AI Takes the Front Line: Orchestration, Upheaval, Risk, and Readiness

Your go-to rundown on AI’s impact on the future of work

Author

Henry Nothhaft, Jr.
June 06, 2025

Your go-to rundown on AI’s impact on the future of work—delivered every Friday. Each edition highlights three to five must-read stories on everything from job disruption and upskilling to cultural shifts and emerging AI tools—all in a crisp, Axios-style format.

Was this newsletter forwarded to you? Sign up to subscribe.

In the wake of the week…

Agentic AI is moving fast from concept to capability, from chatbot to co-worker. Our lead story charts its emergence as a new operating layer in customer experience. But that evolution has consequences. Vista’s CEO predicts a major shakeout in white-collar jobs, like tomorrow, and new research shows most companies are still scrambling to build the workforce alignment and trust needed to navigate what comes next. Meanwhile, AI hallucinations continue to slip into critical operations, from support queues to courtrooms, becoming a growing liability for enterprises without proper safeguards.

But not all signs point to upheaval. New research from APQC lays out a clear foundation for AI success in HR, emphasizing the value of starting small, focusing on real outcomes, and building the muscle memory that makes transformation sustainable. And in Extra Credit, Trip.com’s CEO offers a practical example of how AI and talent development can work in tandem, not at odds.

Let’s dive in. 👇

Agentic AI Becoming the New Front Line for Customer Experience

Agentic AI is moving beyond automation and into orchestration of customer experience. Unlike traditional chatbots that respond to individual queries, agentic AI takes initiative, learning, reasoning, and acting across systems to resolve issues and improve outcomes without waiting for a prompt.

Research from Cisco shows that enterprise buyers now expect more than half of all CX interactions to be handled by AI agents within 12 months. The trend is accelerating as platforms like Salesforce deepen their data intelligence capabilities through acquisitions such as Informatica, while vendors like Quiq build agentic systems that operate with human-like autonomy.

The new model is personalized, proactive, and increasingly predictive. AI agents can summarize issues, resolve problems before escalation, and deliver real-time insights and recommendations tailored to each customer.

By the Numbers

  • 56% of enterprise CX interactions are expected to be handled by agentic AI within one year (Cisco)

  • $25.73 billion projected market size for hyper-personalization by 2025 (Thompson)

  • 93% of enterprise buyers believe agentic AI will improve service quality and relevance (Cisco)

  • 52% expect agentic AI to lead to increased customer spend

  • 86% believe agentic AI will help vendors become long-term strategic partners

Why It Matters

AI is influencing customer expectations. They want services that anticipate needs, reduce friction, and respond in context. Agentic AI supports this evolution by applying memory, reasoning, and real-time decision-making across complex workflows. The impact is already visible in areas such as onboarding, support automation, and post-sale engagement.

Leadership Insight

Customer experience leaders should think of agentic AI as more than just a smarter chatbot. It is a system that can manage real tasks across the customer journey, tracking issues, solving problems, and coordinating follow-up across tools and teams. To get started, focus on one or two high-friction areas where better responsiveness or personalization would move the needle. When designed well, agentic AI handles the repetitive work and frees up your people to focus on high-value interactions that strengthen customer relationships.

Thanks to a growing ecosystem of low-code and no-code platforms, teams can begin experimenting with agentic AI without heavy engineering support. Tools like CustomGPT, Cognosys, and Relevance AI offer drag-and-drop interfaces for designing task-driven agents that integrate with customer data, CRM systems, and third-party APIs. For CX leaders, this means proof-of-concept pilots can be launched quickly and refined over time, bringing AI from concept to execution without waiting on a full platform overhaul.

The Bottom Line: Agentic AI is raising the stakes in customer experience. The opportunity lies in using it to elevate service quality and customer value. Leaders who start with clear use cases and accessible platforms can move quickly, deliver results, and lay the groundwork for more intelligent and responsive service.

