Meta canceled a second round of layoffs after AI agents failed to match human productivity, with code changes up 220% but user-facing features only 36% higher, while Microsoft found AI deployment yielded no business improvement and cut internal AI budgets by over a third, amid surveys showing 36% of large firms suffered material AI incidents.
The article analyzes why individual AI productivity gains (80%) fail to translate into organizational improvements (6%), attributing the gap to bottleneck shifting per Theory of Constraints, AI's inability to grasp company-specific tacit knowledge, employee resistance, and mismatched hierarchical structures, then proposes end-to-end process redesign, knowledge explicitization, and cultural transformation as solutions.
Even China's highly industrialized short drama industry—the world's largest—cannot achieve full automation, yet many believe AI programming and infrastructure can be fully automated; the article cites high AI project failure rates and organizational barriers to argue that cognitive biases from vendor hype drive unrealistic expectations.
Analysis of Fed, McKinsey, MIT, and academic studies shows AI delivers measurable task-level productivity gains (14–55%) but enterprise-level EBIT impact remains concentrated among few firms that redesign workflows, while macro data shows profit gains driven by AI supply chain, not broad efficiency.
This article analyzes how AI automation shifts human work from task execution to responsibility-bearing roles, defining four new responsibilities — Result Owner, Professional Judge, Digital Capability Manager, and Frontline Discoverer — and outlining a four-step organizational transition path with real-world case studies.
AI AdoptionAI workforce transformationDigital Employees
The article analyzes why enterprise LLM deployments fail to yield measurable financial returns despite impressive demos, citing MIT research showing 95% of organizations see no ROI. It identifies three inflated expectations — demo-as-product, tool-as-organizational-change, and AI-as-requirement — and proposes "expectation engineering" practices: baseline metrics, can't-do lists, phased acceptance, and translating disappointment into problem lists.
AI AdoptionExpectation EngineeringExpectation Management
The article analyzes why individual productivity gains from AI coding tools fail to improve team delivery speed, using Huolala's year-long retrospective to reveal three pitfalls—unclear tool boundaries, context overload, and adversarial code review—and proposes standardizing tools, defining usage boundaries, marking AI-generated code, and integrating AI into workflows to turn personal gains into organizational efficiency.
A hotel case study shows that universal AI adoption accelerates unverified suggestions across departments; sustainable collaboration requires a named activity owner, explicit resource commitments, a shared protocol distinguishing suggestions from commitments, and a process to halt and re-verify when conditions change.
AI Adoptioncase studycross-functional collaboration
Accenture and MIT data reveal 88% of Chinese enterprises have moved past AI pilots but only 14% achieve significant value; the gap stems from missing business ontology, data quality, and organizational readiness — not model capability — and success requires evaluating AI against existing business KPIs, not new metrics.
Shopify CEO Tobi Lütke's mandate for universal AI adoption led to 'slop grenades'—AI-generated work that takes seconds to produce but hours for recipients to verify, while METR experiments show AI-assisted coding tasks took 19% longer due to hidden verification costs, revealing a productivity illusion where management sees gains but frontline workers bear the verification burden.
The article explains the divergent AI experiences using Andrej Karpathy's four-layer funnel model, showing how task complexity, domain verifiability, and access levels create vastly different perceptions of AI utility, and provides a practical framework for selecting verifiable tasks to test AI agents effectively.
Stack Overflow's 2024-2025 developer survey comparison reveals AI tool usage rose from 62% to 79%, but trust and satisfaction dropped, particularly among learners, while AI agents see early adoption, remote work shifts toward hybrid models, and developers still value human judgment and community learning over full automation.
Despite 70% of BI products adding AI chat features, 90% of deployed ChatBI projects see under 30% adoption because semantic gaps, inconsistent metrics, and missing data governance make NL2SQL unreliable; true intelligent analysis requires unified definitions, semantic layers, and organizational adaptation, not just better models.
The article argues AI's initial easy adoption in high-tolerance, online creative work is saturating, and the next phase requires deep integration with enterprise software infrastructure to handle cross-system SOPs, accuracy, and stability, giving established software companies an advantage over pure model providers.
Although 88% of enterprises now use AI, only 5.5% see a measurable EBIT impact; the gap stems not from technology but from outdated workflows and structures, and McKinsey, BCG, and Deloitte data show that without process and organizational redesign AI investments merely burn cash.
