Where Agentic AI fits in HR according to Jeff Smith BlackRock Alum

Harnessing AI sits at the top of the 2026 priority list in Gartner’s survey of 426 CHROs across 23 industries, and the technology in question is no longer a chatbot. Agentic AI describes autonomous systems that plan and execute multi-step HR workflows, from onboarding provisioning to leave processing, without waiting for a human prompt at each step. Deloitte’s research names the risk in the rollout. Among 100 C-suite leaders studied, 59% are taking a tech-focused approach to AI, and those organizations are 1.6 times more likely to miss the returns they expected than peers taking a human-centric approach. Buying autonomy is easy; deciding what deserves it is the hard part. Jeff Smith, BlackRock’s former HR leader, who spent more than a decade leading human resources there after senior roles at Time Warner and AOL, has argued for years that technology’s job is augmentation. “In its best form, it will enhance decision making through providing insights and data analysis in a faster, more efficient way to ideally be processed and used by humans,” he says of AI. “AI should augment human abilities.” That principle draws a boundary through the middle of the HR department. From answering questions to running workflows Earlier HR automation waited for instructions. Agentic systems accept an outcome, then sequence the steps themselves: a new-hire agent that provisions accounts, schedules orientation, assigns training, and chases the missing tax form without a coordinator touching any of it. Josh Bersin’s task-level analysis of more than 250 HR jobs finds 30% to 40% of existing roles can be automated with relatively low effort, concentrated in scheduling, records, employee services, and content delivery work. He projects employee-to-HR staffing ratios stretching from the traditional 100:1 toward 300:1 or beyond as agents absorb the tactical layer. The capability gap underneath the ambition Enthusiasm is outrunning readiness by a wide margin. Just 5% of HR teams feel fully prepared to implement AI effectively, and 40% of CHROs identify insufficient AI knowledge within their own teams as the biggest obstacle to integration, even as nearly half of them prioritize AI investment over the next two years. Gartner’s 2026 CHRO research points to where the gains actually live: redesigning the HR operating model carries the highest predicted impact on AI productivity, at 29%. An agent dropped into an unchanged process automates the process’s existing flaws at machine speed. That is the precise failure mode Deloitte’s human capital trends survey of more than 9,000 leaders warns about when it urges organizations to redesign work around human-machine collaboration rather than bolting machines onto old workflows. The Jeff Smith BlackRock augmentation line Smith’s augmentation principle sorts HR work by a single question: does the task require judgment about people, or execution around them? His longstanding technology brief covers the second category. “It is critical to have exceptional technology to make processes better and more efficient, for culture and risk management, and to help provide data and insight to make decisions,” he has argued. Execution around people is where agents earn their keep: Provisioning and onboarding logistics: Accounts, equipment, scheduling, and compliance steps that follow a known sequence every time. Leave and records processing: High-volume, rules-based administration where speed and accuracy beat discretion. First-line employee services: Policy questions and routine requests, with clean escalation when a question stops being routine. Data assembly: Pulling, joining, and summarizing workforce data so a human starts at the insight instead of the spreadsheet. What has to stay with people Judgment, ethics, and exceptions Performance evaluation, compensation calls, terminations, accommodations, and investigations carry consequences for individual lives and legal exposure for the company. Smith’s line holds even when an agent could technically run these flows, because the point of augmentation is better human decisions, not fewer humans deciding. He has also said job displacement deserves ethical scrutiny. Culture and connection Culture work resists delegation for a simpler reason: it lives in human contact. “It is easier to have random moments of learning or mentoring or coaching or ideation and innovation if people are walking the halls and interacting with each other,” Jeff Smith, BlackRock alum, has observed. An agent can schedule the mentoring session. It cannot be the mentor, and a department that forgets the difference automates away the relationships its influence depends on. The HR job after the handoff Bersin already documents coordinators and schedulers becoming agent builders and managers, and Korn Ferry projects an entire new category of AI-manager roles connecting business problems, culture, and machine output. Oversight becomes the job: setting the outcomes agents pursue, auditing what they produce, owning the escalations, and maintaining the data quality that keeps autonomous systems truthful. Accountability sharpens at the same time, because an error a human clerk once made occasionally, an agent operating at scale can make thousands of times before anyone notices. Sequencing matters as much as sorting. Reversible, rules-heavy workflows like leave processing make better first deployments than anything touching pay or performance, because early errors surface where they can be corrected quietly. Trust compounds the stakes: employees judge HR automation against their own paycheck and their own leave request, and one mishandled case erases the goodwill a thousand fast transactions built. Jeff Smith, BlackRock veteran, attaches a prerequisite. “I think getting the basics right and executing them is far more important before you are innovating,” he says. Agents inherit the process they are given. Clean processes, accurate data, and clear expectations are what make autonomy safe to grant. A boundary worth drawing slowly HR departments will spend 2026 under pressure to hand agents everything measurable, and the winners will be the ones that sort deliberately. Routine execution goes to the machines, with oversight designed in from the first deployment. Judgment, ethics, and culture stay with people, staffed and funded accordingly. The principle remains the one Jeff Smith, BlackRock’s former HR leader, has repeated across a decade of technology cycles: AI should augment human abilities. A department that holds that line gets faster workflows and better decisions at the same time. A department that crosses it discovers, one automated misjudgment at a time, why the human parts of human resources were never the inefficiency. DISCLAIMER – “Views Expressed Disclaimer – The information provided in this content is intended for general informational purposes only and should not be considered financial, investment, legal, tax, or health advice, nor relied upon as a substitute for professional guidance tailored to your personal circumstances. The opinions expressed are solely those of the author and do not necessarily represent the views of any other individual, organization, agency, employer, or company, including NEO CYMED PUBLISHING LIMITED (operating under the name Cyprus-Mail).