The search industry is rapidly producing new standards, conventions, and frameworks.
Terms like MCP, A2A, ARD, or LLMs.txt were dropping LinkedIn and industry publications.
As overwhelming as another set of acronyms may seem, most of these standards attempt to solve different problems at different layers of the emerging agent ecosystem.
Mapping These Protocols
These levels affect different parts of the agent’s journey, and impact them in different ways. The map (below) considers whether the protocol is action-oriented (providing agency – driven by the agent) or information-oriented (providing information – considered to be publisher-driven).
This is not just one axis; it helps to think of it as organized like this:
- Action – knowledge.
- Agent – publisher.
These different levels are at different levels of maturity and adoption and mean different things depending on your job and where you spend your time every day.
The five-minute version
Each of these frameworks can be complex and constantly changing, so treat the below as a quick primer to see what/when/how each one works.
In most organizations today, the key command is:
- Understand the landscape.
- Improve availability.
- Highlight skills where appropriate.
- Monitor emerging standards.
Many businesses are already concerned with step four while still struggling with steps one and two.
Official Details
If you think that any of these agreements is what you are looking for, I would strongly suggest that you read their documentation as my (quick) summary above (and the space change of this space), then it is a sure way to find what you need.
Wait, Are Some of These Contestants?
Some of these standards overlap, while others solve adjacent problems. In some cases, their creators present them as alternatives for certain situations, making it easy to think that they are in direct competition when they often complement each other.
A2A, for example, says it’s a better option than MCP servers in some cases, and ARD/WebMCP both seem to have more of the same goal.
This table should help you understand why/when some standards may be used over others.
Overlap is about discovery, persuasion, and orchestration.
What Should We Really Focus On?
AI agents will not interact with websites via a single protocol just as browsers interact with websites via a single HTML tag. The future is likely to include a set of complementary (sometimes conflicting) standards, each addressing a different aspect of the interaction between content, capabilities, systems and functions.
The key is not to adopt every new acronym that appears. It’s about seeing each problem it’s trying to solve, understanding if it’s important to your organization, and keeping an eye on the levels that get real adoption rather than just creating a conversation.
After reading this, if you think you might have to wait a little longer to settle the dust, you’d be forgiven!
That said, there are a few areas I’ll be keeping an eye on:
- If you’re in ecommerce, check out UCP. Shopping and checkout are fast, and this may be one of the first standards many retailers encounter.
- If you expect AI agents to complete actions on behalf of your clients, keep an eye on WebMCP and ARD. Together, they represent one of the clearest directions to go for presenting website capabilities to agents.
- If you’re concerned about AI systems that detect and understand large or complex websites, OKF is the standard I’ll be watching very closely.
Standards do not become standards because they are technically superior. They become standards because enough of the ecosystem accepts them.
Some of the protocols discussed here may become the foundation of the future web, while others may merge, evolve, or quietly disappear. Understanding what they are today is more important than betting on when someone will eventually “win.”
Additional resources:
This post was originally published on Chris Green Search Marketing (SEO/AEO).
Featured image: Rawpixel.com/Shutterstock
