias, and fairness, and align with business performance. Regularly conduct A/B testing and model validation [42] [43]. Quality Evaluation Automatio...

  2. [10]

    These benefit from continuous refinement based on user feedback and evolving needs

    Analysis Traditional software components like user interfaces and customer-facing features are particularly well-suited to Agile. These benefit from continuous refinement based on user feedback and evolving needs. Agile breaks development into smaller, predictable sprints, ena...

  3. [11]

    Balancing two traditional software developments with experiment-driven AI component research is necessary

    Recommendations Several tailoring recommendations strengthen the development approach and address the gaps in managing iterative development for AI software projects. Balancing two traditional software developments with experiment-driven AI component research is necessary. The...

  4. [12]

    How to Train an Accurate and Efficient Object Detection Model on any Dataset,

    G. Zalesskaya, B. Bylicka and E. Liu, "How to Train an Accurate and Efficient Object Detection Model on any Dataset," in Proceedings of the 18th International Joint Conference on Computer Vision, Imaging and Computer Graphics Theory and Applications (VISIGRAPP

  5. [13]

    Establishing feedback loops and metrics that enable continuous refinement of AI models is also essential

    in AI research, including Cross -Industry Standard Process for Machine Learning (CRISP -ML(Q)) [47], Machine Learning Workflow (MLW) [48], Data Science Workflow (DSW) [49], Machine Learning Lifecycle (MLLC) [50], Process for Developing and Deploying AI models (PDDA) [51], MLOp...

  6. [14]

    Analysis AI software projects demand specialized expertise, particularly in fields like artificial intelligence (AI) and machine learning (ML): • Multidisciplinary Teams: teams with diverse, hyper- specialized skills in AI and machine learning develop AI projects . These teams...

  7. [15]

    Recommendations The following table presents tailoring recommendations for specialized expertise by performance domains. Table 4 Specialized Expertise Tailoring Guidance Domain Recommendations Team Multidisciplinary Teams [4]: Structure multidisciplinary teams with clearly def...

  8. [16]

    Its impact goes beyond technology, influencing individuals, societies, and the environment

    Analysis AI is revolutionizing industries, offering immense potential and raising significant ethical concerns. Its impact goes beyond technology, influencing individuals, societies, and the environment. Recent controversies, such as Clearview AI's unauthorized facial recognit...

  9. [17]

    [34] [35]. Based on the literature review and the case studies, we see the following key ethical considerations for AI projects: • Bias and fairness: Bias in AI refers to systematic algorithm errors that lead to unfair or discriminatory outcomes for certain groups, often based...

  10. [18]

    Domain Recommendations Promote transparency: Communicate AI components, potential impacts, and ethical considerations to stakeholders, ensuring informed participation

    Recommendations Below, we provide tailoring recommendations for the PMBOK Guide to better address the unique ethical aspects of AI projects: Table 5 Ethical Aspects Tailoring Guidance Domain Recommendations Stakeholder Ethical stakeholder engagement: Involve ethicists, regulat...

  11. [19]

    Artificial intelligence in information systems research: A systematic literature review and research agenda,

    C. Collins, D. Dennehy, K. Conboy and P. Mikalef, "Artificial intelligence in information systems research: A systematic literature review and research agenda," International Journal of Information Management, vol. 60, 2021

  12. [20]

    Grand View Research, "Artificial Intelligence Market Size, Share & Trends Analysis Report By Solution, By Technology (Deep Learning, Machine Learning, NLP, Machine Vision, Generative AI), By Function, By End -use, By Region, And Segment Forecasts, 2024 – 2030," 2024

  13. [21]

    Data Issues in Industrial AI System: A Meta-Review and Research Strategy,

    X. Li, C. Yang, C. Møller and J. Lee, "Data Issues in Industrial AI System: A Meta-Review and Research Strategy," 2024

  14. [22]

    Collaboration Challenges in Building ML -Enabled Systems: Communication, Documentation, Engineering, and Process,

    N. Nahar, S. Zhou, G. Lewis and C. Kästner, "Collaboration Challenges in Building ML -Enabled Systems: Communication, Documentation, Engineering, and Process," in 2022 IEEE/ACM 44th International Conference on Software Engineering (ICSE), Pittsburg, PA, USA, 2022

  15. [23]

    PMI, PMBOK® Guide: A Guide to the Project Management Body of Knowledge (PMBOK® Guide) – Seventh Edition and The Standard for Project Management (Edition 7) (Paperback), Project Management Institute, 2021

  16. [24]

    Making Sense of AI Systems Development,

    M. Dolata and K. Crowston, "Making Sense of AI Systems Development," in IEEE Transactions on Software Engineering, 2024

  17. [25]

    The Root Causes of Failure for Artificial Intelligence Projects and How They Can Succeed: Avoiding the Anti-Patterns of AI,

