So for people who don’t know you, we connected in the UK ServiceNow Community comms. You’ve been amazing in connecting and partnering with myself and Eclipse AI to drive the AI ship forward for our customers.

You lead the team at ServiceNow responsible for AI Success at Scale,

which really is aiming to help organisations harness the power of AI, automation, and ITSM best practices to drive efficiency and 10x customer experiences.

You have a significant experience in the SaaS space.

Led strategic transformation initiatives across multiple industries, including defence, construction, and retail.

Your passion and expertise spans AI adoption, IT operations, enterprise automation, and service improvement frameworks.

In short, it is brilliant to have you here today, and hopefully for many more discussions to help our community unlock the value of AI.

Scott: It is an exciting era to be in and it is exciting to work with our partners in the pursuit of AI value and success at scale for our customers, so it is great to be working a partner like Eclipse AI who is so passionate about AI as we are at ServiceNow

Chris: Absolutely, and with the dropping of ServiceNow’s AI Agent Studio and Orchestrator, things are exciting in the ServiceNow eco-system. Which leads me to the first thing I would love for us to start of with.

After what seems like forever, AI Agents are now accessible to build and improve your company’s efficiency and value of output overall. But I think there’s quite a bit of confusion in the market about what exactly an AI agent is versus an AI assistant. Could you help us clarify that distinction?

Scott: That’s a great place to start, Chris. There’s definitely confusion about the terminology, and it’s important we get clear on it because it fundamentally changes how organisations approach implementation.

An AI assistant, which most people are now familiar with, is essentially a tool that responds to your queries and requests, but it waits for your input. Think of ChatGPT, Claude, or ServiceNow’s Now Assist - they’re incredibly powerful but fundamentally reactive. You ask, they answer.

AI agents, on the other hand, take a significant step forward. They don’t just respond - they act autonomously on your behalf. They can make decisions, execute tasks, and coordinate work across systems without constant human guidance. They’re proactive rather than reactive.

The difference might sound subtle, but it’s transformative in practice.

Chris: That’s a really helpful distinction. I often explain it to clients by comparing an assistant to a junior colleague who’s clever but needs constant direction, versus an agent who’s more like an experienced team member who you can give an objective to and trust them to handle the details.

Would you say that’s a fair analogy?

Scott: Agreed, Chris. An assistant is there to help you with the task you’re already doing, but an agent can take the entire task off your plate. If we use your analogy, an assistant is like having someone in the room who can answer questions while you’re working, whereas an agent is like delegating the entire piece of work.

And ServiceNow’s AI Agent Orchestrator takes this even further - it’s like having a coordinator who knows which agents to assign different parts of a complex task to, ensuring they all work together seamlessly.

Chris: I’ve noticed a lot of confusion in the market about this. I was recently working with a client who was using Microsoft Copilot, and they were struggling to understand the difference between that and the new Copilot Agent Studio capabilities. They’d built workflows that used Copilot to summarise meeting notes and post them to SharePoint, but they couldn’t see how agents were different from these automated workflows.

Scott: That’s such a common confusion point that we see from ServiceNow. Many organisations have already built workflows around AI assistants, so they naturally ask, “What’s the difference between my current automation and these new agents?”

The key difference is in decision-making autonomy. Traditional workflows, even those incorporating AI assistants, follow predetermined paths that a human designed. Every branch in the decision tree had to be anticipated and programmed in advance.

Chris: Yes and then AI agents, by contrast, can assess situations and determine appropriate actions dynamically. They can decide which tools to use based on context rather than following a rigid script. That’s why I often say we’re moving from a world where technical people build workflows to emulate human decisions, to a world where we can just have a conversation with an agent to build that decision-making capability.

That’s where the democratization aspect comes in, right? Because traditionally, automation required technical expertise - you needed someone who could build those workflows, understand APIs, and write code or complex configurations.

Scott: Exactly. This is what makes the current moment so significant. Up until now, if a business user or process owner wanted to automate something, they typically had to go through IT or technical teams. They’d submit a request, wait for it to be prioritized, and hope it eventually got built.

