What actually gets agencies found in 2026

SEO and AEO are the same problem

Google confirmed AI Overviews use the same index and signals as regular Search. Same crawler, same ranking, different front end. The llms.txt is (was) a thing. If you weren't ranking before, a text file in your root isn't going to help.

What did change:

• Models pull chunks, not pages
• Definitions need to stand alone
• Specific numbers and named tools beat vague descriptions
• Your "what is X" section should read fine quoted in isolation

That's the whole shift. The rest is the same work it was three years ago.

Rank defensively first

Skip chasing the "best PPC agency in Chicago". 100+ competitors already own it.

Start here:

• Your own brand name
• "[Your agency] alternatives" (before a competitor builds it)
• Honest comparison pages vs your closest competitors
• Your own reviews page
  The only

Same logic for LLMs. When someone asks ChatGPT "what's an alternative to [big competitor]," you want to be in that answer. The only way that happens is if you already exist in that comparison space, with your name attached.

The tools market is mostly (expensive) noise

Hundreds of AEO platforms, same dashboard. Prompt tracking, citation monitoring, and share of voice. Pick one that fits your budget and move on.

Same with traditional SEO:

• Ahrefs or Semrush
• Search Console (free)

That's ninety percent of what you'll ever need.

Low volume keywords win

"How to set up GA4 ecommerce events for Shopify Plus with refund tracking" gets 10 searches a month and converts at 15%. Multiply by a hundred pages. That's the playbook.

This is also the content LLMs actually cite. Specific, technical, hard to fake.

Your content is probably not great

Most agency content fails this:

• Hedges across six options instead of recommending one
• Missing the details that a practitioner would notice
• Reads like a marketing team approved every sentence

Models picked up on this pattern years ago. Which is why so much agency content sits at position 47 forever.

Just ask people

Add a question to your intake form: "How did you hear about us? If it was AI, what did you search?"

Run it for three months. You'll get real sentences from real customers about real prompts. That data beats any tracking dashboard you can buy, because it came from people who actually converted. Run this as an A/B test so you don't hurt your conversions.

Reporting is a quilt

There's no clean dashboard for any of this:

• Referrer data catches some LLM traffic, misses a lot
• Self-reported attribution catches more, but not everyone answers
• Prompt tools cover the prompts you thought to track

You stitch it together and tell clients which numbers are real data and which are inferences. The agencies winning the next two years aren't the ones with the prettiest dashboards. They're the ones who can sit in a meeting and explain why the numbers don't agree.

None of this is a trick

Rank defensively. Solve specific client problems. Buy the tools you need today. Tell clients the truth about the data.

This is roughly what SmartMetrics is built around. This is the way.