    J. Ryseff, B. F. De Bruhl and S. J. Newber, "The Root Causes of Failure for Artificial Intelligence Projects and How They Can Succeed: Avoiding the Anti-Patterns of AI," 2024

  18. [26]

    Requirements Engineering in Machine Learning Projects,

    A. Gjorgjevikj, K. Mishev and L. Antovski, "Requirements Engineering in Machine Learning Projects," IEEE Access, vol. 11, pp. 72186 - 72208, 2023

  19. [27]

    [28] and experimentation [5] [29] [30], the need for specialized expertise [4] [31], and ethical considerations [32]

  20. [28]

    Which Project Management Methodology is better for AI -Transformation and Innovation Projects?,

    A. Najdawi and A. Shaheen, "Which Project Management Methodology is better for AI -Transformation and Innovation Projects?," in 2021 International Conference on Innovative Practices in Technology and Management (ICIPTM), Noida, India, 2021

  21. [29]

    Ready for Managing AI Projects? An Analysis of AI Project Management Frameworks,

    L. Wrobel, C. Dietzmann and R. Alt, "Ready for Managing AI Projects? An Analysis of AI Project Management Frameworks," in Proceedings of the 58th Hawaii International Conference on System Sciences | 2025, 2025

  22. [30]

    New Intel® Geti ™ Software Platform Brings the Power of AI to the Entire Team,

    P. Ramos, "New Intel® Geti ™ Software Platform Brings the Power of AI to the Entire Team," 7 11 2022. [Online]. Available: https://medium.com/openvino-toolkit/new-intel-geti-software- platform-brings-the-power-of-ai-to-the-entire-team-38e49c75d9a. [Accessed 1 23 2025]

  23. [31]

    Intel Geti GitHub,

    "Intel Geti GitHub," 4 2025. [Online]. Available: https://github.com/open-edge-platform/geti

  24. [32]

    OpenVINO,

    "OpenVINO," 23 1 2025. [Online]. Available: https://github.com/openvinotoolkit/openvino. [Accessed 23 1 2025]

  25. [33]

    black box problem

    [34] [35] [29]. Below, we formulate the description of these project features (PF): • Data Dependency (PF1): AI projects rely heavily on high-quality, relevant datasets for training and performance. Ensuring data quality, accuracy, completeness, and bias mitigation is critical...

  26. [34]

    OpenVINO™ Training Extensions,

    "OpenVINO™ Training Extensions," Intel, 2025. [Online]. Available: https://github.com/open-edge-platform/training_extensions. [Accessed 2025]

  27. [35]

    Datumaro,

    "Datumaro," 23 1 2025. [Online]. Available: https://github.com/open- edge-platform/datumaro. [Accessed 23 1 2025]

  28. [36]

    Anomalib: A Deep Learning Library for Anomaly Detection,

    S. Akcay, D. Ameln, A. Vaidya, B. Lakshmanan, N. Ahuja and U. Genc, "Anomalib: A Deep Learning Library for Anomaly Detection," in 2022 IEEE International Conference on Image Processing (ICIP), Bordeaux, France, 2022

  29. [37]

    Anomalib,

    "Anomalib," 1 23 2025. [Online]. Available: https://github.com/open- edge-platform/anomalib. [Accessed 23 1 2025]

  30. [38]

    OpenVINO™ Explainable AI Toolkit - OpenVINO XAI,

    "OpenVINO™ Explainable AI Toolkit - OpenVINO XAI," 2025. [Online]. Available: https://github.com/openvinotoolkit/openvino_xai. [Accessed 2025]

  31. [39]

    Data Collection and Quality Challenges in Deep Learning: A Data -Centric AI Perspective,

    S. Whang, Y. Roh, H. Song and J. -G. Lee, "Data Collection and Quality Challenges in Deep Learning: A Data -Centric AI Perspective," The VLDB Journal, vol. 32, p. 791–813, 2023

  32. [40]

    Algorithmic progress in language models,

    A. Ho, T. Besiroglu, E. Erdil, D. Owen, R. Robi, Z. C. Guo, D. Atkinson, N. Thompson and J. Sevilla, "Algorithmic progress in language models," 2024

  33. [41]

    [42]. The following tailoring guidance can supplemented to better address the unique aspects of AI project management [37] [42] [43] [22] [44]: Table 2 Uncertainty & Experimentation Aspect Tailoring Guidance Domain Recommendations Stakeholder Educate on AI projects: Regularly ...

  34. [42]

    Domain Recommendations AI Quality Standards and Transparency: Establish clear quality standards for AI -powered software deployments

    [43]. Domain Recommendations AI Quality Standards and Transparency: Establish clear quality standards for AI -powered software deployments. Employ Explainable AI techniques to enhance the predictability and trustworthiness of AI model outputs [37]. Post-deployment refinement: ...

