AI Takes On HR: Leaner Teams, Chatbots, and Who Gets the Gains

HR at an AI Crossroads: Keep People, Gain Productivity

During the pandemic, RingCentral hired 4,000 people to meet demand. Its HR team of 300 has since been cut by nearly half. Head of HR Alvin Lam says AI tools like the company's chatbot, Ringo, could help maintain service even with fewer staff. This is the trade-off many HR leaders face: where to apply AI, how fast to move, and how to protect the human core of the function.

What's actually happening in HR

Across 1.2 million US companies, just over 9 percent report using generative AI in production, and the number is climbing. CEOs at large firms like Salesforce, Amazon, and JPMorgan Chase have been clear: AI will remove some jobs and improve productivity. HR sits right in the blast radius-and the opportunity zone.

By January 2024, nearly two-thirds of organisations using AI in HR were applying it to talent acquisition, according to the Society for Human Resource Management. Leadership and development and performance management followed. Candidates use generative AI to produce polished résumés at scale; employers respond with AI screening to separate signal from noise.

Inside companies, AI chatbots now handle routine queries. At IBM, 94 percent of staff questions are resolved by AskHR, which began using generative AI in August 2024 to synthesize answers from policy documents. The outcome: fewer tickets, faster response, more time for higher-value work.

The lines you can't cross

HR's use of AI is governed by data protection rules and newer laws like the EU AI Act, which labels some HR systems "high risk" due to their impact on careers and rights. Review the high-risk criteria and documentation duties before piloting tools that affect hiring, promotion, or pay. EU AI Act overview

Litigation risk is real. In the US, Derek Mobley has sued Workday, alleging its screening algorithm discriminated on the basis of age, race, and disability; Workday says the case is without merit. HR tech leaders acknowledge the implications of such cases could be broad as more advanced tools roll out.

The ROI case-and the headcount reality

HR leaders are being asked to prove returns from AI. That means linking AI-enabled screening to quality-of-hire data and business outcomes, not just time saved. Job postings for HR roles have cooled more than the broader market since generative AI arrived, though causation is hard to prove.

IBM reports fewer people in HR than in 2016 and a 40 percent reduction in HR spend over four years, supported by consolidating nearly 9 in 10 systems. A McKinsey report places HR among the top four functions reporting cost reductions from generative AI in late 2023, ahead of marketing/sales and product development. McKinsey: State of AI 2024

Don't hollow out HR: a five-move plan

• Be strategic, not trendy. Map your talent and service priorities first. Select use cases that advance those goals, not just "what the tool can do."
• Start with support and admin work. Deflect FAQs via chatbots, automate workflows, and clean up policies. Reinvest the saved time into workforce planning, org design, and manager enablement.
• Engage senior leaders early. Set guardrails, agree on success metrics, and align on risk tolerance for hiring, promotion, and performance decisions.
• Expose cross-functional waste. Use AI to spot duplicated processes across HR, Finance, and IT (e.g., onboarding, access requests) and standardize the flow.
• Redesign the work. Don't bolt AI onto broken processes. Simplify steps, clarify decision rights, then apply automation and assistant tools.

A practical playbook for HR teams

• Audit the work. List top demand drivers (tickets, requisitions, policy searches). Quantify time, cost, and error rates.
• Pick two pilots. Example: Applicant screening triage and employee self-service for policies/benefits. Define a 60-90 day timeline.
• Set metrics up front. Time-to-fill, quality-of-hire at 6/12 months, deflection rate, SLA adherence, employee satisfaction, bias metrics.
• Build guardrails. Human-in-the-loop on high-stakes decisions, adverse impact testing, data retention rules, prompt/content logs, vendor due diligence.
• Upskill your HR staff. Train on prompt quality, policy retrieval, bias testing, and exception handling. Tie skill growth to career paths and pay.
• Communicate clearly. Tell candidates and employees where AI is used, what data is processed, and how to appeal decisions.
• Close the loop. Compare AI vs. human outcomes. Keep the winner. Scale with documentation and regular audits.

If you need structured training paths for HR teams adopting AI, explore curated programs by role and skill at Complete AI Training.

What stays human

Expect automation to touch most HR tasks; some leaders forecast up to 80 percent of functions will be automated over time. The final slice-employee relations, sensitive terminations, senior coaching, culture shaping-stays with people, often by choice, not technical limits.

Vendor and governance questions to ask

• What data do you use for training and inference? How is it segregated and secured?
• Show adverse impact testing and explainability for high-risk use cases. How often do you re-test?
• How are prompts, outputs, and decisions logged for audits? What is the retention period?
• What is your incident response plan for model drift, hallucination, and data leakage?
• Can we run a head-to-head trial against our baseline and own the evaluation data?

The decision HR must make

You can let AI cuts happen to your team, or you can direct the change and share the gains with employees through better work, skill premiums, and shorter low-value workloads. The companies that win will do both: protect the human core and prove the math on productivity. That's the mandate-and the moment-for HR.