iNDX CO+ Update - June 23-July 5: A New Way To Reach Customers In The AI Era
I am seeing an emerging new rail for reaching customers. It is early, but I think it is important enough that every business should start paying attention now.
The shift is simple: more consumers are starting to use AI as their main portal to the digital world. Instead of going to websites, apps, search engines, social feeds, or inboxes first, they are asking AI to bring the right information to them.
That creates a big question for every brand: How do we reach customers when the customer’s AI becomes the front door?
Part of the answer is GEO/AEO, or generative engine optimization and answer engine optimization. I covered this in the Master Findability framework. This is the new form of SEO: structuring your content so AI systems can find it, understand it, trust it, and use it in answers.
But I think there is another piece emerging that may become just as important as email lists and social posts.
That piece is skill packs (discussed in our last email: https://indx.earth/community-summaries/indx-co-update-june-8-june-22-wow-we-just-took-a-2-year-leap-forward-for-paddlenet-and-every-business-everywhere-and-get-yourself-to-the-colab-in-october-heres-why.html) connected to MCP.
Think of MCP as a door that lets an AI model access a brand’s content, tools, products, offers, events, or data. A business creates an MCP endpoint. A user adds that endpoint to their AI tool of choice. Then, when the user asks a question, or when their AI runs a workflow for them, the AI can open that door and pull in the right information.
This could become a new customer subscription rail.
Today, we send emails to subscribers. Those emails might include content, products, offers, event dates, resources, updates, and announcements. But the customer still has to open the email, scan it, filter it, and decide what matters.
What if their AI did that for them?
One customer might only want offers. Another might only want events. Another might want technical resources. Another might want product updates based on what they already bought.
Instead of every customer receiving the same email and doing the filtering themselves, their AI could query the brand, pull the latest information, filter it around that customer’s needs, and present only what matters.
That is powerful.
It is less work for the customer. It is more relevant. It should create better engagement. And for the business, it could mean less pressure to keep creating new messages for every channel. If the content is already published and structured correctly, AI can retrieve and use what it needs.
I am already seeing this emerge in the tech world. Tools like Codex Desktop and Claude Desktop already let users add MCP servers or custom connectors. This is still early, but the pattern is clear.
I have not yet seen enough around attribution and tracking. Businesses will need to know what customers are seeing, what the AI is pulling, and whether it is working. I expect that layer will emerge.
But you can already see where this goes.
What if a customer adds your MCP endpoint and connects it to their email address or customer profile? What if their AI can access their transaction history, preferences, event registrations, or past purchases? Now the AI can filter your content around that specific customer.
That is the long-promised idea of one-to-one marketing, but finally with the technical pieces starting to come together.
We have email. We have social. We have text. Now we may have MCP-based access, a far more superior way to communicate with our customers
My goal is not to turn everyone into a developer. My goal is to help you see where this is going so you can start planning now.
Do not look at this and say, “This is too technical. I cannot afford this. I cannot build this.”
This will get easier. It will get cheaper. The tools will improve. The important work right now is to think strategically:
- What do you want customers to receive from you?
- What information should their AI be able to access?
- What content, offers, events, products, or services should be structured now so this becomes easier later?
I am already working on this for PaddleNet and thinking through how I might use it in my pet business.
You should start thinking about it for your business too.
If you want to brainstorm, reply to this email and we can set up a virtual meeting, or join one of my open office hours. The more we talk through these emerging tools together, the faster we all learn. We cannot do this alone. We have to crowdsource the learning and help each other. That is how we win in the AI era: individually in our businesses, and collectively across our industries, communities, and economy.
Findability: Main Category
Here are recent links I found helpful.
6 content audit workflows to build in Claude (or Codex or any agent)
Good practical workflows for using Claude, Codex, or another AI agent to run content audits faster. Helpful if you want to start turning content review into a repeatable AI-assisted process.
https://community.indx.earth/t/6-content-audit-workflows-to-build-in-claude-or-codex-or-any-agent-/184
https://searchengineland.com/content-audit-workflows-claude-481099
Free alternative to Semrush and Ahrefs
Open SEO looks like a useful open-source alternative to expensive SEO tools like Semrush and Ahrefs. It also includes an MCP server and skills, which means it can connect into AI agent workflows. You can self-host it or run it through Cloudflare.
https://community.indx.earth/t/free-alternative-to-semrush-and-ahrefs/183
https://github.com/every-app/open-seo
Llms.txt is under pressure
Ahrefs reported that most llms.txt files in its study received no requests in May 2026. The implication is clear: keep llms.txt, but do not rely on it alone. The stronger strategy is still structured, crawlable, semantically organized content with clear entity, trust, provenance, and discovery signals.
https://community.indx.earth/t/llms-txt-is-under-pressure/182
https://ahrefs.com/blog/llmstxt-study
Announcing the Agentic Resource Discovery specification
Google announced a new Agentic Resource Discovery specification with support from several major platforms. This reinforces the direction of the Master Findability Framework: make content easier for AI agents to discover, understand, and use.
https://community.indx.earth/t/announcing-the-agentic-resource-discovery-specification/181
https://developers.googleblog.com/announcing-the-agentic-resource-discovery-specification
When AI recommends your brand, buyer behavior changes
This is useful data confirming what many of us already suspected: AI recommendations can influence downstream search, website visits, and product discovery. The practical implication is that being visible to AI systems is becoming part of the customer acquisition path.
https://community.indx.earth/t/when-an-ai-platform-recommends-your-brand-to-someone-new-that-person-becomes-182-more-likely-to-search-you-on-google-117-more-likely-to-visit-your-site-and-185-more-likely-to-view-your-products-on-a-retailer-s-page-within-the-week/180
https://scrunch.com/blog/prompt-to-purchase-pipeline-how-ai-influences-buyer-behavior
Content Marketing Is Getting Technical. That’s Not the Problem
This is a good look at the leading edge of SEO and GEO. The technical work is becoming more accessible, but the deeper issue is business strategy. If a business has to constantly chase every new technical tactic, that is hard to sustain. The better path is to build durable differentiation through strong content structure, clear entities, trust signals, and AI-ready publishing.
https://community.indx.earth/t/content-marketing-is-getting-technical-that-s-not-the-problem/179
https://projects.eddiesoehnel.com/adminprojects/seo-geo-OPEN
https://inatoncheva.substack.com/p/content-marketing-is-getting-technical