With AI agents, particularly with tools like ServiceNow’s AI Agent Studio, we’re democratizing that capability. A process owner can now build an agent through conversation, explaining what they want it to do in natural language. The technical complexities are abstracted away.

It’s comparable to how spreadsheets democratized financial modeling decades ago. Before spreadsheets, if you wanted financial projections or complex calculations, you needed to go through the IT department. Spreadsheets put that power directly in the hands of business users. AI agents are doing the same for process automation.

Chris: That’s a brilliant analogy. You’re absolutely right - this is as significant as the spreadsheet revolution, but for process automation.

One thing I’ve been thinking about recently is how this shift affects the way we conceptualize use cases. I was telling a client recently that they already have all the use cases they need - they don’t need to invent new ones.

Scott: This is something I’m passionate about. The message for me, and I’ve written it in our notes, is that “you already have the use cases.” Don’t sit down and go, “what use cases can we build for AI agents?” The use cases already exist, and you’ve been building them for years.

What you’ve been doing is bringing people early in their careers to do repeatable tasks. That’s the thing - it’s the same process. You take all the jobs that are repeatable, and you give them to people who are early in career, and they do those jobs. What we’re saying is you’re going to do the exact same process, but it’s for agents instead.

Chris: That’s such a powerful reframing. Instead of seeing AI agents as this exotic new technology that requires special use cases, we’re simply identifying the repetitive, rule-based work that already exists in the organization and thinking about how to hand that off to agents.

Scott: Absolutely. And there’s a crucial point here about how we communicate this to teams. We need to be really careful about the wording, because if you say it the wrong way, it sounds like job replacement rather than elevation.

What we’re really talking about is elevating people from repetitive tasks to more strategic work. We need to make it clear that we’re not looking to reduce headcount - we’re looking to increase capacity for innovation and higher-value activities.

Chris: I couldn’t agree more. This is something I’m passionate about as well. The cognitive shift is incredible, and people don’t realize that it’s coming. There’s people who underrate AI, and there’s people that overhype it. And the overhypers are dangerous to a degree. But the underrating of this is dangerous because you’ll get caught sleeping.

What it’ll do is remove laziness from the workplace. If you are doing a job that AI can do cheaper and 24/7, and you’re not growing or have any interest in growing, that’s who will lose the job. But if you’re in an organization that sees this as an opportunity to get more value out of you, to grow - that’s where the real transformation happens.

Scott: That’s a profound point. Every company wants to grow, and a lot of companies need people to grow, as growth is highly linked to how many people you have in your business. If you have people that are maximizing their potential all the time, you are getting prime product for what you’re paying for as an enterprise.

And as an individual, if you are being pushed to your potential, you are growing to become the best person you can be. It’s a win-win when approached correctly.

Chris: I’ve always had a mantra, even before AI came along, that the best way to grow is to make yourself dispensable. I read it in a book by Branson when I was at university. He said if you make yourself dispensable, you have to grow into a place where for a period of time you’re indispensable. And as long as you’re having this cycle of dispensable, indispensable, you are elevating yourself constantly.

This is an opportunity for individuals to do that, for enterprises to do that. You take tedious work away from people, and you activate their potential. You make them the managers of those AI agents. They learn how to orchestrate AI agents. They manage the policy.

Scott: That’s such a powerful mindset. And it’s refreshing to hear that perspective, especially in contrast to some other voices in the industry. Marc Benioff from Salesforce has made statements about stopping hiring software engineers - which isn’t the message that helps growing businesses.

The message from ServiceNow has consistently been about how we 10x our customers’ work experience for all employees. It’s about elevation, not replacement. And that distinction in approach makes all the difference in how organizations adopt and benefit from these technologies.

Chris: You’re absolutely right. And I think that’s what makes this moment so important - it’s not just a technical shift, it’s a philosophical one. We’re moving from automation as a cost-cutting exercise to automation as an enabler of human potential.

When we talk about democratizing automation through AI agents, we’re really talking about democratizing opportunity - giving everyone in the organization the tools to focus on what humans do best: creativity, empathy, strategy, and innovation.

