ServiceNow Knowledge 2025 Keynote: Bill McDermott and Jensen Huang on Agentic AI

⬅️ Back to Videos

📋 VIDEO INFORMATION

  • Title: 🎥 Video: ServiceNow Knowledge 2025 Keynote
  • Episode: “Put AI to work for people – now”
  • Host: Bill McDermott (Chairman & CEO, ServiceNow)
  • Guests: Kevin Clark (Chairman & CEO, Aptiv), Jennifer Morgan (CEO, UKG), Amit Zavery (President & COO, ServiceNow), Cindy Hoots (CIO, AstraZeneca), Jensen Huang (CEO, NVIDIA), Ben Gilbert & David Rosenthal (Acquired Podcast)
  • Duration: Approximately 1 hour and 23 minutes

📓 Watch the full keynote here: https://www.youtube.com/watch?v=yYnzj9LFYXA

📺 Watch here

🎣 HOOK

Bill McDermott opens ServiceNow’s largest conference ever by declaring that AI is “civilization’s greatest opportunity of this century” and a $22 trillion market that will eliminate $4 trillion in operating expenses by 2030. Yet he immediately confronts the brutal reality: 40% of office productivity vanishes as employees swivel-chair between 17 different applications daily, and only one in four digital transformation projects ever delivers positive ROI. The stakes? A $10 trillion “tax on inefficiency” strangling the US economy alone. This is a surgical dismantling of 20th-century enterprise architecture and the unveiling of what McDermott calls “the only platform built for the intelligence supercycle.”

💡 ONE-SENTENCE TAKEAWAY

ServiceNow positions itself as the enterprise “AI operating system” by orchestrating autonomous agents across siloed systems through a unified control tower, turning fragmented workflows into real-time, collaborative action while giving humans superpowers to focus on what matters most.

📝 SUMMARY

🔍 The Problem: Digital Dystopia

Bill McDermott frames the enterprise crisis as a “destruction of time”, that is the most precious asset. For 60 years, companies have battled the same enemy: siloed org structures and legacy architectures that create rigid, disconnected processes. The result is catastrophic waste, with knowledge workers losing five hours daily to mindless phone activity and office workers wasting 40% of their productivity app-switching. McDermott argues that 21st-century problems like geopolitical conflicts, skills gaps, global tariffs and more, cannot be solved with 20th-century architectures. The cost of inaction is a $10 trillion efficiency tax equivalent to 7% of US GDP.

🛠️ The Solution: ServiceNow AI Platform

The core announcement is the ServiceNow AI Platform, described as the “central nervous system of the 21st-century enterprise.” McDermott introduces a “north-south, east-west” architecture: north-south for full-stack integration from infrastructure to user interface, east-west for cross-functional workflows spanning IT, HR, Finance, CRM, and Security. The platform rests on three pillars: one platform, one architecture, one data model. At its heart is the Configuration Management Database (CMDB), providing a real-time, consolidated view of every enterprise asset.

🤖 The Agentic Layer

McDermott declares the traditional org chart will collapse as AI agents become digital workers. The platform introduces:

  • AI Agent Orchestrator: The only enterprise-wide agent orchestrator that ensures the right agent handles the right task at the right time
  • AI Control Tower: A unified command center to manage, govern, and secure all AI agents across the enterprise…even third-party agents
  • AI Agent Fabric: Native collaboration between agentic systems, replacing traditional APIs with intent-based communication

📈 Real-World Validation: Aptiv, UKG, AstraZeneca

Aptiv’s Transformation: Kevin Clark explains how the automotive OEM evolved into an edge intelligence company powering vehicles, aircraft, and 5G networks. ServiceNow automates 50% of their 200,000 employees’ service requests globally, and their supply chain digital twin will be fully mapped by year-end…turning procurement from 20-30 minute manual processes to seconds.

UKG’s Frontline Connection: Jennifer Morgan reveals that 80% of the workforce is frontline hourly workers, yet only 23% feel connected to enterprise technology. The ServiceNow-UKG partnership creates a “single point of interaction” connecting branch managers and frontline workers directly to enterprise systems, handling scheduling, pay questions, and customer service without ticket-based delays.

