What I like best is that the AI Agents run natively on the Now Platform, so they operate on the same governed data model and workflows we already trust. The agents have direct context from the CMDB, CSDM, and live records, which means they act on real platform state rather than a disconnected layer bolted on top. The AI Agent Orchestrator handling coordination across specialised agents is a big plus, and because everything stays inside the platform's existing RBAC, ACLs, and audit trail, governance and compliance are intact by default rather than being an afterthought. For regulated clients I work with, that combination of context, orchestration, and built-in guardrails is what makes the agents usable in production and not just in a demo. Review collected by and hosted on G2.com.

More user friendly and security concernsThe main friction is cost predictability. Consumption runs on Assist units, and forecasting usage across a large agent deployment is hard before you see real production volumes, which makes budgeting conversations with finance teams difficult. Licensing is also a moving target. The shift to the Foundation, Advanced, and Prime tiers changed how entitlements map to AI capabilities, and keeping clients aligned on what they actually own versus what needs an upgrade takes real effort. Beyond cost, the agents are most effective when the underlying data foundation is mature. On clients with a weak CMDB or incomplete CSDM, the agents underperform and you end up doing foundational data cleanup first. Skill coverage out of the box is still uneven across non-ITSM workflows, so custom agent build effort is higher than the marketing suggests. Better native observability into why an agent made a given decision would also help in regulated environments where we need to defend that reasoning to auditors. Review collected by and hosted on G2.com.

Because it’s built directly into the main system, I can work with all my existing data without having to rely on any messy extra tools. The AI/Intelligence feature is fantastic, since it genuinely understands what’s going on in my files and can handle tough jobs on its own. Review collected by and hosted on G2.com.

What I dislike the most is how difficult it is to figure out the usage limits and consumption patterns. Keeping track of how many tasks the system is handling behind the scenes feels confusing, making it really hard to understand our overall system activity levels. Review collected by and hosted on G2.com.

In this era of automation, ServiceNow AI Agents have outdone other platforms by letting AI take care of the first few steps before I even have to think about them. There’s also no need for constant supervision. Overall, the platform feels more like a perfect team member than just a tool. Review collected by and hosted on G2.com.

Agents are effective and responsible for sure, but it took us time to start trusting them before giving them more responsibility. I have also felt that AI occasionally lacks the real business context behind any unusual request. Review collected by and hosted on G2.com.

What I like most about ServiceNow AI Agents is how they cut down on manual effort by automatically taking care of tasks such as ticket assignment and offering resolution suggestions. They integrate smoothly into existing workflows without causing disruption, and they continue to improve over time, which helps make my work faster and more efficient. Review collected by and hosted on G2.com.

One thing I dislike about ServiceNow AI Agents is that sometimes the recommendations are not fully accurate, especially in complex or uncommon scenarios. It still requires manual verification, which reduces complete dependency on AI.

Also, initial setup and fine-tuning can take time and effort to get the best results. Without proper configuration and quality data, the performance may not be as effective as expected. Review collected by and hosted on G2.com.

Embedding adaptive decision-making directly into core data structures has eliminated many of the rigid limitations I used to run into. With this structural capability in place, systems can self-heal safely, handle and resolve complex incidents, and govern enterprise actions with high precision while keeping latency minimal. On top of that, the platform’s customer support is professional, and their responses are consistently very fast. Review collected by and hosted on G2.com.

The biggest issue for me is that monitoring agent activity can become difficult when several agents are running at the same time. Keeping track of their actions and outcomes sometimes requires extra effort. Review collected by and hosted on G2.com.

I really appreciate that ServiceNow AI Agents save our team a lot of time on repetitive work, like manually reviewing each ticket or searching for the right information. These AI agents provide relevant context and recommend the next action, which has improved our efficiency. The biggest value for us is how it removes representative administrative work, reducing manual effort in handling incident service requests and employee support. This automation, with its useful suggestions, has improved response times, reduced ticket backlogs, and freed up our support team to focus on more pressing issues. Setting it up was pretty easy, especially since we were already using ServiceNow. Expanding it to Microsoft Teams was also straightforward. Review collected by and hosted on G2.com.

I don't like that the quality of the results depend heavily on your data. If your knowledge articles are not well maintained, the recommendations won't always be accurate. Some advanced automations take time to configure and test before they are ready for production. So there is a bit of a learning curve during implementation. Review collected by and hosted on G2.com.

ServiceNow AI work in ServiceNow and get the required data from CMDB and figure the solution to get the work done. The best thing is that it not only answer the quarries it just plan and take action automatically. The user interface is very easy to use for us when any serious customer case come. Review collected by and hosted on G2.com.

It’s very hard to predict the exact price before using it, because adding more calling tools and third-party integrations can make the bill much higher. Review collected by and hosted on G2.com.

What stands out most is how straightforward it is to build agents. I can stand up a working agent for a workflow without heavy custom development, which keeps the focus on the process rather than the plumbing. We mainly use it for workflow automation, and it fits into our existing ServiceNow flows cleanly, so handoffs that used to require manual steps now run on their own. The clearest benefit has been faster resolution times: routine requests that previously waited in a queue get handled and moved along much sooner, which has taken real pressure off the team. Review collected by and hosted on G2.com.

The main gap I’ve run into is the documentation. When I’m working on more advanced agent configurations, the available material doesn’t always go deep enough, so I often end up figuring things out through trial and error or by digging through community threads instead of finding a clear, authoritative reference. It’s more of an annoyance than a blocker, and it hasn’t stopped me from getting agents up and running. Review collected by and hosted on G2.com.

What I like most about ServiceNow AI Agents is that they save a lot of time on the small things that add up during the day. We started using the AI capabilities mainly for ticket summaries and suggested responses, and honestly, that reduced a lot of context switching for our team. During handoffs between shifts, agents no longer have to read through long incident histories to understand what’s going on. The summaries are usually accurate enough to get you up to speed quickly. I also like that the AI is built into existing workflows instead of feeling like a separate tool we have to manage. Review collected by and hosted on G2.com.

One thing we’ve noticed is that the AI output still needs human review, especially for more complex requests or incidents with a lot of context. For straightforward tasks it’s helpful, but sometimes the summaries or suggested responses miss nuances that an experienced agent would catch. There was also a bit of a learning curve in figuring out where the AI added value and where manual handling was still the better option. It’s definitely useful, but it isn’t something you can put on autopilot and forget about. Review collected by and hosted on G2.com.

I use ServiceNow AI Agents to improve efficiency in customer support and service experience. It automates repetitive tasks, streamlines ticket handling, and supports knowledge management, reducing the overall workload and improving response time. Its reliability is a big deal in my department as it manually routes tickets and ensures consistent handling of common requests. I appreciate how it helps make the support process feel stable and predictable. Setting it up was fairly straightforward and just required some coordination with all the teams. I also like using it alongside other support and workflow tools, as it typically works together with the main ServiceNow platform for ticketing and incident management, complementing communication tools like emails or chat systems. Overall, I haven't run into any issues, and the experience has been smooth and reliable. Review collected by and hosted on G2.com.

It's not always accurate with the response, but no AI is. Review collected by and hosted on G2.com.