Scott: That’s exactly it. And this democratization has another critical dimension - it addresses one of the biggest bottlenecks in digital transformation. When only technical teams can build automation, you create an inevitable backlog. Business teams identify opportunities faster than technical teams can implement them.

By putting these capabilities directly in the hands of the people who understand the processes best, we’re removing that bottleneck. It’s like giving everyone in the organization a superpower to reshape their own work.

Chris: I think we’ve given our listeners a really clear distinction between AI assistants and agents, and painted a powerful picture of why this matters right now. Before we move on to more specific implementation considerations, any final thoughts on this fundamental shift we’re discussing?

Scott: Just that I encourage everyone listening to think about this not as a technical project but as a business transformation opportunity. The organizations that will benefit most are those that approach AI agents as a way to reimagine work - not just automate what already exists.

Start by looking at the processes that cause the most friction in your organization. Where do things get stuck? Where do your talented people spend time on routine tasks instead of creative ones? Those are your starting points for this journey.

Chris: Excellent advice. And with that, let’s shift gears to talk about some practical implementation strategies…

Technical

Chris: Now that we’ve established the fundamental distinction between AI assistants and agents, I’d like to dig into what makes effective agent architecture. There’s a critical insight I’ve been sharing with clients that completely transforms how they think about implementing AI agents.

Scott: I’m intrigued - what’s this insight?

Chris: It’s what I call the “radial approach” to agent design. Most organizations are still trying to build AI agents using the same linear workflow thinking they’ve always used. They’re essentially trying to hammer the square peg of agent technology into the round hole of traditional process design.

Scott: That’s an interesting way of putting it. Can you explain what you mean by “radial approach” versus the traditional linear workflows?

Chris: Absolutely. In traditional workflow automation, we build linear processes with predefined steps - if this happens, then do that, and so on. It’s like a flowchart with decision branches, but fundamentally it’s a fixed path with predictable endpoints.

The radial approach flips this completely. Instead of designing the system to channel the agent, we let the agent channel the system design. The agent sits at the hub of a wheel, with spokes extending outward to different resources - tools, knowledge bases, APIs, and even other agents. Rather than following a predetermined path, the agent dynamically decides which spokes to traverse based on the task at hand.

Scott: That’s a powerful mental model. So rather than trying to anticipate every possible path in advance, you’re creating an environment where the agent can make contextual decisions about how to accomplish its goals?

Chris: Exactly. It’s the difference between hiring someone and giving them a rigid script to follow versus hiring someone, explaining their objectives, and showing them what resources are available to accomplish those objectives.

Let me share a practical example. I was working with a client who was trying to implement an agent to handle employee onboarding. Initially, they were trying to map out every possible scenario and decision point in the process - essentially recreating their existing workflow but with an AI touch. It was becoming unwieldy and inflexible.

When we shifted to a radial approach, we instead focused on defining what resources the agent needed access to - the HR system, the IT ticketing system, the knowledge base of policies, the email templates, and so on. Then we simply instructed the agent: “Your job is to ensure new employees have everything they need on day one.”

The difference in results was remarkable. The radial approach created an agent that could adapt to unique situations, learn from experience, and even suggest improvements to the process itself.

Scott: That’s fascinating. It sounds like this approach better leverages what AI is good at - being adaptive and learning from context - rather than trying to force it into rigid structures. And I can see how that connects nicelu with ServiceNow’s AI Agent Orchestrator.

Chris: The Agent Orchestrator is built for exactly this kind of approach. It serves as that central hub that understands which specialized agents have what capabilities, what resources are available, and how to coordinate action across the system. It’s designed for radial thinking rather than linear processing.

But here’s where we hit an interesting challenge that not many people are talking about yet - what I call the “task length barrier.”

Scott: The task length barrier? What’s that?

Chris: It’s a fascinating phenomenon revealed when I read a paper by the METR research. As tasks grow in complexity and duration - specifically, tasks that would take a human expert more than 80 hours - AI performance degrades exponentially. Even the most advanced models like Claude 3.7 Sonnet show significant performance drop-offs on complex, sustained tasks.

