186 American Academic Scientific Research Journal for Engineering, Technology, and Sciences ISSN (Print) 2313-4410, ISSN (Online) 2313-4402 https://asrjetsjournal.org/index.php/American_Scientific_Journal/index Leveraging Artificial Intelligence and Data Analytics for Decision-Making in IT Project Management Ievgenii Lysenko * IT Project Manager at PwC, Vinnytsia, Ukraine Email: yevhenii.lysenko@vntu.net Abstract The article focuses on enhancing the justification of managerial decisions in large IT projects, characterized by persistent instability and escalating budgetary and schedule risks, through the systematic integration of artificial intelligence and advanced data analytics. The objectives of this study are twofold: first, to comprehensively describe the mechanisms for deploying predictive models and generative AI tools at every phase of the IT project life cycle; and second, to empirically validate the claimed effects using the PwC CEE IT practice. The novelty of the work lies in the combination of predictive machine learning, Monte Carlo simulations, AI scoring and the Copilot generative planner in a single decision-making loop, as well as in the fact that the author described in detail his experience of implementing and adapting these technologies in the real PMO process of PwC CEE IT. This article will be helpful to project leaders, PMO analysts, and developers of decision-support systems in IT project management. Keywords: artificial intelligence; data analytics; IT project management; predictive analytics; generative AI; Monte Carlo simulation; decision-making. 1. Introduction Despite widespread process standardization and the adoption of PMBOK and PRINCE2 frameworks, large IT initiatives continue to exhibit alarmingly high instability. Significant cost overruns result in direct capital losses and erode stakeholder trust, making the justification of decisions a central concern in modern project management. ------------------------------------------------------------------------ Received: 4/20/2025 Accepted: 6/1/2025 Published: 6/13/2025 ------------------------------------------------------------------------ * Corresponding author. https://asrjetsjournal.org/index.php/American_Scientific_Journal/index American Academic Scientific Research Journal for Engineering, Technology, and Sciences (ASRJETS) - Volume 102, No 1, pp 186-194 187 The advent of mature AI tools and advanced data analytics is transforming this landscape. Integrating predictive models into management platforms can reduce the share of failed projects by an average of 30% through more accurate risk forecasting and earlier detection of deviations [1]. Corporate centralized data repositories and machine-learning algorithms establish a continuous feedback loop, where each completed iteration enriches the next, diminishing reliance on managers’ subjective experience. Consequently, the digital “co-pilot” extends initiation and planning processes beyond traditional manual timeframes, freeing experts to focus on strategic analysis and cross-functional coordination. Against this backdrop, the present study pursues two interconnected goals: (1) to systematically describe how end-to-end integration of AI and data analytics strengthens decision justification at every phase of the IT project lifecycle; and (2) to empirically verify these effects using the author’s PwC CEE IT practice as a case study. 2. Materials and Methodology This investigation into the integration of AI and data analytics for informed decision-making in IT project management is grounded in the analysis of 15 key sources, including academic articles, McKinsey and Capterra reports, industry publications, and the author’s empirical case study at PwC CEE IT. The theoretical foundation comprises Thamma [1], who demonstrated reduced project-failure rates via predictive models; Bloch and his colleagues [2] and the CHAOS Report [4], detailing the principal causes of budget overruns; and Chandrasekaran [5], who outlined the “Projects 5.0” concept incorporating digital twins and AI scoring to improve ROI. Methodologically, the study combined: ● Comparative analysis of technologies—juxtaposition of traditional PM frameworks (PMBOK, PRINCE2) with generative-AI and advanced-analytics platforms (Planisware Generative AI [9], Microsoft Copilot [12], TrueProject [11]) to assess their impact on planning speed, quality, and early risk detection [1,9,11]. ● Systematic review of industry reports—meta-analysis of McKinsey's “Projects 5.0” data [5], Capterra AI- feature ROI findings [7], and Shell predictive-maintenance case studies [6], to evaluate the economic efficiency of a data-driven transition. ● Empirical case study—collection and analysis of PwC CEE IT project metrics from April through November 2024: budget deviations, days to launch, and person-hour expenditures; application of Monte Carlo simulation to quantify the impact of managerial interventions [8,10]. ● Content analysis of surveys and interviews—synthesis of feedback from 300 capital-program executives and 1,440 project managers on generative-AI use in PMOs, and their requests for AI-literacy training [5], [15]. 