A Gallup 2026 survey of 120,000 employees shows that active manager involvement dramatically boosts AI tool effectiveness, while lack of leadership leads to poor adoption, highlighting that management practices—not the technology itself—are the key to successful AI integration.
Enterprises that roll out the AI productivity tool WorkBuddy to all employees often cannot quantify any cost reduction or efficiency improvement because six management gaps—data governance, process integration, system integration, requirement translation, incentive alignment, and strategic focus—prevent the tool from being embedded into real business value streams.
The article analyzes Zhou Hongyi’s internal memo to 360’s managers, showing that resistance to AI‑driven transformation typically comes from middle managers who fear loss of power, and proposes concrete steps—assessment metrics, new roles, and direct reporting—to turn them into change engines.
The article argues that avoiding AI incurs hidden costs far beyond money—lost time, missed opportunities, talent shortages, and weakened competitiveness—by showing how AI reshapes cost structures, delivers exponential business value, and creates efficiency, talent, and capital gaps across industries.
Kuaishou’s analysis reveals that increasing developer count reduces AI efficiency gains, identifies three layers of human friction, three organizational gaps, and four AI‑native principles, and demonstrates a four‑quadrant framework that boosted AI code generation and demand delivery by up to fourfold.
The article outlines a step‑by‑step framework for enterprises to adopt AI responsibly, emphasizing business‑first pilot projects, human‑AI collaboration, process redesign, organization‑wide training, and risk controls rather than costly, blanket technology purchases.
Global retail loses $1.7 trillion annually to inventory errors, while 74% of retail AI projects never leave the pilot stage; the article explains three core pain points in stores and shows how Biying Tech’s AI‑driven "digital employees" dramatically improve staffing, decision‑making, and supply‑chain efficiency.
AI AdoptionDigital EmployeeSupply Chain Automation
In the first half of 2026, major technology companies eliminated over 180,000 positions yet poured more than $320 billion into AI infrastructure, a shift the article dissects by detailing the task‑level automation logic, the four‑layer AI architecture, real‑world deployment cases, and the limits of current AI replacement.
The article outlines a typical five‑step pattern of AI adoption in companies—empty budgets, token teams, costly hardware, failed pilots, and scapegoating—highlighting that while AI projects mostly flop, the role of “AI transformation consultant” thrives, and the real risk is being blamed when projects collapse.
Amid a steep decline in chip stocks and growing AI‑bubble concerns, US firms such as Coinbase, Lindy, Airbnb and DoorDash are migrating to open‑source Chinese large models like Kimi K3, GLM 5.2 and Qwen, citing 60%‑plus token share, 80% of AI startups using them, and cost reductions of 60‑90% that reshape the global AI economics.
The article argues that AI has become a national strategic priority and a mandatory driver of enterprise transformation, outlining the shift from digitalization to digital intelligence, the anxieties CEOs face, and a seven‑point framework for leaders to become effective chief digital intelligence officers.
The 2026 report shows that while AI usage among small and medium enterprises is rising, most firms only use isolated tools; true transformation requires deep workflow integration, solid digital foundations, security readiness, and strategic process redesign.
Over the past year five AI product managers left their positions, revealing that merely knowing large‑model tech is insufficient; true success requires owning unique data, deep business understanding, controllable delivery, and demonstrable cost or revenue impact.
The article outlines four progressive stages of enterprise AI adoption—Tool AI for personal efficiency, Scenario AI for specific tasks, Process AI for end-to-end workflows, and AI-Native where AI becomes an integrated, governed part of business operations—emphasizing that true AI maturity depends on how deeply AI participates in business processes, not just model count.
A recent interview with two FDE partners reveals that fewer than half of enterprise AI projects succeed, with failures clustering around unrealistic expectations, misplaced budgeting, and misaligned employee incentives, and the article outlines concrete steps to avoid each trap.
AI AdoptionAI project managementEmployee incentives
Despite 95% of generative AI pilots failing to scale and only 6% of firms realizing real business value, most internal AI projects stumble because of data silos, workflow fragmentation, and misaligned KPIs; the article analyses these root causes and proposes a three‑layer data‑process‑governance framework to turn AI from a flashy demo into genuine productivity.