This is why many early agent implementations have been disappointing. Organizations try to build a single, general-purpose agent to handle complex end-to-end processes, and it inevitably falls short.

Scott: Well ServIcenow’s Orchtestrator plays perfectly into the solution to solving this, because surely it is about divisible tasking and domain specific agents that are tightly coordinated?

Chris: Exactly. Vertical agents. I wil preach them till I die or they become irrelevant. And Servicenow has architectured AI Agents in their platform perfectly. Vertical agents are specialized for specific domains or functions rather than being general-purpose. They have deeper knowledge and capabilities within a narrower context. Think of a finance agent that’s expert in reconciliation processes, or an HR agent specifically designed for leave management.

The METR data is striking - while horizontal, general-purpose agents typically achieve around 47% success rates on complex tasks, properly designed vertical agents can achieve 90%+ in their domains of expertise.

But here’s the really exciting part - the task length barrier gets defeated by what I call a “community of AI agents” or a “cognitive village.”

Scott: I am loving this cognitive village concept and it does sound very much in line with what we built at ServiceNow in the AI Studio. If I get you rights, it’s the idea that instead of having one super-intelligent agent trying to do everything in your ServiceNow processes, you create a community of specialized agents in the Studio working together, each with their own expertise and responsibility and coordinated by the Orchestrator. Just like in a human village, you have specialists - the baker, the blacksmith, the teacher - in an AI cognitive village, you have specialized agents that collectively accomplish what would be impossible for any single agent.

Chris: You read my mind. ServiceNow’s architecture is perfectly suited for this approach. The AI Agent Orchestrator serves as the coordinator - like the village mayor - while individual agents built in the AI Agent Studio handle specialized tasks. Let me give you a real-world example.

I recently worked with a financial services client who was struggling with their quarterly compliance reporting process. It was complex, involving data gathering from multiple systems, regulatory analysis, document preparation, and multi-level approvals. Their initial attempt to build a single agent to handle the entire process was failing - running straight into that task length barrier.

We redesigned it as a cognitive village with specialized agents:

• A data collection agent that knew how to gather information from various systems
• A compliance analysis agent specialized in regulatory requirements
• A document preparation agent focused on creating standardized reports
• A coordination agent to manage approvals and timelines

Scott: So in the ServiceNow world, Orchestrator managed the entire process, delegating tasks to the right specialist at the right time. Chris: Ye, and the results were remarkable - what had been a three-week process involving dozens of people was reduced to three days with minimal human intervention.

Scott: That’s a compelling example. It sounds like there’s a parallel to how human organizations work - we don’t expect any single person to be an expert in everything.

Chris: Exactly! It’s strange that we somehow expected AI to work differently. No human could be simultaneously expert in HR processes, finance, IT security, and customer service - yet we’ve been trying to build general-purpose AI that can do all of these things. The cognitive village approach simply mirrors how human organizations have always functioned.

Scott: I can see how this would be powerful in ServiceNow implementations specifically. Our platform naturally spans multiple departments and processes - IT, HR, facilities, finance, customer service. Building specialized agents for each domain that can work together through the Orchestrator seems like the natural approach.

Chris: That’s precisely why ServiceNow is so well-positioned for this. You already have the platform that connects these different domains. The AI Agent Orchestrator leverages that existing foundation to create coordinated, intelligent automation across departmental boundaries.

Let me highlight another advantage of the cognitive village approach: it makes governance and security much more manageable. Rather than giving a single agent broad access to everything, each specialized agent only needs access to the specific systems and data relevant to its function. It’s the principle of least privilege applied to AI.

Scott: That’s a critical point for enterprise adoption. Security and governance are top concerns for our customers. The ability to precisely control what each agent can access and do is essential for responsible implementation.

Chris: Absolutely. And there’s one more benefit that’s often overlooked - specialized agents are much easier for non-technical users to create and maintain. If you’re in HR, you understand HR processes deeply. You can create an effective HR agent without needing to understand finance or IT operations.