3. Results and Discussion The stability of outcome metrics remains a critical weakness in traditional IT project management. A comparative analysis of 1,471 large initiatives revealed that, among the 45% of projects over USD 15 million that exceeded budget, 13% of overruns were due to loss of focus, 11% to execution challenges, 9% to American Academic Scientific Research Journal for Engineering, Technology, and Sciences (ASRJETS) - Volume 102, No 1, pp 186-194 188 requirement changes, and 6% each to skill gaps and unexplained causes (Fig. 1). Moreover, 7% of budget- overrun projects also missed deadlines. Actual returns averaged 56% below expectations [2]. The primary triggers are unclear objectives, unrealistic schedules, and constant scope changes. Figure 1: Distribution of reasons for budget overruns in large IT projects [2] When scaled to megaprojects, overruns intensify: one analysis [3] documented average cost overruns of at least 79% and schedule delays of 52% relative to the original baselines. The CHAOS Report further indicates that only 31% of IT projects are completed successfully, 50% are finished with resource overruns and reduced functionality, and 19% are terminated prematurely [4]. This volatility stems from escalating complexity in technical and organizational interdependencies. Microservice architectures, cloud infrastructures, and numerous third-party APIs increase the likelihood of integration defects, making it more challenging to trace critical paths. Under limited observability, managers rely on expert estimates that fail to capture nonlinear scale effects and task interlocks, systematically understating early budget and schedule baselines. Poor data reliability and timeliness exacerbate the issue. Most organizations consolidate labor and performance metrics post hoc, depriving teams of the ability to make dynamic corrections. Cognitive biases—such as optimism, strategic risk underestimation to secure funding, and decision latency—further delay the detection of problems. The resulting vicious cycle of complexity, fragmented information, and biased judgments heightens the likelihood of overrun. Predictive analytics and AI methods aim to break this cycle precisely. The financial instability revealed by traditional approaches underpins the economic case for data-driven models. American Academic Scientific Research Journal for Engineering, Technology, and Sciences (ASRJETS) - Volume 102, No 1, pp 186-194 189 While megaprojects currently overrun budgets and schedules by 30–45 %, a survey of over 300 capital-program executives found that systematic adoption of “Projects 5.0” digital principles—predictive models, digital twins, and end-to-end data exchange—reduces costs and durations by 30–40 %, more than doubling portfolio-wide ROI [5]. Gains arise from early risk detection, critical-path optimization, and dynamic resource allocation, thereby reducing failure probability and freeing capital for higher-priority initiatives. Industry leaders confirm these effects. Shell’s AI-driven predictive-maintenance deployment reduced direct repair expenses by 20%, saving approximately USD 2 billion annually [6]. A Capterra meta-analysis of 2,500 managers found that 90% of AI-feature users achieved a positive return on investment (ROI) and 63% reported significant productivity gains [7]. These findings illustrate that data-driven management yields scalable benefits across sectors. Predictive analytics transforms expert estimates into statistically validated economic forecasts. Strategic viability is measured by net present value (NPV). Pre-construction excellence practices across various capital projects have boosted NPV by at least 20% through refined cash-flow inputs and dynamic risk accounting [3]. McKinsey analysts note that IRR can double in partial “sell-down” transactions without real value creation, whereas modified IRR and NPV retain accuracy; thus, combining NPV with MIRR offers reliable return metrics. Monte Carlo methods simulate up to 10,000 schedule scenarios, producing empirical distributions of completion times and quantifying the effects of interventions. In study [8], a 20% staffing increase reduced the deadline- miss probability from 50% to under 15%. Each managerial action is evaluated by its delta in failure probability and the present value of penalty