  35. [43]

    Sammicheli, Scrum in AI: Artificial Intelligence Agile Development with Scrum and MLOps, Leanpub, 2023

    P. Sammicheli, Scrum in AI: Artificial Intelligence Agile Development with Scrum and MLOps, Leanpub, 2023

  36. [44]

    Analysis of Software Engineering for Agile Machine Learning Projects,

    K. Singla, J. Bose and C. Naik, "Analysis of Software Engineering for Agile Machine Learning Projects," 2019

  37. [45]

    The AI research parts are less predictable and may span over a few agile sprints or rapidly change direction

    [46]. The AI research parts are less predictable and may span over a few agile sprints or rapidly change direction. The lifecycle shall adapt to the uncertain nature of the research . Project team can leverage a set of emerging process models

  38. [46]

    Requirements Engineering for Machine Learning: A Review and Reflection,

    Z. Pei, L. Liu, C. Wang and J. Wang, "Requirements Engineering for Machine Learning: A Review and Reflection," in 2022 IEEE 30th International Requirements Engineering Conference Workshops (REW), 2022

  39. [47]

    Machine learning with requirements: A manifesto,

    E. Giunchiglia and F. Imrie, "Machine learning with requirements: A manifesto," Neurosymbolic Artificial Intelligence, pp. 1-13, 2024

  40. [48]

    Ahmad, M

    K. Ahmad, M. Abdelrazek, C. Arora, M. Bano and J. Grundy, Requirements engineering for artificial intelligence systems: A systematic mapping study, vol. 158, Elsevier, 2023

  41. [49]

    How does Machine Learning Change Software Development Practices?,

    Z. Wan, X. Xia, D. Lo and G. C. Murphy, "How does Machine Learning Change Software Development Practices?," in IEEE Transactions on Software Engineering, 2021

  42. [50]

    Piorkowski, S

    D. Piorkowski, S. Park, A. Y. Wang and D. Wang, How AI Developers Overcome Communication Challenges in a Multidisciplinary Team: A Case Study, 2021

  43. [51]

    Top 10 Ethical Considerations for AI Projects [Blog post],

    R. Schmelzer and K. Walch, "Top 10 Ethical Considerations for AI Projects [Blog post]," 15 1 2025. [Online]. Available: https://www.pmi.org/blog/top-10-ethical-considerations-for-ai- projects

  44. [52]

    Ethically Aligned Design: A Vision for Prioritizing Human Well -being with Autonomous and Intelligent Systems – Version II Request for Input,

    IEEE Standards Association, "Ethically Aligned Design: A Vision for Prioritizing Human Well -being with Autonomous and Intelligent Systems – Version II Request for Input," 2018

  45. [53]

    OECD AI Principles overview,

    "OECD AI Principles overview," [Online]. Available: https://oecd.ai/en/ai-principles

  46. [54]

    Ethics guidelines for trustworthy AI,

    Smuha, N. (Coordinator), "Ethics guidelines for trustworthy AI," High-Level Expert Group on AI established by European Commission,, [Online]. Available: https://ec.europa.eu/newsroom/dae/document.cfm?doc_id=60419

  47. [55]

    Comparing PMBOK and Agile Project Management software development processes,

    P. Fitsilis, "Comparing PMBOK and Agile Project Management software development processes," in Advances in Computer and Information Sciences and Engineering, 2008

  48. [56]

    State -of-the-Art Review of Taxonomies for Quality Assessment of Intelligent Software Systems,

    C. Arina, A. Jabborov, A. Kruglov and G. Succi, "State -of-the-Art Review of Taxonomies for Quality Assessment of Intelligent Software Systems," in 2022 3rd International Informatics and Software Engineering Conference (IISEC), Ankara, Turkey, 2022

  49. [57]

    Evaluating Large Language Models: A Comprehensive Survey,

    Z. Zishan Guo, R. Jin, C. Liu, Y. Huang, D. Shi, S. L. Yu, Y. Liu, J. Li, B. Xiong and D. Xiong, "Evaluating Large Language Models: A Comprehensive Survey," 2023

  50. [58]

    A survey of safety and trustworthiness of large language models through the lens of verifcation and validation,

    X. Huang, W. Ruan, W. Huang, G. Jin, Y. Dong, C. Wu, S. Bensalem, R. Mu, Y. Qi, X. Zhao, K. Cai, Y. Zhang, S. Wu, P. Xu, D. Wu, A. Freitas and M. A. Mustafa, "A survey of safety and trustworthiness of large language models through the lens of verifcation and validation," Artif...

