What Is Agentic Resource Discovery (ARD)? The Beginning of AI That Finds Businesses for Customers
2026-06-30
By Morten Siert Eriksen
For twenty years, getting found meant ranking on a search engine. The AI ecosystem is slowly changing this. More and more customers use an AI assistant like ChatGPT, Claude or Gemini to just handle it: "find me a plumber who can come this week," "book a table somewhere good nearby," "who can redesign our website?"
For that to work, AI needs a way to discover what's actually out there — the tools, services and businesses that can help with a given task. That's exactly what a new open standard called Agentic Resource Discovery (ARD) is being built to do. It's early days, but it points to a real shift in how customers and businesses find each other.
Topics
From searching to asking
The way people find things online is changing in a fundamental way.
For two decades the pattern was the same: type a few words into a search engine, scan a page of links, click around, compare, and decide. The business that ranked well got the click. Today a growing number of people skip all of that and simply ask an AI assistant to do it for them — and increasingly, to act on it too.
While this is practically working today, AI assistants can only suggest what they’re trained to know, or what it can find on the Internet. If businesses want to deliver a custom AI experience to their customers through the AI assistants, the customers will have to know about the tools, resources and custom AI Agents up front and have these pre-connected to their device. This effectively prevents the customer from ever finding new, aspiring brands, products and services. "A customer cannot use a capability it does not know about".
That gap — between what customers ask AI to do and what AI can actually find — is what the new standard ARD is dealing with. (how customers use AI to find businesses)
For two decades the pattern was the same: type a few words into a search engine, scan a page of links, click around, compare, and decide. The business that ranked well got the click. Today a growing number of people skip all of that and simply ask an AI assistant to do it for them — and increasingly, to act on it too.
While this is practically working today, AI assistants can only suggest what they’re trained to know, or what it can find on the Internet. If businesses want to deliver a custom AI experience to their customers through the AI assistants, the customers will have to know about the tools, resources and custom AI Agents up front and have these pre-connected to their device. This effectively prevents the customer from ever finding new, aspiring brands, products and services. "A customer cannot use a capability it does not know about".
That gap — between what customers ask AI to do and what AI can actually find — is what the new standard ARD is dealing with. (how customers use AI to find businesses)
What is Agentic Resource Discovery (ARD)?
Agentic Resource Discovery — ARD for short — is an open standard that gives AI assistants a way to ask one simple question: what's available for this task?
Instead of being stuck with a fixed set of built-in tools, an assistant can search a discovery service and get back a shortlist of matching options. This can be tools, services, AI agents representing businesses and brands. This way clothing brands can expose their custom size-calculator to customers asking if the design fits them, and local carpenters can have customers see their availability and make an appointment after prompting for carpenters in the area.
ARD is not replacing MCP. An MCP server can be a resource made available through ARD, so customers can more easily find them. Where MCP servers unlocked integrations to the AI assistants, ARD unlocks the ability for anyone to find and use the MCP servers (and other resources as well).
ARD was recently announced and released and is at the moment supported by contributors like Google, Microsoft, Salesforce, Nvidia, Cisco, Databricks, GitHub, GoDaddy, Hugging Face, ServiceNow and Snowflake. Like with other standards, more supporters and better adoption will hopefully come.
In plain terms: ARD is the emerging standard which is meant to help businesses be found by customers using AI assistants.
Instead of being stuck with a fixed set of built-in tools, an assistant can search a discovery service and get back a shortlist of matching options. This can be tools, services, AI agents representing businesses and brands. This way clothing brands can expose their custom size-calculator to customers asking if the design fits them, and local carpenters can have customers see their availability and make an appointment after prompting for carpenters in the area.
ARD is not replacing MCP. An MCP server can be a resource made available through ARD, so customers can more easily find them. Where MCP servers unlocked integrations to the AI assistants, ARD unlocks the ability for anyone to find and use the MCP servers (and other resources as well).
ARD was recently announced and released and is at the moment supported by contributors like Google, Microsoft, Salesforce, Nvidia, Cisco, Databricks, GitHub, GoDaddy, Hugging Face, ServiceNow and Snowflake. Like with other standards, more supporters and better adoption will hopefully come.
An agentic resource published once should be discoverable by any user — and a user should be able to find useful resources well beyond the small set they already know about.