The article dissects AI vendors' bold claims that 80% of code is AI‑written, reveals how these volume metrics replace dismissed pseudo‑indicators, contrasts them with mixed research findings on productivity and quality, and warns that such numbers can drive budgets, OKRs, and even layoffs.
AI AdoptionAI code generationDeveloper Productivity
Most enterprises can only feed AI with less than 15% of their knowledge assets because 80% of their data is hidden in PDFs, emails and chat logs, making data governance the true competitive moat and requiring a three‑step approach to turn dormant data into AI‑ready fuel.
The article analyzes how enterprises have shifted from fearing AI underuse to worrying about AI misuse, identifies five critical gaps—knowledge, data, process, governance, and value—and presents a four‑type AI‑employee framework and an HR‑style management platform to turn AI into reliable, production‑grade staff.
The article analyzes how enterprise AI rollout transforms middle managers from traditional experience‑based overseers into architects of reusable organizational capabilities such as Skills and Agents, reshaping responsibilities, decision‑making and the overall structure of the company.
The 2026 AI Skills Shift analysis reveals a rapid surge in AI‑centric job postings, sizable salary premiums, widespread role reshaping, generational risks, and a strategic move from using AI as a tool to embedding it as a core capability in the workforce.
The article argues that most enterprise AI projects fail because leaders treat AI as a novelty to showcase rather than a strategic tool for business‑process redesign, citing real‑world cases of AI‑driven customer service and approval automation that increased complaints and missed cost‑saving goals.
Many people adopt AI tools but see little impact because their mindset and methods are misaligned; this article breaks down ten common cognitive gaps—from poor business judgment and ROI misunderstanding to inadequate scenario analysis, tool awareness, and expectation management—that determine whether AI truly adds value.
When planning an AI application, you must first ask why you want to build it, then assess the intertwined risks and opportunities, quantify them, evaluate cost‑benefit and competitive barriers, and answer the core question of the business threat and growth lever the AI can address.
The guide defines the AI Champion role as the internal catalyst who turns AI from a personal toy into a stable productivity lever, outlines six core responsibilities, required capabilities, success and failure case studies, and provides a detailed weekly‑to‑monthly practice framework for enterprise AI transformation.
A 60‑year‑old fisherman in Anhui won a prize of 100 million AI tokens, mistook it for a scam, and sparked a discussion about the official Chinese term “词元”, its nationwide adoption, massive token usage statistics, and what this reveals about AI penetration in everyday life.
The article examines why veteran employees find AI tools too hard while newcomers deem them cumbersome, identifies generational language gaps as the root cause, and presents a three‑step “dual‑track collaboration” protocol—including tailored prompt libraries, cross‑generational communication scripts, and a weekly SOP—to bridge the divide and accelerate AI adoption.
The report shows AI entering a "Frontier Firm" era, with organizations moving from pilots to enterprise‑wide deployments, 82% of leaders targeting 2026 for strategic transformation, and a rapid rise of digital labor agents that create capacity gaps, reshape job structures, and raise governance challenges.
Despite a surge of AI‑savvy “super‑individuals” who double or triple personal output, most firms see no rise in overall efficiency because 59 % pursue a technology‑first strategy that raises failure risk by 1.6 times, highlighting the need for organization‑first reforms, collaborative incentives, and dedicated AI capability units.
The article argues that traditional digitalization has hit its ceiling and only firms that treat AI as a core productivity engine—led personally by CEOs, aligned on strategy, and executed through fast wins, robust mechanisms, and AI‑native organization—can avoid collapse and achieve a second growth curve.
The article recounts a failed enterprise AI rollout, analyzes why strong tools were ignored—fear of responsibility, learning cost, and lack of psychological safety—and offers a three‑step protocol with low‑barrier prompts and conversion scripts to break the ice and boost adoption.
The article explains, using the diffusion S‑curve, loss‑aversion, status‑quo bias, and capital constraints, why ordinary people systematically fail to become early adopters of emerging technologies such as AI, showing that structural factors—not lack of information—exclude them from high‑return periods.
The article argues that the failure of many enterprises to benefit from AI is not due to a lack of technology but to insufficient digital foundations, disorganized processes, poor data quality, cultural resistance, and a shortage of skilled talent, turning AI projects into costly showpieces.