  51. [59]

    R. V. Yampolskiy, AI: Unexplainable, Unpredictable, Un controllable, CRC Press, 2024

  52. [60]

    Continuous experimentation on artificial intelligence software: a research agenda,

    A. Nguyen-Duc and P. Abrahamsson, "Continuous experimentation on artificial intelligence software: a research agenda," in ESEC/FSE 2020: Proceedings of the 28th ACM Joint Meeting on European Software Engineering Conference and Symposium on the Foundations of Software Engineering, 2020

  53. [61]

    AI Risk Management Framework,

    NIST, "AI Risk Management Framework," January 2023. [Online]. Available: https://www.nist.gov/itl/ai-risk-management-framework

  54. [62]

    Quality issues in machine learning software systems,

    P.-O. Côté, A. Nikanjam and R. Bou, "Quality issues in machine learning software systems," Empirical Software Engineering, vol. 29, pp. 149:1-47, 2024

  55. [63]

    Agile, Traditional, and Hybrid Approaches to Project Success: Is Hybrid a Poor Second Choice?,

    A. Gemino, B. Horner Reich and P. M. Serrador, "Agile, Traditional, and Hybrid Approaches to Project Success: Is Hybrid a Poor Second Choice?," Project Management Journal, no. 52, 2020

  56. [64]

    One size does not fit all: Choosing the right project approach,

    S. C. Burgan and D. S. Burgan, "One size does not fit all: Choosing the right project approach," in PMI® Global Congress 2014 , North America, Phoenix, AZ, 2014

  57. [65]

    Towards CRISP -ML(Q): A Machine Learning Process Model with Quality Assurance Methodology,

    S. Studer, T. Binh Bui, C. Drescher, A. Hanuschkin, L. Winkler , S. Peters and K. -R. Müller, "Towards CRISP -ML(Q): A Machine Learning Process Model with Quality Assurance Methodology," Mach. Learn. Knowl. Extr., vol. 3, no. 2, pp. 392-413, 2021

  58. [66]

    Software Engineering for Machine Learning: A Case Study,

    S. Amershi, A. Begel, C. Bird, R. DeLine, H. Gall, E. Kamar, N. Nagappan, B. Nushi and T. Zimmermann, "Software Engineering for Machine Learning: A Case Study," in 2019 IEEE/ACM 41st International Conference on Software Engineering: Software Engineering in Practice (ICSE-SEIP), 2019

  59. [67]

    Human -AI Collaboration in Data Science: Exploring Data Scientists' Perceptions of Automated AI,

    D. Wang, J. D. Weisz, M. Muller, P. Ram, W. Geyer, C. Dugan, Y. Tausczik, H. Samulowitz and A. Gray, "Human -AI Collaboration in Data Science: Exploring Data Scientists' Perceptions of Automated AI," in Proceedings of the ACM on Human -Computer Interaction, 2019

  60. [68]

    Assuring the Machine Learning Lifecycle: Desiderata, Methods, and Challenge,

    R. Ashmore, R. Calinescu and C. Paters, "Assuring the Machine Learning Lifecycle: Desiderata, Methods, and Challenge," in ACM Computing Surveys (CSUR), Volume 54, Issue 5, 2021

  61. [69]

    Artificial intelligence in operations management and supply chain management: an exploratory case study,

    P. Helo and Y. Hao, "Artificial intelligence in operations management and supply chain management: an exploratory case study," Industry Experiences of Artificial Intelligence: Benefits and Challenges in Operations and Supply Chain Management, vol. 33, no. 16, pp. 1573- 1590, 2022

  62. [70]

    From DevOps to MLOps: Overview and Application to Electricity Market Forecasting,

    R. Subramanya, S. Sierla and V. Vyatkin, "From DevOps to MLOps: Overview and Application to Electricity Market Forecasting," Applied Sciences, vol. 12, no. 19, 2022

  63. [71]

    AI and Privacy Risks: The EU AI Act vs. US NIST AI Risk Management Framework, TTLF Working Papers No. 125,

    E. K. Cortez, "AI and Privacy Risks: The EU AI Act vs. US NIST AI Risk Management Framework, TTLF Working Papers No. 125," Transatlantic Technology Law Forum, 2024

  64. [72]

    Amazon Scraps Secret AI Recruiting Tool that Showed Bias against Women,

    J. Dustin, "Amazon Scraps Secret AI Recruiting Tool that Showed Bias against Women," in Ethics of Data and Analytics , Auerbach Publications, 2022

  65. [73]

    Artificial Intelligence Index Report 2024,

    N. Maslej, L. Fattorini, R. Perrault, V. Parli, A. Reuel, E. Brynjolfsson, J. Etchemendy, K. Ligett, T. Lyons, J. Manyika, J. C. Niebles, Y. Shoham, R. Wald and J. Clark, "Artificial Intelligence Index Report 2024," AI Index Steering Committee, Institute f or Human-Centered AI...

Pith tools

Reviewed August 7, 2026 · model on record in the stance chip above.