There's a shift happening in how people find local businesses. It's not dramatic. It's not headline news. But if you run a local service business — a law firm, a financial advisory practice, an HVAC company, a med spa — it's going to quietly shift where your best leads come from: away from traditional search results and toward a handful of AI-recommended businesses.
People are starting to ask instead of search. And that changes almost everything about how local discovery works.
The Shift From Searching to Asking
For twenty-five years, finding a local business meant typing something into Google and picking from a list. The game was visibility — show up high enough on the list and you'd get the click.
That game still exists. But a parallel game has started running alongside it.
When someone asks ChatGPT "who's the best estate planning attorney near me in Coeur d'Alene," they don't get a list. They get a name. Maybe two. The AI synthesizes what it knows, weighs its confidence, and gives a single recommendation it's willing to stand behind.
Meta Ray-Ban glasses have sold over 7 million units. Google, OpenAI, and Apple are all building or refining screenless and wearable AI devices. When someone asks their glasses "find me a good financial advisor near Spokane" while walking to their car — they hear a name in their ear. There's no screen. There's no list. There's one answer.
I wrote about the structural mechanics of this in The Shortlist Problem. The short version: audio interfaces don't accommodate lists. They pressure AI systems toward one high-confidence recommendation. The businesses that make that recommendation win the lead. The businesses that don't aren't ranked lower — they're simply not in the conversation.
What Determines Who Gets Recommended
In practice, AI systems tend to favor local businesses that are easy to understand, verify, and trust. The signals they rely on are well documented.
Structured data.
JSON-LD schema tells AI systems exactly who you are, what services you offer, where you operate, and what makes you credible. Businesses with custom per-page schema give AI systems clear, machine-readable facts to work from.
So what: pages with clear schema are easier for AI to cite with confidence than pages without it.
Machine-readable identity files.
llms.txt is a plain-language brief that tells AI crawlers who you are, what you do, and which pages matter most. agent.json is a structured identity file that encodes your business for autonomous AI agents. Together, they're the "about us" and "business card" your website hands to AI systems before they decide who to recommend. Most local business websites have neither.
So what: without these files, AI systems have to guess who you are and what you do. They usually don't bother guessing.
Directory consistency.
AI systems cross-reference your business information across Google Business Profile, Bing Places, Apple Business Connect, Yelp, and BBB. Inconsistencies in your name, address, or phone number create ambiguity.
So what: if your naming is messy across directories, the AI has to guess which version is right — and it usually just skips the guess entirely.
Review signals.
Volume, recency, and specificity of reviews all factor in. Detailed, location-specific reviews carry more weight than generic five-star ratings.
Site performance.
AI crawlers prioritize sources that respond quickly and cleanly. This is where most small business owners have a blind spot — and it's worth examining honestly.
The WordPress Conversation Worth Having
Let's be clear about something: WordPress is not the enemy. A well-maintained WordPress site, properly configured with good hosting, updated regularly, and kept lean, can perform well. Some do.
The honest question isn't "is WordPress capable?" It's "is your WordPress site actually being maintained that way?"
For most small and medium-sized local businesses, the answer is no — and not because the owners don't care. It's because they're running a business. Plugins accumulate. Theme updates get skipped. Hosting plans don't get reviewed. A site that loaded in 1.2 seconds at launch is loading in 3.8 seconds two years later, and nobody noticed because the degradation was gradual.
Performance matters to AI crawlers for a straightforward reason: they're scanning enormous numbers of pages and they prioritize sources that respond quickly and cleanly. A slow, bloated site doesn't get penalized in some formal sense — it just gets less attention, less thoroughly indexed, and less confidently cited.
According to web performance data, the median WordPress site with active plugins loads in 3–5 seconds on mobile. The sites on the faster end of that range are the ones with dedicated technical attention. Most local business sites aren't getting that attention — not because the owners are negligent, but because it's genuinely hard to sustain without ongoing technical resources.