AI AdoptionOrganizational ChangeProcess Optimization
In the AI era, the author argues that early, hands‑on adoption is essential—not for hype but to avoid distorted judgments, build repeatable workflows, and develop lasting capabilities that outlast rapidly changing tools.
McKinsey’s report shows that agentic AI, built on large models, can take on about 60% of marketing tasks—automating content creation, audience testing, and media planning—while boosting revenue 10‑30%, increasing execution speed 10‑15×, cutting costs, and outlining a five‑step workflow transformation with associated risks and governance recommendations.
The author, Mitchell Hashimoto—co‑founder of HashiCorp and creator of Terraform and Ghostty—shares a step‑by‑step, candid account of adopting AI agents, detailing six phases from abandoning chatbots to continuously running agents, the concept of “harness engineering,” and practical insights on when and how to integrate AI into a developer workflow.
The 2026 AI Index Report shows that generative AI has been adopted by 53% of the global population within three years, delivering $172 billion in annual consumer value, reshaping productivity, science, medicine, and education while exposing job shifts and policy gaps.
The 2026 Stanford AI Index report, over 400 pages long, reveals that AI resources keep growing while model concentration rises, the performance gap between US and Chinese models has nearly vanished, top‑tier models are converging, 88% of enterprises have adopted generative AI, and concerns mount over transparency, environmental impact, and uneven labor‑market effects.
The article analyzes the current AI hype, highlights common misconceptions about AI replacing jobs, explains the true boundaries of AI capabilities, and offers practical advice for leaders on how to integrate AI into business processes without falling into costly missteps.
Despite the rapid maturity of AI coding assistants like Cursor and GitHub Copilot, many companies see less than 20% active usage because they treat AI as a checklist item rather than redesigning workflows, culture, and documentation to create a true human‑AI partnership.
Despite hype, many AI features see rapid drop‑off after launch; this article shows product managers how to avoid building AI for its own sake by rigorously defining scenarios, aligning user needs, assessing risks, and applying a five‑step framework to ensure AI delivers real value and sustained engagement.
Enterprises often find themselves stuck in a paradox where small AI use‑cases are dismissed as insignificant while large AI initiatives are deemed too risky or costly, revealing deep misalignments in digital transformation strategy, organizational structure, and value assessment mechanisms.
A McKinsey‑based analysis reveals that while 88% of enterprises have deployed AI, 86% of leaders admit their organizations aren’t ready for large‑scale adoption, and only 15% see revenue growth from generative AI, prompting a strategic shift for CTOs/CIOs from tech caretakers to value architects in the 2026 AI era.
Enterprises seeking AI‑driven efficiency often stumble into five common traps—poor selection, perfectionism, over‑control, fighting AI in its strong suits, and unvalidated delivery—each dramatically cutting ROI unless a disciplined, human‑centric process is applied across the AI lifecycle.
NVIDIA's 2026 AI State Report shows AI moving from optional projects to essential enterprise infrastructure, with 64% of firms already using AI, clear revenue growth and cost‑reduction benefits, rising budgets, open‑source adoption, and persistent challenges around data, talent, and ROI measurement.
The article analyzes how breakthrough AI models in late 2025 turned code generation from a supplemental tool into a primary developer, reshaping software engineering value, demanding new engineering guardrails, redefining roles, and outlining eight practical checks to safely adopt AI coding at scale.
An AI product manager discovers that chasing model hype without understanding the business creates embarrassment, and learns to map SOPs, align AI with real metrics like GMV, and turn AI into a profit‑driving tool rather than a self‑indulgent showcase.
Surveys reveal that while 72% of developers use AI tools daily and AI‑assisted commits have risen to 42% with a 76% boost in code output, 96% still distrust AI‑generated code, only 0‑20% of tasks can be fully delegated, and 27% of the new output consists of work that previously didn’t exist.
The article analyzes why product managers are tougher to recruit than project managers, highlighting higher skill demands, broader responsibilities, the need for cross‑disciplinary expertise, rapid market changes, and the accelerating impact of AI on product development workflows.
The article analyzes rapid AI advancements, outlines five stages of technological impact, presents concrete metrics of model efficiency, and offers practical strategies for professionals to adapt, emphasizing the urgency of embracing AI tools across all knowledge‑based industries.
This article outlines a step‑by‑step framework for discovering genuine AI‑applicable use cases, covering common pitfalls, a four‑stage screening process, demand validation techniques, AI‑fit assessment, and ROI evaluation to ensure the chosen scenario is both solvable by AI and worth the investment.