Upheaval Ahead: AI Will Force a Talent Shakeout, Warns Vista CEO

If agentic AI is becoming the new operating layer, as we just explored, then this is the disruption that follows. At the SuperReturn private equity conference, Vista Equity CEO Robert F. Smith offered a stark forecast: “Next year, 40% of this crowd will have an AI agent. The other 60% will be looking for work.” AI was the dominant theme as investors evaluated how to drive value, often through automation, across portfolios.

Smith’s prediction aligns with warnings from Anthropic’s Dario Amodei, who says up to half of all entry-level white-collar roles could vanish within five years. Some leaders, like Thoma Bravo’s Orlando Bravo, see a productivity upside. But the underlying trend is clear: organizations are restructuring now, betting AI agents will soon be able to handle more than support tasks.

By the Numbers

  • 1 billion: Global knowledge workers Smith says will see job changes

  • 200,000: Banking jobs global firms may cut in 3–5 years (Bloomberg Intelligence)

  • 6%: Unemployment among recent college grads, up from prior years

  • 10–20%: Potential unemployment spike from AI, per Amodei

Why It Matters

Unlike past waves of tech disruption, this one is preemptive. Businesses are not waiting for AI to match human performance; they are reorganizing in anticipation of that day. That gamble may pay off for early adopters or backfire if AI falls short of expectations. Either way, the transition is already underway, and leadership must prepare for deeper shifts in workforce structure, morale, and trust.

Leadership Insight

Leaders must channel urgency into responsibility. Use moments like this as a prompt to strengthen workforce dialogue, clarify where AI adds value, and invest in skills and mobility, not just efficiency. The future of work should not be something done to employees, but with them.

The Bottom Line: AI is already transforming white-collar work. Business leaders are still debating whether that means replacement or augmentation. Real leadership requires a deliberate commitment to bring people along.

AI Hallucinations Pose Real Enterprise Risk

Generative AI models often generate confident, plausible-sounding responses that are completely false, what researchers call hallucinations. In low-stakes contexts, that tradeoff may be acceptable. In customer service, compliance, legal filings, and critical business workflows, it can create substantial risk.

Recent cases span the spectrum: an Air Canada chatbot hallucinated refund policies and cost the airline a court judgment. In courtrooms, AI-generated case law has triggered sanctions. In hiring, bots have quietly introduced unrealistic job requirements, drying up candidate pools. A single false recommendation, misquote, or citation can damage trust, trigger compliance violations, or undermine leadership credibility.

Despite industry promises, hallucination rates are not declining. In fact, recent leaderboard data shows newer “reasoning” models hallucinate more than earlier versions. Accuracy remains a lower priority than speed, cost, and scale for most providers.

By the Numbers

  • 33% hallucination rate for OpenAI’s o3 model when summarizing biographies

  • 48% hallucination rate for o4-mini, up from 16% in o1 (OpenAI)

  • 30+ legal cases in May 2025 involved AI-generated false citations (Charlotin)

Why It Matters

Generative AI is inherently probabilistic. That quality is what makes it flexible, creative, and fast. It is also why it can invent facts with confidence. Most hallucinations do not spark headlines. They slip into policy language, customer communications, and business decisions quietly eroding accuracy, credibility, and quality.

Leadership Insight

Executives must treat hallucinations like any other risk that is to be managed. The solution is not to ban AI, but to design guardrails that keep it in its lane. Use retrieval-augmented generation (connecting AI to your own vetted data like policies, product specs, or knowledge bases) to anchor outputs in trusted data. Set confidence thresholds and error escalation workflows. Keep a human in the loop to review or approve outputs and maintain clear documentation when AI is used in customer-facing or compliance-related outputs.

Critical thinking and internal verification practices are now core competencies. Do not delegate judgment to a system that was not designed and hasn’t been optimized for accuracy.

The Bottom Line: Hallucinations are part of how generative AI works. The risk is not in using AI, but in using it without oversight. Leaders should focus on clear guardrails, trusted data, and human checkpoints, especially where compliance and trust are on the line.