AstraZeneca’s Mission: Cindy Hoots shares how ServiceNow saved 90,000 hours in onboarding alone and reduced research supply requests from 20-30 minutes to seconds; critical when the company aims to deliver 20 new medicines by 2030. AI agents now monitor lab equipment, auto-procure supplies, prep sales reps with personalized doctor briefs, and even navigate insurance approvals for patients via cross-organizational agent communication.

The NVIDIA Partnership & Technical Core

Jensen Huang joins stage to announce Nemotron April, a new reasoning model co-built with ServiceNow that can read multi-modal PDFs, break down complex problems, and conduct autonomous research. Huang explains that NVIDIA chose ServiceNow because “every company needs AI agents” and ServiceNow is already “in every company’s IT department.” Their six-year partnership integrates NVIDIA’s AI stack with ServiceNow’s workflow data fabric, creating a “data flywheel” that continuously improves agent performance on enterprise data.

The Human Element

Despite the technology focus, McDermott repeatedly emphasizes: “There is no artificial intelligence without human intelligence.” He introduces ServiceNow University; a free, gamified, AI-powered learning featuring Idris Elba, aiming to empower 3 million AI learners by 2027. The vision is humans moving “up the chain” to focus on creative, strategic work while agents handle complexity and chaos.

🧠 INSIGHTS

Core Insights

  • Agentic AI Requires Platform Monogamy: McDermott argues that AI’s full potential is only realized when it spans “every system, every department, every function” making point solutions and multi-platform strategies obsolete. The ServiceNow AI Platform is positioned as the inevitable convergence point for enterprise software.

  • The North-South, East-West Architecture: This framework reveals ServiceNow’s strategic depth. East-west integration eliminates functional silos (IT, HR, Finance); north-south integration replaces the entire stack from infrastructure to UX. This isn’t workflow automation; it’s enterprise re-architecture.

  • AI Control Tower as the New Org Chart: The Control Tower doesn’t just manage agents; it becomes the operating system for human-AI collaboration. McDermott explicitly states organizations will onboard, measure performance, and govern AI agents like human workers, fundamentally restructuring how teams operate.

  • Real-Time Empathy at Mass Scale: The keynote’s emotional peak is AstraZeneca’s patient journey, where an AI agent resolves insurance denial in seconds by autonomously communicating with pharmacy, doctor, and insurance agents. This demonstrates agentic AI’s ability to remove complexity while preserving human dignity “I didn’t have to disrupt a single minute of my life.”

  • The Trust Currency: Both McDermott and Huang emphasize their six-year partnership succeeded because “trust is the ultimate human currency.” In an era of AI skepticism, they model how deep technical integration requires human alignment first, with hundreds of engineers working shoulder to shoulder as “one company.”

How This Connects to Broader Trends/Topics

  • The End of Best-of-Breed: The keynote directly attacks the “patchwork of point solutions held together with duct tape” that defines modern enterprise software. By acquiring Moveworks (employee experience) and Logic.ai (CPQ), ServiceNow is executing a consolidation play that mirrors Microsoft’s strategy of betting that AI makes integrated suites superior to specialized tools.

  • Frontline Worker Empowerment: Jennifer Morgan’s UKG segment highlights a critical societal shift: connecting the 80% hourly workforce to enterprise AI. This addresses the “great disconnection” where field employees lack technology access, potentially reducing turnover and improving customer experience; a $1 trillion opportunity in frontline productivity.

  • Regulated Industry AI Adoption: AstraZeneca’s Cindy Hoots explicitly frames AI governance as an ethical imperative, noting they were the first pharma company to publish an AI ethics statement. As healthcare, finance, and manufacturing adopt agents, ServiceNow’s emphasis on compliance and structured workflows becomes a competitive moat against startup disruptors.

  • The Reasoning Model Race: Huang’s announcement of Nemotron April positions NVIDIA against OpenAI’s o1 and Google’s Gemini in the reasoning wars. But unlike consumer models, April is purpose-built for enterprise ambiguity; able to read PDFs, conduct multi-step research, and operate within ServiceNow’s security model, potentially leapfrogging general-purpose LLMs for B2B use cases.