This brings us back to the democratization theme we discussed earlier. The cognitive village approach doesn’t just solve technical challenges - it enables distributed ownership of automation across the organization.

Scott: I love how this all connects. The radial design approach creates adaptable agents, vertical specialization overcomes the task length barrier, and the cognitive village model enables complex processes to be automated through collaboration between specialized agents.

It sounds complex when described technically, but it’s actually a very intuitive approach that mirrors how humans naturally organize work.

Chris: That’s exactly right. And what excites me most is that this approach is accessible to non-technical users. You don’t need a computer science degree to understand these principles - they’re fundamentally about how work gets done, not about the underlying technology.

Scott: I think this gives the listeners a clear framework for thinking about agent implementation. Rather than trying to build a single “super agent,” focus on creating specialized agents for specific functions, and then use the Orchestrator to coordinate them.

And from ServiceNow’s perspective, this aligns perfectly with our platform strategy. We’ve always been about connecting workflow across departments and systems, and AI agents extend that capability in powerful new ways.

Chris: Exactly. And what’s truly exciting is that we’re just at the beginning of what’s possible with this approach. As organizations build more specialized agents and learn how to orchestrate them effectively, we’ll see entirely new possibilities emerge - just as specialized roles in human society enabled achievements that would have been impossible for generalists.

Scott: This has been an incredibly insightful discussion of the technical foundations. I wondered if we could talk about implementation strategies as things are not all technical implementations. but before we move on to implementation strategies, any final thoughts on this architectural approach?

Chris: Just that organizations shouldn’t be intimidated by these concepts. The beauty of ServiceNow’s implementation is that it makes these sophisticated architectural patterns accessible. You don’t need to understand the technical details of how the AI works - you just need a clear vision of what you want it to accomplish and the right approach to structuring your agents.

The technology is ready. The question is whether organizations are ready to think differently about how they approach automation.

Scott: That’s a perfect segue into our next topic - let’s talk about practical implementation strategies and real-world use cases…

Implementation - Real-World Implementation Segment

Chris: Now that we’ve explored the technical foundation in a way that hopefully makes sense to business leaders, let’s get practical. How do organizations actually get started with implementing AI agents effectively?

Scott: That’s where the rubber meets the road, isn’t it? And I think many organizations stumble right at the beginning because they’re not sure where to start.

The first challenge is identifying the right use cases. And as we touched on earlier, I believe organizations already have the use cases—they just need to recognize them.

This is something I feel strongly about. Don’t sit down with a blank sheet and ask, “What could AI agents do for us?” Instead, look at what your organization is already doing repeatedly that could be handled by agents.

Chris: I find that the most promising opportunities typically fall into three categories:

First, high-volume routine tasks that follow consistent patterns but still require judgment. Think of categorizing incidents in IT, processing leave requests in HR, or validating supplier invoices in finance.

Second, cross-departmental processes where work frequently gets stuck during handoffs. Employee onboarding is a classic example—it touches HR, IT, facilities, and often finance.

Third, information gathering and synthesis tasks where people spend more time collecting data than analyzing it. This includes status reporting, compliance documentation, and knowledge management.

Scott: Those are great categories. I’d add that organizations should look particularly at places where early-career employees are doing repetitive work. Those are natural candidates for agent automation.

Chris: Great point. One approach I’ve used successfully with clients is what I call the “frustration audit.” Simply ask team members: “What tasks do you find most tedious or frustrating in your daily work?” Those pain points often reveal prime opportunities for agent implementation.

Scott: I love that approach. When we’ve seen the most successful implementations at ServiceNow customers, they’ve typically started with those pain points rather than trying to implement the most technically impressive solution.

Chris: Let me share a specific example from a UK utility company I worked with recently. They had a process for handling infrastructure change requests that involved multiple teams—the requestor, technical reviewers, security, and implementation teams.

The process was full of friction points—emails getting lost, approvals delayed, documentation inconsistent. When we mapped it out, we identified three distinct agent opportunities:

1. A request processing agent to gather all necessary information upfront
2. A documentation agent to ensure complete and consistent documentation
3. A coordination agent to track approvals and keep stakeholders informed

What made this successful was that we didn’t try to automate the entire process at once. We started with the documentation agent, which had a clearly defined scope and could show immediate value.