avoidance, shifting decision-making from opinion to quantitative risk. At the portfolio level, these simulations feed automated scorecards. Platforms that leverage analytic hierarchy processes and AI ranking report a doubling of portfolio ROI after transitioning from manual selection to formal, weighted criteria. Hence, the “NPV/MIRR + Monte Carlo + AI scoring” pipeline creates a closed-loop decision framework, where each initiative undergoes quantitative vetting of profitability, schedule, and risk, and outputs are consolidated into a managed portfolio with forecasted economic returns. Generative AI further elevates tactical planning by converting text prompts into ready-made artifacts. For example, Planisware Enterprise’s Generative AI module produces a Work Breakdown Structure (WBS) and detailed schedule in seconds, rather than hours of manual work, and simultaneously generates an initial risk register [9]. Watson AI reduced baseline-plan preparation time by 25% in large IT deployments [10], freeing resources for what-if analysis via natural-language queries to Co-Pilot, which instantly recalculates critical paths and cost/time deltas—tasks that formerly required manual network adjustments and repeated Monte Carlo runs. Pilot implementations of Microsoft Copilot in mid-sized enterprises report up to 40% increases in output efficiency, as staff shift time from routine plan editing to analytical interpretation and stakeholder engagement [11]. Thus, generative AI transforms planning from a deterministic, labor-intensive task into an interactive dialogue, testing hypotheses in real time with immediate economic evaluation. American Academic Scientific Research Journal for Engineering, Technology, and Sciences (ASRJETS) - Volume 102, No 1, pp 186-194 190 Continuous risk analytics replaces periodic reporting. TrueProject-level platforms stream Cost Performance Index and Schedule Performance Index in real time, highlighting anomalies well before cost or schedule breaches [11]. 4castplus control towers aggregate EV, CV, SV, and Estimate-at-Complete, compute rolling- window deviations, and trigger ML alerts when trends exit confidence intervals, enabling teams to detect variances two reporting cycles earlier than traditional EVM [12]. McKinsey’s Project Delivery Hub integrates historical CPI/SPI, weather, logistics, and contract data into an XGBoost model that generates weekly threat ratings. Its deployment on an oil-and-gas site reduced schedule lag and unlocked over USD 75 million through early bottleneck escalation [13]. Engineering and business metrics converge through DevOps DORA measures—deployment frequency, lead time, change-failure rate, and mean time to resolution (MTTR)—streamed from CI/CD pipelines into the same risk-management system as financial indicators. Google’s Accelerate State of DevOps study shows that elite performers deploy code 973 times more often and recover from incidents 6,570 times faster than low performers, directly correlating with profitability and market share gains [14]. This “from-code-to-cash” continuum consolidates technical, temporal, and financial data in a continuous stream, with machine-learning algorithms bridging the gap between DevOps teams and executive leadership. To convert these data streams into uncertainty reduction, PMOs must adopt formal Data Governance, defining unified data definitions, provenance, and validation rules to eliminate noise and prevent false alarms. The Data Steward, as the “guardian of data quality, compliance, and context,” is pivotal; without accountability, AI initiatives fail to scale effectively. Yet, regulations alone are insufficient; McKinsey finds that 48% of employees cite formal generative AI training as key to increased daily usage. In comparison, 46% of executives identify a skills gap as the primary barrier to the rollout of AI production [15]. A survey of 1,440 U.S. managers found that 90% receive inquiries about new generative AI tools at varying frequencies: less than quarterly (5%), quarterly (5%), weekly (15%), several times a week (28%), to multiple times a day (16%); only 10% never receive such queries (Fig. 2) [15]. American Academic Scientific Research Journal for Engineering, Technology, and Sciences (ASRJETS) - Volume 102, No 1, pp 186-194 191 Figure 2: Frequency of team inquiries about using new gen-AI tools at work [15] To prioritize effectively, manage risks, optimize resources, and align functions, organizations need a generative AI roadmap. A U.S. C-suite survey revealed that none lack an AI-adoption plan; 21% are developing one, 53% have a plan under refinement, and only 25% report having a fully articulated roadmap (Fig. 3) [15]. Figure 3: Cross-Sectional Assessment of Generative AI Implementation