🏗️ FRAMEWORKS & MODELS

The ServiceNow AI Platform Architecture

McDermott’s “north-south, east-west” metaphor describes a platform that:

  1. East-West Integration: Connects every business function (IT, HR, Finance, CRM, Legal, Security) on a single data model, eliminating silos
  2. North-South Stack: Unifies infrastructure, data, workflow, AI, and user experience layers, replacing legacy systems
  3. CMDB Core: Provides a real-time, unified view of all enterprise assets as the “single source of truth” for agent orchestration

AI Agent Orchestration Framework

Amit Zavery details a three-layer system:

  1. AI Agent Orchestrator: The “enterprise-wide conductor” that assigns tasks, monitors performance, and ensures handoffs between specialized agents
  2. AI Control Tower: A unified command center providing governance, security, and compliance monitoring across all agents (ServiceNow and third-party)
  3. AI Agent Fabric: Enables agents to communicate natively using intent-based protocols (MCP, A2A), replacing traditional API integration with conversational interoperability

The Data Flywheel Model

Jensen Huang describes the continuous improvement cycle:

  1. Onboard: Deploy agents on ServiceNow’s workflow data fabric
  2. Train: Agents learn from enterprise data via zero-copy access
  3. Evaluate: Measure agent performance on real workflows
  4. Improve: Refine models (like Nemotron April) on proprietary data
  5. Deploy: Scale improved agents across the enterprise

💬 QUOTES

  1. “AI is civilization’s greatest opportunity of this century. And it’s the gateway to prosperity and the absolute requirement for survival.”

  • Bill McDermott, framing the stakes for enterprise leaders

  1. “The greatest dilemma of the digital age is the destruction of time. Think about it. Time is our most precious asset and on average, people waste five hours a day on mindless activity on their phone.”

  • Bill McDermott, diagnosing the productivity crisis

  1. “21st century problems cannot be solved with 20th century architectures.”

  • Bill McDermott, justifying platform consolidation

  1. “There is no artificial intelligence without human intelligence. Humans are now moving up the chain because the thinking machines handle the complexity in the chaos that humans never wanted to do in the first place.”

  • Bill McDermott, on the human-AI partnership

  1. “Service now has become the central nervous system of the 21st century enterprise.”

  • Bill McDermott, positioning the platform

  1. “If you always do what you always did, you’ll always get what you always got. I’m for change.”

  • Bill McDermott, on abandoning legacy CRM approaches

  1. “We are the only ones that do what we do, the way we do it.”

  • Bill McDermott, claiming platform uniqueness

  1. “Every minute matters. It’s crucial that in order to get life-saving medicines to patients, as quickly as possible, that we’ve streamlined and optimized every part of our business.”

  • Cindy Hoots, AstraZeneca CIO, on AI’s human impact

  1. “Trust is the ultimate human currency.”

  • Bill McDermott, explaining the NVIDIA partnership foundation

  1. “Service now is destined to be the best platform, the operating system of enterprise AI agents. And the reason for that is because every company in the world needs AI agents. And those AI agents are going to be curated, managed, made better by the IT department.”

  • Jensen Huang, validating ServiceNow’s strategy

  1. “We want to create that single point of interaction that really connects our joint customers who are out in the field with our customers back into the enterprise.”

  • Jennifer Morgan, CEO of UKG, on frontline worker empowerment

  1. “I want my service now.”

  • Audience chant led by Bill McDermott and Jensen Huang, cementing brand loyalty

🎯 HABITS

Product Development Habits

  • Platform-First Partnerships: ServiceNow co-engineers solutions with customers (Aptiv, AstraZeneca) and partners (NVIDIA, UKG) by embedding hundreds of engineers side-by-side, treating customer environments as co-development labs rather than sales targets.

  • Acquisition as Integration: The planned acquisitions of Moveworks and Logic.ai demonstrate a habit of buying specialized AI capabilities not for standalone products, but to deeply integrate reasoning and employee experience into the core platform architecture.

  • Demo-Driven Narrative: Every major announcement is anchored in a live demo (Brittany the lab manager/sales rep/patient) that follows a single persona through multiple workflows, making abstract architecture tangible and emotionally resonant.

Leadership Habits

  • CEO-as-Storyteller: Bill McDermott consistently uses personal anecdotes (his conversation with an unnamed Fortune 500 CEO, his first meeting with Jensen) and visceral metaphors (“duct tape and chewing gum,” “swivel-chair hell”) to make technical strategy accessible and urgent.

