ServiceNow Knowledge 2025: Enterprise AI Workflow Deployments from Eight Named Companies

Executive Summary

• ServiceNow Knowledge 2025 (May 2025, Las Vegas, 25,000+ attendees) produced quantified AI deployment metrics from eight named companies — fewer than Salesforce Dreamforce but stronger on workflow transformation evidence, particularly in IT service management and HR operations.
• EY’s deployment is the most detailed: 12,000 AI actions per day across 400,000 employees, 103,000 resolution notes generated in the first month, 70% accepted without edits, saving an estimated 66,000 hours and $2.3 million annually. This is one of the largest named Now Assist deployments with public metrics.
• Orica’s deflection rate jump from 18% to 94% in eight weeks is the single most dramatic before/after metric in the Knowledge 2025 corpus — though it measures virtual agent containment, not end-to-end resolution quality.
• The pattern across all eight deployments reinforces the same finding from AWS re:Invent, Microsoft Ignite/Build, Google Cloud Next, and Dreamforce: IT service desk and HR service delivery produce the fastest, most measurable AI ROI. No company disclosed metrics for AI in revenue-generating or customer-facing operations at the same specificity level.
• ServiceNow’s own Enterprise AI Maturity Index (n=4,500 executives, Oxford Economics) found maturity scores dropped 9 points year-over-year — a useful counterweight to vendor optimism and a data point that maps directly to the 5%-vs-95% thesis.

The Production Metrics

IT Service Management and Operations

HR Service Delivery

Operations and Supply Chain

Vendor Self-Deployment (“Now on Now”)

Platform Financial Context

ServiceNow’s Q4 FY2025 earnings (January 2026) frame the commercial scale behind these deployments:

• Now Assist surpassed $600 million in ACV, with Q4 net-new ACV more than doubling year-over-year
• 35 Now Assist deals above $1 million in Q4 alone
• AI Control Tower deal volume nearly tripled quarter-over-quarter
• Knowledge 2025 generated $1.2 billion in new sales pipeline
• Total quarterly revenue: $3.6 billion (up 20.5% YoY); 98% renewal rate

These are SEC-filed figures and represent the strongest commercial validation among enterprise workflow platforms. The $600M ACV for Now Assist alone exceeds many standalone AI companies’ total revenue.

The Maturity Paradox

ServiceNow’s Enterprise AI Maturity Index 2025 (n=4,500 executives, Oxford Economics partnership) provides a useful counterweight to the deployment success stories above:

• Average enterprise AI maturity score dropped 9 points year-over-year (44 → 35 on a 100-point scale)
• 55% of enterprises have rolled out 100+ AI use cases — but only 30% deploy them across multiple functions
• 67% report AI increased gross margin (average 11% increase) — yet maturity scores still declined

ServiceNow interprets this as “awareness of what good looks like has risen faster than execution.” The more honest read: companies are deploying AI use cases at scale but siloing them within functions, creating the same fragmentation pattern the platform promises to solve. This maps directly onto BCG’s finding that only 5% capture substantial financial gains — the other 95% have AI running but not connecting.

Methodology Notes and Source Credibility

Overall credibility: MEDIUM. Knowledge 2025 is a vendor conference. Every customer on that stage was selected by ServiceNow, and every metric was approved for public sharing. No control groups. No independent verification. The survivorship bias is identical to AWS re:Invent, Google Cloud Next, and Dreamforce — these are the wins, not the median outcome.

What makes the Knowledge data more useful than some vendor conferences: EY disclosed specific operational numbers (12,000 actions/day, 70% acceptance rate, $2.3M value) that are specific enough to falsify. Orica’s 18% → 94% deflection rate has a clear baseline. AstraZeneca’s 90,000 hours saved comes from a named CIO on-stage. These are not marketing brochure numbers.

These case studies are vendor-published and represent selected wins with no control group and no independent verification. Cross-reference against: METR RCT (experienced developers 19% slower), CMU study (40.7% code complexity increase), Atlan 200-deployment analysis (median +159.8% ROI requires workflow redesign first).

The Enterprise AI Maturity Index (n=4,500, Oxford Economics partnership) carries MEDIUM-HIGH credibility — it is ServiceNow-commissioned but executed by an independent research firm, and the finding (maturity declining) runs counter to vendor interest.

Key Data Points

What This Means for Your Organization

The ServiceNow Knowledge data reinforces a pattern now visible across all five major vendor conferences this research has covered: IT service desk and HR service delivery are the proven entry points for enterprise AI workflow automation. Every company that disclosed specific metrics at Knowledge 2025 started there — not in revenue-generating operations, not in strategic planning, not in product development.

The maturity paradox is the more important finding. Half of the enterprises surveyed have deployed 100+ AI use cases, yet maturity scores declined. The implication: deploying AI is not the hard part. Connecting deployments across functions, measuring actual business outcomes, and redesigning workflows around AI outputs — that is where most organizations stall. Orica’s 18% → 94% deflection improvement did not come from installing Now Assist; it came from rewriting knowledge base content and restructuring escalation paths to work with the AI.

For organizations evaluating ServiceNow’s AI capabilities or any workflow AI platform, the honest question is not “does the technology work?” — the Knowledge 2025 evidence says it does, within the IT and HR domains where it has been tested. The question is whether the organization is prepared to redesign the workflows surrounding the technology. If that question raises specifics worth working through, the conversation is open — brandon@brandonsneider.com.

Sources

• Pariseau, B. “ServiceNow shops share AI copilot results, prep for agents.” TechTarget, May 2025. https://www.techtarget.com/searchitoperations/news/366623782/ServiceNow-shops-share-AI-copilot-results-prep-for-agents (Credibility: HIGH — independent trade press, named sources, specific metrics)
• Lopez, M. “AI Agents in Action: ServiceNow’s Knowledge 2025 Vision.” Lopez Research, May 2025. https://www.lopezresearch.com/ai-agents-in-action-servicenows-knowledge-2025-vision-for-enterprise-workflow-transformation/ (Credibility: MEDIUM-HIGH — independent analyst, on-site reporting)
• “5 game-changing lessons from AI award winners.” ServiceNow Blog, May 2025. https://www.servicenow.com/blogs/2025/5-lessons-ai-award-winners (Credibility: MEDIUM — vendor blog, but AI Pacesetter Award winners disclosed specific metrics)
• “ServiceNow Q4 FY 2025 Earnings Highlight AI Platform Momentum.” Futurum Group, Jan 2026. https://futurumgroup.com/insights/servicenow-q4-fy-2025-earnings-highlight-ai-platform-momentum/ (Credibility: HIGH — SEC-filed financial data)
• “The Top Takeaways From ServiceNow Knowledge 2025.” CRI Advantage, May 2025. https://criadvantage.com/knowledge-2025-takeaways/ (Credibility: MEDIUM — partner recap)
• ServiceNow Enterprise AI Maturity Index 2025. ServiceNow/Oxford Economics, May 2025. https://www.servicenow.com/content/dam/servicenow-assets/public/en-us/doc-type/resource-center/white-paper/wp-enterprise-ai-maturity-index-2025.pdf (Credibility: MEDIUM-HIGH — vendor-commissioned but Oxford Economics executed; n=4,500)

Brandon Sneider | brandon@brandonsneider.com April 2026