ServiceNow's AI Vision: Interview with Chris Bedi

AI Impact Across the Enterprise

According to Chris Bedi, the impact of machine learning, generative AI, and agentic AI can be felt throughout the company. The goal is for every human decision to be supported by a recommendation based on machine learning, to the point where making a decision without that input would feel strange. While decisions are sometimes deferred to the model when accuracy is high, it is still not common.

A significant challenge lies in adoption, not the technology itself. User experience is essential to making AI an integral part of the workflow. Due to a focus on adoption, there are now hundreds of models in production, creating millions of predictions daily.

Examples of AI Implementation

In field sales, a team previously dedicated to answering compensation questions could provide responses in up to four days. Now, generative AI can provide answers in eight seconds, allowing those team members to concentrate on more strategic tasks such as redesigning compensation models.

Agentic AI is also suitable for customer support. After AI was trained using standard operating procedures and a knowledge base, there was an 18% improvement in resolution time for high-volume, low-complexity cases and a 54% improvement for lower-volume, more complex cases. The AI added more value in complex cases by running diagnostics, retrieving logs, and executing scripts.

Measuring AI's Impact

The impact of AI is measured in these ways: AI’s impact contributes to growth, margin, and revenue per employee at the enterprise level. Additionally, the focus is on how AI saves people time through summarization, meeting notes, and automation, increasing capacity without replacing people.

Agentic AI is already handling 20% of support cases, and that percentage is increasing. The long-term objective is the “AI employee,” defined as software capable of performing 80% or more of a specific job. ServiceNow has approximately 2,100 unique roles. If an AI employee is created for even one of those roles, it may eliminate the need to hire for that role again, instead hiring for new roles that don't yet exist. The path involves a combination of agents capable of handling 80% of a job's tasks. Digital employees will first appear in support, HR, and procurement.

A concept that resonates with customers is the idea of a “zero-headcount department,” such as a support function without any full-time employees. While achieving zero may be unrealistic, a 70% reduction in headcount is appealing.

AI employees will collaborate with humans. An AI recruiter could support a human recruiter, but with greater scalability. Another important change is proactivity. Currently, most Gen AI is reactive, waiting for prompts. However, AI employees, especially those powered by agentic AI, will operate independently.

Productivity has already increased by 20% across customer, HR, and IT support because of AI agents. Currently, most companies are leveraging AI to increase capacity by automating simple tasks and providing summaries. Digital employees and full-scale AI operations are gaining traction.

AI is a priority. There has been a shift from pilots and proof-of-concepts to scaled, production-level use cases. ServiceNow aims to scale intentionally. Agentic AI has been implemented in several areas, and the expectation is to have 10–15 digital employees by the end of the year. There has already been over $350 million in enterprise value derived from AI.

The focus is on AI, data, and workflows as the fundamental components. Metrics are defined at the company and department levels, and projects that advance those metrics are prioritized. Teams are given space to experiment, such as through AI hackathons or sandbox environments. Innovation is encouraged, but in a structured manner, balancing focus and exploration.

Enterprises should remain intellectually curious and learn from startups. Clear on-ramps for startups should be created. Interoperability is important because enterprises need it. One of the biggest challenges is adoption. It's important to make tools seamless and an integral part of the daily workflow. The focus should be on executing well because poor UX can hinder adoption.