  • Trust-Based Ecosystem Building: McDermott explicitly credits Jensen’s early kindness when NVIDIA was “nobody” and frames their partnership as “one company model,” demonstrating a habit of investing in relationships years before they become strategically critical.

  • Metrics Anchoring: Every claim is backed by specific, large-scale numbers: 65 billion workflows, 90,000 hours saved, 200,000 employees, 60,000 research requests; establishing credibility through empirical scale rather than vaporware promises.

Personal Habits

  • Presence as Performance: Jensen Huang literally sprints on stage after landing “Maverick style,” while McDermott opens with crowd energy checks and personal shout-outs (Fred Luddy, board members), making executive presence a theatrical production that energizes the audience.

  • Continuous Learning Investment: The announcement of ServiceNow University with Idris Elba reflects McDermott’s habit of making education a core company mission, aiming to “empower 3 million AI learners by 2027” and treating upskilling as a product rather than an HR function.

📚 REFERENCES

  • Aptiv: Automotive OEM turned edge intelligence provider powering vehicles, aircraft, and 5G networks via Wind River OS; case study showing 50% automation of global service requests

  • UKG: Enterprise SaaS for HCM, payroll, and workforce management; represents the 80% hourly frontline workforce opportunity and demonstrates cross-platform agent communication

  • AstraZeneca: Pharmaceutical giant aiming for 20 new medicines by 2030; provides regulated industry validation with 90,000 hours saved in onboarding and 60,000 research requests automated

  • NVIDIA: AI hardware and software leader; co-developed Nemotron April reasoning model and provides the AI factory stack integrated with ServiceNow’s workflow fabric

  • Wind River: Aptiv’s real-time operating system for mission-critical devices; shows ServiceNow’s ability to connect physical edge computing with enterprise workflows

  • Logic.ai: AI-powered CPQ (configure-price-quote) acquisition target; demonstrates strategy of integrating specialized AI into native sales workflows

  • Moveworks: Employee experience AI acquisition target; shows commitment to removing “swivel-chair” complexity for knowledge workers

  • Nemotron April: New reasoning model co-announced by NVIDIA and ServiceNow; specifically designed for enterprise multi-step problem solving, PDF interpretation, and web research

  • MCP and A2A Protocols: Agent-to-agent communication standards; ServiceNow’s support signals commitment to open ecosystem over walled garden

  • ServiceNow University: Free, gamified AI learning platform featuring Idris Elba; represents strategic bet on creating talent pipeline and customer lock-in through education

✅ QUALITY & TRUSTWORTHINESS NOTES

  • Empirical Scale: Claims are substantiated with specific metrics from production deployments: 65 billion workflows executed, 90,000 hours saved at AstraZeneca, 50% automation rate at Aptiv’s 200,000-employee global workforce, and 60,000 research requests fulfilled in seconds rather than 20-30 minutes.

  • Cross-Institutional Validation: Multiple independent enterprise leaders (Kevin Clark of Aptiv, Jennifer Morgan of UKG, Cindy Hoots of AstraZeneca) provide corroborating testimony of measurable ROI, not just vendor promises, with each representing different regulated industries (automotive, workforce management, pharmaceuticals).

  • Technical Transparency: Jensen Huang explicitly details the three technical requirements for enterprise agents (multi-modal content understanding, reasoning, team collaboration) and announces Nemotron April as a jointly-developed, auditable model; avoiding vague “AI magic” claims.

  • Regulatory Proactivity: AstraZeneca’s Cindy Hoots notes they were “the first pharmaceutical company to publish the ethical use of AI statement,” and ServiceNow introduces specific AI governance solutions for regulated functions like HR, demonstrating awareness of compliance risks rather than ignoring them.

  • Partnership Longevity: The six-year NVIDIA collaboration, with hundreds of co-located engineers and shared product roadmaps, establishes this as a mature integration rather than a marketing-driven press release partnership.

  • Acknowledged Limitations: McDermott explicitly states that for 60 years companies have failed at digital transformation (only 1 in 4 projects succeed), acknowledging the historical difficulty of the problem they’re solving rather than pretending it’s trivial.

Crepi il lupo! 🐺