Where Edge Infrastructure Actually Helps
This is the honest case for edge hosting — not that it beats WordPress on a spec sheet, but that it removes the maintenance variable that causes most small business sites to drift.
A Next.js site deployed on a platform like Vercel's global edge network is fast by default and stays fast. There are no plugins to update, no theme conflicts to resolve, no hosting plan to optimize. The performance you get at launch is the performance you keep with almost no ongoing technical babysitting. For most local firms, that's the difference between "fast for a month" and "fast for years."
For a local attorney or financial advisor who is not a technical person and doesn't have a developer on retainer, this matters practically. You get consistent, predictable site performance without having to think about it.
Edge caching extends this further. Your site's files — including llms.txt and agent.json — get distributed to server locations around the world. When an AI agent queries your site, it's hitting a nearby server instead of routing all the way to your origin. Response times are faster and more reliable regardless of where the query originates or what kind of device is making it. It's one consistent signal among many — and consistent signals compound over time in ways that inconsistent ones don't.
The Honest Case for Acting Now
I want to be straightforward about what we know and what we're reasoning toward, because I think that distinction matters.
What we know: AI assistants are already answering local business queries. Wearable devices that rely on audio interfaces are already in the market and growing. Audio interfaces pressure AI systems toward single recommendations. Structured data, identity files, directory consistency, and site performance are all documented inputs into how AI systems evaluate and recommend local businesses.
What we're reasoning toward: As wearable adoption grows, the share of local business discovery that happens through AI recommendation rather than traditional search will increase. The businesses that have built clean, structured, machine-readable infrastructure before that shift compounds will be better positioned than the ones that build it after.
That's not a guaranteed outcome. It's a well-reasoned bet — and the key is that the same work helps you today in ChatGPT, Perplexity, and Google AI Overviews, even if wearables grow slower than expected. It's not a wager on one specific future. It's a foundation that pays off across multiple scenarios.
What This Actually Looks Like in Practice
The Four-Layer Authority Engine that KodeCite builds into every site:
Layer 1 — Edge Speed
Next.js deployed on Vercel's global edge network. Consistent sub-second performance on real devices — not because someone is actively maintaining it, but because the architecture produces it by default and keeps producing it without ongoing technical babysitting.
Layer 2 — Per-Page Schema
Custom JSON-LD on every page — homepage, service pages, location pages, articles, FAQs — encoding your specific expertise, geography, and authority as machine-readable facts. Zero invalid items at launch, verified before go-live.
Layer 3 — llms.txt
Your plain-language brief for AI crawlers. Who you are, what you do, where you operate, which pages matter most. The "about us" your site hands to every AI system that visits before it decides who to recommend.
Layer 4 — agent.json
Your structured identity file for autonomous AI agent discovery. The "business card" that tells AI agents exactly how to describe you, recommend you, and — as agent-to-agent commerce develops — interact with your business on behalf of users.
The Practical Starting Point
If you're a local service business owner wondering where you actually stand, the starting point is a clear-eyed audit of your current presence.
How does your business currently appear when someone asks ChatGPT or Perplexity about your category in your market? What does your PageSpeed score look like compared to your top three local competitors? Do you have any structured data — and if so, is it valid? Is your NAP data consistent across every major directory?
Most local businesses have never looked at these things. That's the gap — and it's also the opportunity, because most of your competitors haven't looked either.
Not sure where you stand?
Get a free AI Scaffolding Audit — we'll show you exactly how your business currently appears in ChatGPT, Perplexity, and Google AI Overviews. By the end of the audit, you'll know whether you can get away with tuning what you already have or whether you're better off with a rebuild. No pitch. No pressure.
Get Your Free AEO AuditI'm not trying to sell you on a crisis that doesn't exist. I'm trying to give you an honest picture of where local business discovery is heading — and the same foundation my clients are already building. The window is open. Whether you walk through it is your call.
— Mark, KodeCite.ai