Agentic AI is the latest generation of AI technology changing the way businesses operate. These autonomous AI agents can work on people’s behalf to complete tasks and go through multiple steps to achieve their goals.
Common tasks for AI agents range from handling customer service enquiries to automating admin for talent onboarding. But AI agents can be put to work and benefit every department of the organisation.
While their capabilities are phenomenal and beyond anything we used just a couple of years ago, it’s important to remember that AI agents have their limitations too. In a word, they are built to complete specific tasks.
So an AI agent that’s great at analysing sales data to alert a manager about inventory requirements is very different from one that schedules patients’ appointments at a local hospital. Each AI agent has its own set rules and scope: the data sources it can draw from, the applications it can use, and so on. And other platforms have their own data sets and their own AI agents too.
Two heads are better than one
Those limitations are put in place for security reasons, as well as technological practicalities. IT teams want to keep the impact of AI agents within certain bounds to avoid potential compliance issues or the risk of data breaches. And a tight scope ensures human overseers can more easily manage the agents’ actions, clearly identifying where it’s making a positive impact as well as quickly fixing any issues.
But the reality is that everything in the IT world relies on access, not just to a single source of data, but to multiple sources. No system, including the ServiceNow platform, has every piece of information that an organisation needs to operate effectively. It’s essential to everyday work that your technology can talk to other systems and get the data required for a particular job.
So the fact that AI agents operate specifically within certain tasks and applications creates a new need among businesses: for greater AI agent integration. As organisations deploy more AI agents across their workflows, the key next step is to connect them to others outside their walls.
What is Agent-to-Agent collaboration?
The good news is that this capability is coming soon. Thanks to a new industry-wide protocol known as Agent-to-Agent or A2A collaboration, organisations will be able to work more seamlessly with each other.
The A2A protocol is an open standard, developed by Google and announced in April 2025, that will enable AI agents to communicate, collaborate, and share information across different platforms and frameworks.
This will allow agents to discover each other's capabilities, interact across applications and data sets, and collaborate on tasks without exposing their internal states. The aim is to promote a more interconnected, powerful, and innovative AI ecosystem.
Google’s aim is to have the A2A protocol available for any business to use later in 2025. As an open-source technology, supported by ServiceNow and dozens of other major industry players, it will be the de facto standard for how AI agents interact.
How ServiceNow can support you?
The pathway is set for the future of enterprise IT. Every vendor and every application will have AI agents that can communicate with each other.
For core platforms like ServiceNow, which form the bedrock of how lots of organisations run today, this creates incredible new opportunities to integrate with other digital infrastructure systems. The challenge is to ensure all of these integrations are smooth and seamless, simplifying rather than complicating how organisations work.
Fortunately, ServiceNow offers a series of tools to help you make sure your AI investment is working effectively.
ServiceNow’s AI Control Tower provides a dashboard where you can monitor and optimise your AI systems, all in one place. AI Agent Orchestrator ensures your digital workers are operating in harmony with each other. And AI Agent Fabric underpins how your agents interact with other companies’, via zero copy integration. Read a full explanation of these ServiceNow AI tools here.
In short, the support you need for greater agentic AI collaboration is already in place. That sets up ServiceNow users well to reap the coming benefits of A2A.
Get ready for AI Agents on steroids
The public launch of the A2A protocol will be a watershed moment for agentic AI. While there’s no firm date for it yet, we already know that by the end of this year there will be hundreds of AI agents available on ServiceNow, ready to connect with other key vendors and their applications.
The onus is on businesses and AI users to be ready for it all. Those that are prepared will have a competitive advantage. Those caught napping will risk getting left behind when A2A becomes available.
So what should you do to get an edge over your peer group? Here are two simple steps we recommend.
First, keep your ServiceNow platform on the latest release and upgrade as soon as you can. That will ensure you can benefit from these new A2A capabilities immediately as they’re rolled out.
Second, embrace A2A in a spirit of experimentation. Agentic AI integrations are there to be used. Start trying out AI Agent Orchestrator and AI Agent Fabric now, and see how they can make your productivity better.
Start small, learn fast, and scale boldly. Agentic AI is only just beginning.
If you're looking to get started with ServiceNow and AI, or you want to make sure you're getting maximum value out of your AI deployments, learn more on our ServiceNow and AI page.