From Urgency to Execution: Laying the Groundwork for AI in HR

If this week’s lead story on agentic AI left you in a cold sweat, new research from APQC offers relief and a path forward. Based on a global survey of 600 HR leaders, the report identifies the foundational moves that separate AI-mature organizations from those stuck in pilot purgatory. The takeaway: success does not come from chasing the flashiest tools, but from building systematically around the right data, people, and strategy.

The most advanced HR teams are not trying to do everything at once. Instead, they are starting small in high-leverage areas like recruiting, performance management, and workforce analytics, then scaling with intention.

❝

Start where you can practically have success and demonstrate value.

APQC’s Elissa Tucker

By the Numbers

  • 127% ROI achieved by HR functions at the highest AI maturity level

  • 20 days shaved off time-to-fill in recruiting among AI-mature firms

  • $470,529 revenue per employee reported at optimized organizations, up from $234,762

  • 20.6 FTEs reduced per $1B in revenue via automation and AI

  • 95% retention rates achieved in some cases, up from 80%

Why It Matters

Agentic AI may be the destination, but this research clarifies the journey. The organizations seeing real value have avoided “shiny object syndrome” by aligning their AI strategy with practical business goals, existing data, and internal readiness. For HR leaders under pressure to “get AI in the mix,” this evidence shows that methodical beats reactive.

Leadership Insight

Treat AI readiness as a core capability. Build literacy, clarify your tech stack’s current AI functionality, and tie early wins to measurable business outcomes. The fastest way to fall behind, other than inaction, is to overreach and stall, losing hearts & minds in the process.

The Bottom Line: Sustainable AI transformation in HR depends on sequencing: data infrastructure, workforce preparedness, and strategic integration. Leaders who build on these foundations now will gain speed and scale later, without burning trust or budget.

Extra Credit

For the overachievers: These are the stories that didn’t crack the top group this week but are too important to ignore—quick hits on what’s happening and why it matters.

How Trip.com’s CEO Blends AI with Talent Strategy

Key Takeaway: Jane Sun, CEO of Trip.com, outlines how her company uses AI to enhance both customer experience and internal operations, from resume screening and training to call center support and personalization at scale.

Why It Matters: Trip.com’s approach reflects a growing reality: competitive advantage in AI adoption will come from combining operational automation with thoughtful talent development, not treating them as separate domains.

Source
HBR on Leadership: Building an AI-Powered, Talent-Friendly Organization [podcast]

Resistance Rising: CEOs Confront Employee Pushback on AI

Key Takeaway: Nearly half of CEOs report employee resistance or hostility toward AI, exposing a growing gap between executive ambition and frontline readiness.

Why It Matters: As our lead story emphasized the rise of agentic AI, this data serves as a stark reminder: strategy must be matched with workforce alignment. Without trust, training, and change management, even the best tools stall at implementation.

Axios Shares AI Survival Toolkit

Key Takeaway: Axios has compiled a practical toolkit for navigating the next wave of AI disruption, covering top tools, advanced prompting, and job risk zones, with input from CEOs, college students, and the public.

Why It Matters: As our lead story explored how agentic AI is reshaping work from the frontlines, this guide offers individuals a proactive way to build fluency, boost productivity, and prepare for what lurks on the horizon.

ChatGPT Expands Enterprise Features to Compete in Productivity Stack

Key Takeaway: OpenAI has rolled out connectors for Google Drive, Dropbox, and HubSpot, plus meeting recording and transcription tools, aimed squarely at competing with Microsoft, Google, Zoom, and Notion.

Why It Matters: After recent enterprise pushes from Microsoft and Google, OpenAI is shoring up ChatGPT’s value as an all-in-one AI assistant for knowledge workers. The platform’s growing integration and research capabilities signal a shift from chat interface to enterprise productivity layer.

This edition of The Big Shift: AI @ Work may have been edited with the assistance of ChatGPT, Claude, Copilot, Gemini, Perplexity, or none of the above.

Want to chat about AI, work, and where it’s all headed? Let’s connect. Find me on LinkedIn and drop me a message.