Scott: That incremental approach is crucial. We see organizations trying to boil the ocean with their first agent implementation, and it rarely works. Start with a contained use case, demonstrate value, and build from there.

Chris: Exactly. Now, let’s talk about implementation strategy. I believe there are three critical components to successful agent implementation.

First, you need the right technical foundation. This includes ensuring your data is accessible and structured appropriately, establishing integration points with relevant systems, and defining clear boundaries for agent operations.

Second, you need appropriate governance frameworks. This covers who can create agents, what approval processes are needed, how agents are tested before deployment, and how their performance is monitored.

Third, and perhaps most importantly, you need a thoughtful change management approach. This includes training for both agent creators and users, clear communication about the role of agents, and measurement of business outcomes rather than just technical metrics.

Scott: Those are excellent points. On the governance side, ServiceNow has built robust capabilities directly into the platform. AI Agent Studio includes built-in guardrails, permission controls, and comprehensive logging to address security and compliance concerns.

One approach we recommend is to start with “read” actions as autonomous while keeping “write” actions supervised until confidence in the AI agents’ performance is established.

Chris: That’s a sensible approach. I’ve seen organizations implement what I call “observation mode” for new agents, where they recommend actions but require human approval before executing them. As confidence builds, they gradually increase autonomy.

Scott: Exactly. Another critical factor is data readiness. While agents can work with imperfect data—they’re more adaptable than traditional automations—they still need access to the right information in usable formats.

This is where I see many organizations get stuck in what I call the “combi boiler problem.” You’ve got all these people excited about hot water, but if the boiler isn’t powerful enough, everyone ends up with cold showers.

In practical terms, this means ensuring that your agents have access to the systems and data they need. This doesn’t require perfect data—a myth that paralyzes many organizations—but it does require accessible, reasonably structured information and appropriate API access.

And speaking of access to different systems, let’s talk about cross-platform integration. Many of our customers are running hybrid environments with ServiceNow alongside Microsoft tools, Salesforce, and other platforms.

Chris: This is where some truly powerful use cases emerge. Let me share an example of ServiceNow and Microsoft integration that I worked on with a financial services client.

They had a demand review process managed in ServiceNow, but all their documentation was built in SharePoint, and their meetings were held in Teams. Previously, someone would attend the Teams meeting, take notes, manually create documentation in SharePoint templates, and then update the ServiceNow records.

We implemented a solution where an AI agent would:

1. Join the Teams meeting and transcribe it
2. Extract key decisions and actions
3. Populate the SharePoint templates
4. Update the ServiceNow records with links to the documentation
5. Notify stakeholders when approvals were needed

This wasn’t just incremental efficiency—it was transformative. What had been a multi-day process with frequent errors became an automated workflow that completed within minutes of the meeting ending.

Scott: That’s a perfect example of how agents can bridge silos between platforms. And I think it illustrates why ServiceNow is so well-positioned in this space—we’re designed to be the integration platform that connects these different systems.

Chris: Absolutely. The ServiceNow AI Agent Orchestrator becomes the central brain coordinating activities across platforms. And the beauty is that you don’t need to rip and replace existing systems—you’re enhancing and connecting them.

Scott: I think this also highlights the importance of ServiceNow’s platform approach. We’ve built AI capabilities natively into the Now Platform rather than as bolt-on additions, which makes these cross-platform integrations much more seamless.

Chris: Let me share another implementation insight that often gets overlooked: the importance of continuous improvement cycles. The most successful agent implementations I’ve seen treat the initial deployment as just the beginning.

Agents improve through use and feedback. Organizations should establish clear mechanisms to monitor performance, gather user feedback, and continuously refine their agents. This is another advantage of the ServiceNow platform—the comprehensive logging and monitoring capabilities make it easy to identify improvement opportunities.

Scott: That’s an excellent point. Unlike traditional automation, which is relatively static, agents can learn and adapt. Organizations that build feedback loops into their implementation process see significantly better outcomes over time.