A McKinsey survey of 200 M&A professionals shows that teams using Generative AI cut costs by about 20%, accelerate deal cycles by 30‑50%, and value its ability to deliver deeper insights, streamline processes, and speed transactions, while outlining current usage, desired features, future outlook, and practical steps for adoption.
McKinsey’s report reveals that while AI tools are abundant, the real competitive edge lies in cultivating domain‑owner leaders who combine deep business expertise with technical acumen, and it offers concrete case studies and six actionable steps for organizations to develop this “second muscle.”
OpenAI’s 2025 State of Enterprise AI report, based on over one million enterprise customers and 9,000 employee surveys, shows AI usage exploding across companies—with ChatGPT Enterprise messages up 8×, token consumption per user up 320×, significant productivity gains for 75% of employees, and industry adoption growing 6‑11×, highlighting a widening gap between AI leaders and laggards.
A new MIT Sloan‑BCG report reveals that agentic AI, unlike traditional or generative AI, acts as an autonomous teammate that plans, executes, and learns, prompting leaders to overhaul workflows, governance, roles, and investment models to capture its strategic value.
Despite soaring user numbers and investment, AI has yet to produce a truly indispensable consumer app because most products merely overlay old workflows, suffer from weak entry points, modest experience gains, trust and cost barriers, and developers focus on technology rather than solving concrete, high‑frequency pain points.
McKinsey’s latest report reveals that Gulf Cooperation Council (GCC) countries are heavily investing in AI infrastructure, achieving high adoption rates, but only a small fraction of firms have scaled AI projects to generate measurable value, highlighting gaps in strategy, talent, and change management.
The article breaks down AI product management into three distinct tracks—business‑oriented, platform‑focused, and technical—detailing their core responsibilities, required skills, typical industry examples, recruitment profiles, transition pathways, and future demand trends, helping professionals choose the right career direction.
A McKinsey‑QuantumBlack 2025 report reveals that while 88% of firms now use AI and AI agents are gaining traction, only a small fraction achieve enterprise‑scale deployment or measurable EBIT gains, highlighting a gap between AI hype and real business transformation.
The article examines the hype versus real value of AI tools, compares different models and platforms, shares concrete usage scenarios, and offers practical recommendations for selecting models, mastering prompt engineering, building personal AI agents, and adopting an AI‑first mindset.
The article examines how AI's rapid adoption in 2025 workplaces, illustrated by Kevin Cantera’s experience of boosting productivity only to be replaced during layoffs, reveals a deeper power shift where tools dictate processes, employees must master AI fundamentals, and organizations face mixed efficiency outcomes.
A comprehensive report based on nearly 5,000 technology professionals reveals how AI tools act as amplifiers of existing team practices, outlines seven distinct team profiles, and provides data‑driven strategies for leaders to turn AI adoption into measurable performance gains.
The 2025 DORA report reveals that 90% of software professionals now use AI daily, boosting productivity across tasks such as coding, testing, and documentation, while also exposing trust challenges and highlighting the need for cultural and process changes to fully harness AI’s transformative potential.
The Anthropic Economic Index report reveals that AI is being adopted at unprecedented speed, but its early diffusion is highly uneven across regions, enterprises, and task types, highlighting potential economic inequality and workforce impacts.
The 2025 Enterprise AI Agent Research Report reveals that domestic Chinese agents excel in localized tasks and data precision, while international agents dominate in generalization, speed, and iterative efficiency, highlighting six critical adoption metrics and showcasing diverse industry case studies that illustrate the current AI Agent landscape and future opportunities.
The article analyzes why many workers reject AI adoption, examining technological maturity limits, lack of best‑practice guidance, human‑nature learning barriers, and misaligned incentives, while illustrating points with real‑world examples and personal observations from a podcast discussion.
The article introduces the AI Adoption Maturity Model (AIM²), a five‑level framework covering strategy, organization, data, technology, application, and business, and showcases industry case studies that illustrate how enterprises can assess their AI readiness, identify gaps, and accelerate AI‑driven transformation.
This article explains how Siemens harnesses generative AI to create an integrated enterprise knowledge chatbot, detailing the challenges of fragmented data, the “XiaoYu” solution, its ecosystem, and the broader impact of AI on productivity and innovation across the company.