Road-Map Adoption Among C-Suite Executives [15] American Academic Scientific Research Journal for Engineering, Technology, and Sciences (ASRJETS) - Volume 102, No 1, pp 186-194 192 Most companies are strategizing the use of generative AI, although only a quarter have finalized plans. Accordingly, PMOs are instituting AI literacy tracks: project managers learn prompt engineering, predictive model interpretation, and data ethics, while Data Stewards mentor teams to embed standards into project practices. This cultural shift completes the logic chain: cleaned, annotated data fuel predictive models; Co-Pilot– integrated models accelerate planning; trained managers critically evaluate AI outputs in budget, schedule, and risk decisions—transforming the PMO into a management-insights factory, where every digital artifact undergoes semantic and economic validation before executive review. In the author’s practical case at PwC CEE IT over the past year, all incoming digital-service and automation ideas were logged in a standardized Google Sheets template, including labor forecasts, financial impact, scalability, and strategic alignment. An API connector to Microsoft Copilot enables on-demand analytics, aggregating parameters, ranking ROI probability, and generating textual recommendations that serve as inputs for leadership judgments. On average, seven requests per month entered this “Copilot + human” initiation funnel; roughly five (or sixty per year) were rejected or sent for revision due to low forecast impact or scalability concerns. Based on the scale of the company, each such decision freed approximately 480 person- hours, totaling 28,800 hours annually—saving approximately USD 1.15 million at a USD 40/hour labor rate— and reallocating resources to higher-return-on-investment (ROI) projects. Transparency improved as Copilot comments were recorded in a traceable format; stakeholder feedback accelerated from days to minutes; and discussions shifted from “meeting prep” to strategic value assessment. The author’s experience demonstrates how targeted generative-AI use in early lifecycle phases yields measurable financial returns and sets a precedent for deeper PMO analytics adoption. Overall, limited generative analytics integration at decision points resulted in discrete reductions in transaction costs, increased transparency, and faster idea-to-value cycles, yielding over USD 1 million in savings and nearly 30,000 freed project hours—empirically validating the economic rationale for a data-driven approach. 4. Conclusion In the face of ever-growing technical and organizational complexity, large IT projects remain precariously unstable, with schedule delays and budget overruns resulting in significant financial losses and eroding stakeholder trust. Implementing mature AI tools and advanced data analytics enables a shift from subjective expert judgments to systematic, quantitatively substantiated decision-making, extending initiation and planning beyond traditional manual timeframes, reducing failure rates by an average of 30%, and improving resource allocation efficiency. While the author did not place full reliance on AI-generated outputs, instead incorporating them as a supplementary parameter within the experimental framework and prioritizing their own expert judgment, the outcomes revealed a notable convergence between the author’s conclusions and the suggestions provided by the AI. End-to-end analytics and AI scoring across all lifecycle phases introduce new capabilities for controlling budget, schedule, and quality. Monte Carlo methods, simulating up to 10,000 scenarios, can quantitatively assess American Academic Scientific Research Journal for Engineering, Technology, and Sciences (ASRJETS) - Volume 102, No 1, pp 186-194 193 interventions and reduce the probability of deadline misses from 50% to under 15%. Portfolio-level automated scorecards based on NPV/MIRR and AI-weighted criteria double aggregate ROI by replacing manual selection with formal screening. Generative AI modules transform text prompts into ready planning artifacts—WBS to initial risk registers—reducing baseline plan preparation time by 25% and converting planning into an interactive, economically validated dialogue. Empirical assessment of the author’s PwC CEE IT practice confirmed the high economic and operational viability of a data-driven approach: seven average monthly requests through the Google Sheets + Copilot funnel yielded five revisions or rejections, freeing 480 person-hours per initiative and saving ~28,800 hours annually (~USD 1.15 million). Copilot’s traceable recommendations enhanced justification transparency and accelerated stakeholder feedback, shifting focus from meeting preparation to strategic value