Chris: Now, let’s talk about getting started. For organizations just beginning their agent journey, what’s a practical first step?

Scott: I recommend a three-phase approach:

First, the Discovery Phase. Identify 3-5 potential use cases using the criteria we’ve discussed. Assemble a cross-functional team spanning business and IT, and establish clear success metrics.

Second, the Pilot Phase. Select your most promising use case and implement it in a contained environment. Focus on learning as much as on outcomes.

Third, the Expansion Phase. Apply those learnings to implement additional use cases and begin building your internal community of agent creators.

Chris: That’s a solid approach. I’d add one more recommendation: create a Center of Excellence or Community of Practice around AI agents. This creates a forum for sharing knowledge, reusing components, and establishing best practices.

Scott: Absolutely. And this brings us back to the democratization theme. The goal isn’t just to implement a few agents—it’s to build organizational capability for creating and managing agents across departments.

Chris: Exactly. And that capability becomes a significant competitive advantage. Organizations that can rapidly deploy agents to address emerging needs will be fundamentally more agile than those relying on traditional development approaches.

Scott: One thing I would like to double click on there: You don’t need flawless data or complete process documentation to get started. Begin with a well-defined, contained use case, learn from that experience, and expand from there.

The organizations that are waiting for everything to be perfect before they start will be left behind by those who are learning through doing.

And it aligns with what we’re seeing across our customer base—the most successful implementations start small, demonstrate value quickly, and grow organically as the organization builds confidence and capability.

Chris: Absolutely. Now let’s discuss the human and organizational impact of this transformation…

Human and Organizational Impact Segment

Chris: We’ve covered the technical foundations and implementation strategies, but I’d argue the most critical factor for success is how organizations handle the human and organizational impact of AI agents.

Scott: I completely agree. The technology is powerful, but without the right approach to people and culture, implementations often fail to deliver their full potential.

Chris: Let’s talk about what I consider the fundamental mindset shift required—moving from a replacement mentality to an elevation mentality.

Scott: This is so important. When I’m speaking with executives, I often see two distinct approaches emerging. Some view AI primarily as a cost-cutting mechanism—“How can we reduce headcount and lower expenses?” Others see it as a growth enabler—“How can we free our people from routine tasks so they can deliver more value?”

Chris: And the data is compelling on which approach yields better results. Organizations focusing primarily on cost reduction through AI achieve significantly lower ROI than those prioritizing growth and innovation. McKinsey’s research puts the difference at about 75%.

Scott: That’s striking. Why do you think there’s such a difference in outcomes?

Chris: I believe it comes down to engagement and opportunity. When employees perceive AI as a threat to their jobs, they resist adoption. When they see it as a tool that eliminates their most tedious tasks and creates opportunities for more meaningful work, they become champions for the technology.

I’ve always followed a principle I learned early in my career: the best way to grow is to make yourself dispensable. When you make yourself dispensable in your current role by automating or delegating routine tasks, you create the opportunity to grow into something more valuable.

Scott: That’s a profound perspective. And we need to be very thoughtful about how we communicate around AI agents. The messaging matters tremendously.

Chris: Absolutely. Organizations need to be explicit about their intentions—that the goal is to elevate employees, not replace them. This can’t just be lip service; it needs to be backed by tangible policies and actions.

For example, with one client we created what I called an “Elevation Pledge”—a formal commitment that time saved through agent automation would be reinvested in either professional development or innovation projects. This transformed the narrative completely.

Scott: That’s a brilliant approach. And it connects to what we have discussed today about—democratizing creation across the organization. When people feel secure, they’re more likely to embrace the opportunity to become creators rather than just users of technology.

We have setup the framework and methodology in ServiceNow where the most successful implementations will be where they treat AI agents as an empowerment tool for employees at all levels. They establish programs that encourage people to identify automation opportunities in their own work and provide them with the training and support to build agents themselves.

Chris: Yes! This distributed approach accomplishes two things: it multiplies the impact by tapping into the knowledge of people closest to the work, and it changes the relationship between employees and automation from potential adversaries to collaborators.