The 2025 Microsoft AI Education Report, based on surveys of 1,851 educators, leaders, and students across six countries, reveals that 86% of education institutions use generative AI, highlights how learners and teachers leverage AI for tasks, discusses emerging concerns, and offers strategic recommendations for integrating AI responsibly in schools.
The McKinsey report outlines a five‑step framework for CEOs to turn generative AI pilots into real business value by setting a clear North Star, establishing data trust, redesigning workflows, blending autonomous and augmented teams, and empowering employees as change agents.
According to Stack Overflow’s 2025 developer survey of 49,000 respondents from 177 countries, AI tool usage has surged to 84% with many relying daily, yet trust in AI outputs has plummeted, revealing growing skepticism, productivity challenges, and shifting attitudes toward AI agents, remote work, and salaries.
The AI Adoption Survey H1 2025 reveals that nearly half of organizations have deployed AI in production, engineering and R&D lead usage, Chinese LLMs gain overseas interest, and cost, reliability and intelligence remain the top challenges, while tool preferences and multimodal trends reshape the market.
McKinsey’s latest State of AI report reveals that 78% of surveyed firms now use AI in at least one business function, with generative AI adoption rising to 71%, and highlights that CEO governance, workflow redesign, and risk mitigation are critical drivers of measurable EBIT impact, especially for organizations with revenues over $5 billion.
The article explains what AI intuition is, why many overlook AI opportunities, and offers practical steps—starting with hands‑on use, integrating AI into daily testing work, and continuously refining your sense of AI’s capabilities—to boost efficiency and stay competitive.
Mary Meeker’s 340‑page AI trend report highlights five core predictions—including unprecedented AI adoption speed, AI becoming a public utility, the rise of Chinese models, AI’s integration into the physical world, and the emergence of AI‑native users—while warning investors of the high risk of AI projects.
In this interview, an automotive industry expert explains how large language models and advanced operations‑optimization techniques are reshaping vehicle design, production planning, logistics, and customer services, while also discussing implementation challenges, team requirements, and future AI‑driven opportunities.
Amid a global digital surge, 45% of China’s central state‑owned enterprises have deployed the DeepSeek large‑model platform, rapidly integrating AI across energy, power, telecom, construction and other sectors to boost intelligent transformation and operational efficiency.
In this roundtable, moderator Garry and guests including Box CEO Aaron Levie discuss the current AI revolution, the role of model companies, opportunities for startups, pricing models, enterprise adoption, security concerns, and the broader economic impact of AI on businesses and society.
Meta plans to adopt AI for large‑scale code generation by 2025, a move that could reshape software development, spark industry‑wide AI adoption, and force programmers to transition from writing code to collaborating with and managing intelligent coding systems.
The article examines Harvey’s rapid rise to a $3 billion valuation, its AI‑driven legal products, strategic positioning as an “AI assistant” with scenario‑specific solutions, market‑entry tactics, and broader insights on selecting high‑value, AI‑friendly use cases for legal AI.
This report examines the technical evolution of large AI models, outlines industry demand, explains how these models add value across sectors, and discusses future challenges and considerations for large‑model‑driven transformation in a wide range of businesses.
AI Adoptionartificial intelligenceindustry applications
The 2023 Software Quality Assurance industry survey, based on nearly a thousand valid responses, analyzes testers' employment status, satisfaction, background, technical skills, personal growth paths, team structures, AI/LLM usage, and future industry trends, providing data‑driven insights for professionals and managers alike.
The article analyzes how large‑model AI is reshaping medical practice, highlighting rapid technology adoption but significant implementation hurdles due to physician behavior, data silos, safety regulations, talent gaps, and differing maturity across R&D, manufacturing, marketing, and customer‑service domains.
The article analyzes how large‑model medical AI is rapidly adopted yet struggles with implementation due to doctor shortages, behavioral resistance, data silos, safety regulations, and the need for strategic alignment, while contrasting the more supportive innovation ecosystem in the United States.
CoderPad’s 2024 State of Tech Hiring report, based on responses from over 13,000 developers and 5,500 HR professionals across 149 countries, uncovers developers’ expectations, AI adoption, preferred languages and frameworks, hiring priorities, job security concerns, and the growing debate over AI use in recruitment.