In plain terms: ARD is the emerging standard which is meant to help businesses be found by customers using AI assistants.
Why this matters for your business (and customers)
Hopefully our introduction to ARD is starting to paint the picture.
Even though online search engines are adopting AI searches and responses, the interactions between the brand or business are mainly text-based. With ARD, businesses can answer prompts with more than just text. Businesses can expose their custom product catalogs, price calculators, AI customer service agents, checkout flow, booking systems to give the customer a more engaging experience directly in the AI assistant.
For brands and businesses this means another channel of serving new and existing customers — and an entire new way of delivering the experience to the customer.
ARD is built as an open standard. Instead of manually registering with every AI assistant like Gemini, ChatGPT, Claude, etc. and redoing it whenever something changes, you describe your business, services and product once, and it can be discovered across the ecosystem.
Even though online search engines are adopting AI searches and responses, the interactions between the brand or business are mainly text-based. With ARD, businesses can answer prompts with more than just text. Businesses can expose their custom product catalogs, price calculators, AI customer service agents, checkout flow, booking systems to give the customer a more engaging experience directly in the AI assistant.
For brands and businesses this means another channel of serving new and existing customers — and an entire new way of delivering the experience to the customer.
ARD is built as an open standard. Instead of manually registering with every AI assistant like Gemini, ChatGPT, Claude, etc. and redoing it whenever something changes, you describe your business, services and product once, and it can be discovered across the ecosystem.
ARD has potential to change AI
Agentic Resource Discovery was just released. It’s in an early version and still needs improvements. But the potential is huge.
While the first implementations of ARD are appearing now, the scope seems narrow and lacks a consumer-facing angle. As plugins, connectors and apps for AI assistants are mainly professional tools at the moment, ARD feels like a standard facing a tech-savvy audience and not the everyday user of AI assistants.
But the early days scope and limitations are all expected. Take for instance how users search the Internet with AI assistants at the start; the user had to actively select a connector or plugin to allow the AI assistant to use the Internet. Now the AI assistant does this by itself. A shift like this seems very likely for finding tools, resources and other AI integrations in the future.
If ARD is not becoming the way of indexing and finding resources in the AI ecosystem, another standard will. Discoverability is a necessity for everyday users to utilize the full potential of AI. Considering the contributors like Google, Microsoft, Salesforce, Nvidia, Cisco, Databricks, GitHub, GoDaddy, Hugging Face, ServiceNow and Snowflake, ARD is here to stay. And we are very excited about it.
While the first implementations of ARD are appearing now, the scope seems narrow and lacks a consumer-facing angle. As plugins, connectors and apps for AI assistants are mainly professional tools at the moment, ARD feels like a standard facing a tech-savvy audience and not the everyday user of AI assistants.
But the early days scope and limitations are all expected. Take for instance how users search the Internet with AI assistants at the start; the user had to actively select a connector or plugin to allow the AI assistant to use the Internet. Now the AI assistant does this by itself. A shift like this seems very likely for finding tools, resources and other AI integrations in the future.
If ARD is not becoming the way of indexing and finding resources in the AI ecosystem, another standard will. Discoverability is a necessity for everyday users to utilize the full potential of AI. Considering the contributors like Google, Microsoft, Salesforce, Nvidia, Cisco, Databricks, GitHub, GoDaddy, Hugging Face, ServiceNow and Snowflake, ARD is here to stay. And we are very excited about it.
Be ready to be found by AI
Even though it’s early days of Agentic Discoverability, we at Fugentic are already preparing for it. We introduced the term Agentic Discoverability back in 2025 as we expected the AI ecosystem to reach this level of maturity eventually (read our introduction to Agentic Discoverability from 2025 here).
While ARD and preparing a business for the Agentic Discoverability are important, it does require some technical know-how. These challenges Fugentic is aiming to remove, making the Agentic Discoverability a space for any brand or business.
We suggest you sign up to stay up to date. While we are tracking the development of the industry, we are preparing the platform to help you be found by your customers using AI.
While ARD and preparing a business for the Agentic Discoverability are important, it does require some technical know-how. These challenges Fugentic is aiming to remove, making the Agentic Discoverability a space for any brand or business.
We suggest you sign up to stay up to date. While we are tracking the development of the industry, we are preparing the platform to help you be found by your customers using AI.