Scott: At ServiceNow, we’re seeing increased interest in exactly this kind of decentralized model. The AI Agent Studio is designed specifically to enable this broader participation—allowing business users without technical expertise to create agents through conversation.

Chris: This democratization requires a significant cultural shift, though. Many organizations are structured around the premise that technology implementation belongs to IT, while business teams are just consumers of those services.

Scott: Absolutely. And this is where leadership becomes crucial. In successful implementations, we see executive teams actively modeling the behaviour they want to see—learning to create and work with agents themselves rather than delegating it entirely to technical teams.

Chris: It reminds me of the early days of the internet. Organizations that treated web technology as purely an IT concern fell behind those that encouraged broader digital literacy across departments.

The same pattern is emerging with AI agents. The organizations creating the most value are those that view agent creation as a broadly distributed capability rather than a specialized technical function.

Scott: That said, there is still an important role for governance and guardrails. Democratizing doesn’t mean eliminating oversight.

Chris: Absolutely. The most effective models I’ve seen follow what I call “governed democratization”—clear boundaries, standards, and review processes, but with the default position being enablement rather than restriction.

For example, organizations might establish categories of agents with different approval requirements based on their potential impact and risk profile. A departmental knowledge agent might have minimal governance requirements, while an agent that can make financial transactions would have more rigorous controls.

Scott: That makes perfect sense. And this brings us to the cultural transformation required. I see three fundamental cultural shifts that successful organizations embrace:

First, they move from process rigidity to outcome focus. Rather than insisting processes be followed exactly as designed, they focus on whether the right outcomes are being achieved.

Second, they shift from knowledge hoarding to knowledge sharing. In many organizations, personal expertise is a source of job security. In an agent-enabled organization, the value comes from making that expertise accessible to others through agents.

Third, they evolve from control to enablement. They recognize that innovation happens faster when people can experiment within appropriate boundaries rather than requiring permission for every action.

Those are profound shifts. And they don’t happen overnight.

Chris: No, they certainly don’t. It requires intentional effort and leadership. I often recommend organizations start by identifying “agent champions” in different departments—people who are excited about the potential and can help build momentum for adoption.

Scott: That connects back to the implementation approach we discussed earlier—starting small, showing value, and expanding organically. The cultural transformation can follow a similar pattern.

Chris: Exactly. And organizations need to celebrate and share successes along the way. When someone builds an agent that transforms a process, make them a hero. Share their story. Create incentives for others to follow their example.

This has been such an insightful conversation, Scott. I think we’ve covered the essential aspects of AI agent implementation—from the technical foundations to implementation strategies to human and organizational impacts.

Any final thoughts for our listeners who are considering embarking on this journey with ServiceNow AI agents?

Scott: I would encourage everyone to recognize that this isn’t just another technology implementation—it’s the beginning of a fundamental transformation in how work gets done.

The organizations that will thrive are those that approach it with both strategic clarity and human sensitivity—understanding the technical requirements while also nurturing the cultural changes needed for success.

Chris: I love that. Start small, but think big. Show tangible value quickly, but build toward a vision of work where routine tasks are handled by agents while human creativity and judgment are focused on innovation and growth.

Scott: That’s a perfect summary. And remember that ServiceNow and partners like Eclipse AI are here to help you on that journey. We’re not just providing technology—we’re sharing best practices, implementation frameworks, and change management approaches based on what we’ve learned from successful deployments.

Chris: Absolutely. This is a journey we’re taking together, and the possibilities ahead are truly exciting. Thank you, Scott, for sharing your insights today. It’s been a pleasure.

Thank you, Chris. I’ve enjoyed our conversation and look forward to continuing to explore these topics in future episodes.

Chris: To our listeners, thank you for joining us. If you have questions or topics you’d like us to cover in future episodes, please reach out to us through the Eclipse AI website or ServiceNow community forums.

Chris: Until next time, I’m Chris Jones from Eclipse AI.

Scott: And I’m Scott Gamble from ServiceNow. Thanks for listening.