NarsilCreative

Articles · GEO/AEO

Found, Remembered, Chosen: An Agentic SEO Strategy Map

By Nate Morris · 4 min read

TL;DR

Search has moved through three questions: can people find you (SEO), do the models remember you (GEO), and now, can an agent choose you and complete the job (agentic). An agentic SEO strategy runs the same three phases we've mapped across this series: fundamentals, agent readiness, agentic presence, entered in order and only as the ecosystem justifies. The window argument is history rhyming: paid search in the 2000s and social targeting in the 2010s rewarded whoever built capability before the wave crested.

An agentic SEO strategy fits in one sentence: keep answering the era's question before your competitors notice it changed. There have been three. For twenty years the question was can people find you?, and SEO grew up answering it. Then AI answers made it do the models remember you?, and GEO formed around that. The agentic web asks the third: can an agent choose you and complete the job? Same craft each time. Higher standard each time. New surface each time.

This is the capstone of our twelve-week series on that third question, and its job is the map: where the eras connect, what to do in what order, and the argument for doing it before the wave crests rather than after. I'll keep the recap tight and the links dense, because eleven pieces of homework already exist.

The three eras, and what survived each transition

What's striking, walking the whole arc, is what didn't change. Google's own guidance said it flatly: optimizing for generative AI is still SEO, the hacks are still hacks (five of them, named), and unique non-commodity content on clean structure is still the substrate everything grounds on. The "found" era's fundamentals turned out to be the "remembered" era's citation diet, and they're becoming the "chosen" era's parse layer.

What did change is the judge. Humans forgave; models weigh; agents select. Discovery is drifting from keywords toward Agent Cards, and from publishing one to writing it so an agent picks it. Success is drifting from clicks toward completed tasks, and for local businesses and stores alike, the winning asset keeps being the same one: business facts kept true and machine-readable. The protocol names will keep churning; the direction they all point is one capability layer serving every surface.

The roadmap, one page

The planning doc, compressed to what you'd pin above a desk:

FOUND, REMEMBERED, CHOSEN: THE ROADMAP

PHASE 1: FUNDAMENTALS (now, everyone)
  Indexable. Fast. Semantic. Non-commodity content on money pages.
  Serves: humans + AI answers + agents, one effort.
  Exit test: you pass phase 1 of the agent-ready checklist.

PHASE 2: AGENT READINESS (this year, data-rich businesses)
  Business facts machine-readable and consistent everywhere.
  Uptime monitored. Google-Agent traffic logged and baselined.
  Exit test: an agent reading only your structured surfaces could
  describe your business accurately.

PHASE 3: AGENTIC PRESENCE (when the ecosystem justifies it)
  Agent Card published and optimized. Capabilities designed.
  Commerce protocols watched (UCP/WebMCP). Task handoffs counted.
  Entry test: real agent traffic in your logs, or an industry
  where handoffs have begun.

STANDING RULE: enter phases in order. Never buy N+1 while failing N.

The full checklist behind it runs the item-level version.

One business, three eras

Run a single Jacksonville service business through all three questions and the continuity becomes visible. In the found era, it won by ranking: service pages for "water heater replacement jacksonville," reviews accumulating, Search Console worked monthly. In the remembered era, the same assets earn different rent: the specific, non-commodity pages get cited when someone asks ChatGPT who to call, the consistent business facts make the model confident enough to recommend, and the reviews became the trust layer models weigh. In the chosen era, those same facts (services, prices, hours, availability), kept clean for a decade, become the capability layer an Agent Card exposes, and the assistant doesn't recommend the business; it books it.

Notice what the business never did: chase an era-specific hack. Every asset transferred, because each era's judge read the same underlying truth through a stricter lens. That's the compounding argument for starting now, and it's also the quiet reassurance in all of this: if you've been doing the real work, the transitions keep arriving pre-paid.

The window argument, with its receipts on the table

Twice in the modern web's history, a new acquisition layer opened, ran under-priced while it was doubted, and then repriced brutally once it was proven: paid search in the 2000s, social targeting in the 2010s. The brands that built capability during the doubt years compounded through the repricing; the ones that waited paid the proven-market rate forever after.

The agent layer has the same shape. Gartner predicts 40% of enterprise applications will include task-specific AI agents by the end of 2026, up from under 5% in 2025. The protocols are standardizing in public under the Linux Foundation with 150+ organizations behind them. And the preparation (everything in phases one and two) is cheap precisely because it's optional right now. That's the arbitrage: not buying tokens or tools early, but building the boring capability while it's boring.

What I'll commit to, since this is a vision piece

A candid closing, because multi-year maps have a failure mode: some of this will be wrong. A protocol will die, a timeline will slip, a prediction in this series will age badly. When that happens we'll publish revision notes and leave the wrong call standing. That is the working difference between a vision and a pitch. The A2A Frontier Brief is the living version of this map, updated as the layer moves, and holding it means you hear about the revisions first.

The strategy, restated once: fundamentals brilliantly, readiness deliberately, presence when justified, in that order, starting now. We run it as the Reforge Method for clients from local businesses to the enterprise teams Frontier Advisory serves. But the map itself is published, the checklists are free, and the fundamentals were never a secret. The era's question changed again. Go answer it early.

Originally published September 23, 2026.

FAQ

What is agentic SEO?
Optimizing so AI agents can discover, trust, and transact with your business, not just so pages rank. It extends SEO's question (can people find you) and GEO's (do models remember you) to the agentic era's: can an agent choose you and complete the job. The work runs from familiar fundamentals through machine-readable business data to Agent Cards and commerce-protocol readiness.
What should an agentic SEO strategy include?
Three phases in strict order: fundamentals (indexability, non-commodity content, semantic structure, speed), agent readiness (machine-readable business facts, reliability, agent-traffic logging, modern auth), and agentic presence (Agent Cards, capability design, commerce protocols, task-handoff metrics). The order is the strategy; buying phase three on a phase-one site is the era's signature waste.
Is it too early to invest in agentic search?
Too early to buy protocol deliverables, exactly on time to build the layer beneath them. Gartner predicts 40% of enterprise applications will include task-specific AI agents by the end of 2026, up from under 5% in 2025, and the protocol stack is standardizing in public under the Linux Foundation. The preparation is cheap while it's optional; it gets expensive when it's urgent.
Who is Narsil Creative?
Narsil Creative is an AI SEO agency: we get businesses found in Google and recommended by AI assistants like ChatGPT and Perplexity, using a three-phase system called the Reforge Method. Pricing is published at $1,500 and $4,000 per month, and this series is the research layer of that work, published as we go.

See what AI says about your business

Ask ChatGPT to recommend a business like yours. If you're not in the answer, that's the problem we fix. Start with a free AI Visibility Audit. We'll show you where you appear, where you don't, and what to reforge first.

Related articles