evaluation. Thus, the integration of systematic AI and data analytics not only elevates decision justification but also redefines the PMO’s role from a reporting center into a management-insights factory, where every digital artifact undergoes unified validation, from semantic to economic, before informing decisions. The achievements of “Projects 5.0” open new horizons for strategic resource utilization, risk minimization, and sustainable growth of IT initiative portfolios, confirming the successful realization of this study’s goals. References [1] L. Thamma, “How AI will help Project Planning,” Projectmanagement.com, 2023. https://www.projectmanagement.com/wikis/849920/how-ai-will-help-project-planning#_ (accessed Mar. 31, 2025). [2] M. Bloch, S. Blumberg, and J. Laartz, “Delivering large-scale IT projects on time, on budget, and value,” McKinsey, Oct. 01, 2012. https://www.mckinsey.com/capabilities/mckinsey-digital/our- insights/delivering-large-scale-it-projects-on-time-on-budget-and-on-value (accessed Mar. 30, 2025). [3] M. Dussud, “How to improve capital project execution,” McKinsey, Sep. 13, 2023. https://www.mckinsey.com/capabilities/operations/our-insights/seize-the-decade-maximizing-value- through-pre-construction-excellence (accessed Mar. 01, 2025). [4] Radek, “Chaos Report — why this study about IT project management is so unique,” The Story, Nov. 04, 2024. https://thestory.is/en/journal/chaos-report/ (accessed Apr. 02, 2025). [5] S. Chandrasekaran, “Capital Projects 5.0: Reimagining capital-project delivery,” McKinsey, 2021. Accessed: Apr. 03, 2025. [Online]. Available: https://www.mckinsey.com/~/media/mckinsey/business%20functions/operations/our%20insights/capita l%20projects%2050%20reimagining%20capital%20project%20delivery/capital-projects-50- reimagining-capital-project-delivery.pdf [6] Dadang Irsyam, “How AI is Fueling Efficiency: Lessons from Shell’s Gas Industry Transformation,” https://www.projectmanagement.com/wikis/849920/how-ai-will-help-project-planning#_ https://www.projectmanagement.com/wikis/849920/how-ai-will-help-project-planning#_ https://www.projectmanagement.com/wikis/849920/how-ai-will-help-project-planning#_ https://www.mckinsey.com/capabilities/mckinsey-digital/our-insights/delivering-large-scale-it-projects-on-time-on-budget-and-on-value https://www.mckinsey.com/capabilities/mckinsey-digital/our-insights/delivering-large-scale-it-projects-on-time-on-budget-and-on-value https://www.mckinsey.com/capabilities/mckinsey-digital/our-insights/delivering-large-scale-it-projects-on-time-on-budget-and-on-value https://www.mckinsey.com/capabilities/operations/our-insights/seize-the-decade-maximizing-value-through-pre-construction-excellence https://www.mckinsey.com/capabilities/operations/our-insights/seize-the-decade-maximizing-value-through-pre-construction-excellence https://www.mckinsey.com/capabilities/operations/our-insights/seize-the-decade-maximizing-value-through-pre-construction-excellence https://www.mckinsey.com/capabilities/operations/our-insights/seize-the-decade-maximizing-value-through-pre-construction-excellence https://thestory.is/en/journal/chaos-report/ https://thestory.is/en/journal/chaos-report/ https://www.mckinsey.com/~/media/mckinsey/business%20functions/operations/our%20insights/capital%20projects%2050%20reimagining%20capital%20project%20delivery/capital-projects-50-reimagining-capital-project-delivery.pdf https://www.mckinsey.com/~/media/mckinsey/business%20functions/operations/our%20insights/capital%20projects%2050%20reimagining%20capital%20project%20delivery/capital-projects-50-reimagining-capital-project-delivery.pdf https://www.mckinsey.com/~/media/mckinsey/business%20functions/operations/our%20insights/capital%20projects%2050%20reimagining%20capital%20project%20delivery/capital-projects-50-reimagining-capital-project-delivery.pdf https://www.mckinsey.com/~/media/mckinsey/business%20functions/operations/our%20insights/capital%20projects%2050%20reimagining%20capital%20project%20delivery/capital-projects-50-reimagining-capital-project-delivery.pdf https://www.mckinsey.com/~/media/mckinsey/business%20functions/operations/our%20insights/capital%20projects%2050%20reimagining%20capital%20project%20delivery/capital-projects-50-reimagining-capital-project-delivery.pdf American Academic Scientific Research Journal for Engineering, Technology, and Sciences (ASRJETS) - Volume 102, No 1, pp 186-194 194 Medium, Oct. 15, 2024. https://medium.com/%40dirsyamuddin29/how-ai-is-fueling-efficiency-lessons- from-shells-gas-industry-transformation-3e754d4e7ff8 (accessed Apr. 04, 2025). [7] E. Mimms, “AI-Driven Project Management Sees 90% Positive ROI Among Users,” Business Wire, Jul. 08, 2024—https://www.businesswire.com/news/home/20240708556035/en/AI-Driven-Project- Management-Sees-90-Positive-ROI-Among-Users (accessed Apr. 05, 2025). [8] S. H. Tanim and M. S. Ahmad, “AI driven strategic decision-making in IT project management: Enhancing risk assessment, cost control, and efficiency,” World Journal of Advanced Research and Reviews, vol. 25, no. 2, pp. 247–268, 2025.https://journalwjarr.com/node/555 (accessed Apr. 06, 2025). [9] “Planisware Enterprise Generative AI: Accelerating the Project Definition and Risk Assessment,” Planisware, 2023. https://planisware.com/resources/ai-ppm/planisware-enterprise-generative-ai- accelerating-project-definition-and-risk (accessed Apr. 07, 2025). [10] D. S. Adamantiadou and L. Tsironis, “Leveraging Artificial Intelligence in Project Management: A Systematic Review of Applications, Challenges, and Future Directions,” Computers, vol. 14, no. 2, p. 66, 2025.https://www.mdpi.com/2073-431X/14/2/66 (accessed Apr. 08, 2025). [11] S. S. Almalki, “AI-Driven Decision Support Systems in Agile Software Project Management: Enhancing Risk Mitigation and Resource Allocation,” Systems, vol. 13, no. 3, p. 208, 2025.https://www.mdpi.com/2079-8954/13/3/208 (accessed Apr. 12, 2025). [12] “How Microsoft Copilot is Transforming Medium-Sized Enterprises,” Cloudtango, Feb. 10, 2025. https://www.cloudtango.net/blog/2025/02/10/how-microsoft-copilot-is-transforming-medium-sized- enterprises/ (accessed Apr. 10, 2025). [13] D. Ahmoye, “Better tech for closer tracking of capital projects | McKinsey,” McKinsey, Aug. 31, 2021. https://www.mckinsey.com/capabilities/operations/our-insights/project-delivery-hubs-analytics- for-better-megaproject-outcomes (accessed Apr. 11, 2025). [14] “Accelerate State Of DevOps 2021,” Google Cloud. https://cloud.google.com/resources/state-of- devops (accessed Apr. 12, 2025). [15] H. Mayer, L. Yee, M. Chui, and R. Roberts, “Superagency in the workplace: Empowering people to unlock AI’s full potential,” McKinsey, Jan. 28, 2025. https://www.mckinsey.com/capabilities/mckinsey-digital/our-insights/superagency-in-the-workplace- empowering-people-to-unlock-ais-full-potential-at-work (accessed Apr. 15, 2025). https://medium.com/%40dirsyamuddin29/how-ai-is-fueling-efficiency-lessons-from-shells-gas-industry-transformation-3e754d4e7ff8 https://medium.com/%40dirsyamuddin29/how-ai-is-fueling-efficiency-lessons-from-shells-gas-industry-transformation-3e754d4e7ff8 https://medium.com/%40dirsyamuddin29/how-ai-is-fueling-efficiency-lessons-from-shells-gas-industry-transformation-3e754d4e7ff8 https://www.businesswire.com/news/home/20240708556035/en/AI-Driven-Project-Management-Sees-90-Positive-ROI-Among-Users https://www.businesswire.com/news/home/20240708556035/en/AI-Driven-Project-Management-Sees-90-Positive-ROI-Among-Users https://journalwjarr.com/node/555 https://planisware.com/resources/ai-ppm/planisware-enterprise-generative-ai-accelerating-project-definition-and-risk https://planisware.com/resources/ai-ppm/planisware-enterprise-generative-ai-accelerating-project-definition-and-risk https://planisware.com/resources/ai-ppm/planisware-enterprise-generative-ai-accelerating-project-definition-and-risk https://www.mdpi.com/2073-431X/14/2/66 https://www.mdpi.com/2079-8954/13/3/208 https://www.cloudtango.net/blog/2025/02/10/how-microsoft-copilot-is-transforming-medium-sized-enterprises/ https://www.cloudtango.net/blog/2025/02/10/how-microsoft-copilot-is-transforming-medium-sized-enterprises/ https://www.cloudtango.net/blog/2025/02/10/how-microsoft-copilot-is-transforming-medium-sized-enterprises/ https://www.cloudtango.net/blog/2025/02/10/how-microsoft-copilot-is-transforming-medium-sized-enterprises/ https://www.mckinsey.com/capabilities/operations/our-insights/project-delivery-hubs-analytics-for-better-megaproject-outcomes https://www.mckinsey.com/capabilities/operations/our-insights/project-delivery-hubs-analytics-for-better-megaproject-outcomes https://www.mckinsey.com/capabilities/operations/our-insights/project-delivery-hubs-analytics-for-better-megaproject-outcomes https://cloud.google.com/resources/state-of-devops https://cloud.google.com/resources/state-of-devops https://cloud.google.com/resources/state-of-devops https://www.mckinsey.com/capabilities/mckinsey-digital/our-insights/superagency-in-the-workplace-empowering-people-to-unlock-ais-full-potential-at-work https://www.mckinsey.com/capabilities/mckinsey-digital/our-insights/superagency-in-the-workplace-empowering-people-to-unlock-ais-full-potential-at-work https://www.mckinsey.com/capabilities/mckinsey-digital/our-insights/superagency-in-the-workplace-empowering-people-to-unlock-ais-full-potential-at-work https://www.mckinsey.com/capabilities/mckinsey-digital/our-insights/superagency-in-the-workplace-empowering-people-to-unlock-ais-full-potential-at-work