What's the difference between SEO and AEO?

SEO is being found in search results. AEO, answer-engine optimization, is being the answer when someone asks an AI assistant instead of searching. They share a foundation of clean headings, structured data, and consistent facts. AEO adds what assistants quote directly: a plain-language FAQ and an AI information page.

Is this content generation at scale?

No. We draft the specific fixes an audit calls for, from facts already on your pages or in your records. If a fact isn't there, the fix is to add it, not to make it up.

We have dozens of locations. Does this work across all of them?

That's where it helps most. Sites built on one template repeat the same findings, so we audit every location, fix the template issues once, and track each site's score from one place.

Do you change our website directly?

Not on our own. Every draft waits for a person on your side to approve it, and we work with whoever runs your site to put approved changes live.

Do you guarantee rankings?

No, and be wary of anyone who does. What we stand behind is the audit: you'll know exactly what an engine or assistant can read on every page, and what changed after the fixes.

Which businesses is this built for?

We built it first for property management, where one company can run dozens of branch offices, each with its own site on a shared template. It works the same way for any business with many locations and a reputation that lives in search.

GROWTH

Be the answer when someone asks.

More buyers now ask an AI assistant before they open a page of search results. Whether you show up in its answer depends on things a crawler can check: one clear heading, structured data, a plain-language FAQ, and an address and phone number that agree everywhere they appear. We audit those checks across every site and location you run, then draft the fixes from your own facts, ready for a person to approve.

Start the conversation

THE PROBLEM

Search changed. Most sites didn't.

From the outside, the gaps are hard to see. From an audit, they're obvious. Here's what we usually find:

Every location runs on the same website template, so the same mistakes repeat on every site: several H1s on one page, headings that skip levels, titles that get cut off in results.
There's no visible FAQ and no FAQPage or LocalBusiness schema, so an AI assistant has to guess what you do and where you do it, and it often guesses from someone else's page.
The address or phone on the website doesn't match the office record or the Google listing, or a closed duplicate listing still carries the live number. Callers and AI answers both end up in the wrong place.
Each office pays a different outside firm for 'SEO', and nobody at the center can say which firm, what it costs, or whether it's working.

WHAT WE DO

Audit first. Then fixes you can approve.

Scored site audits

A fixed set of checks run against each site's home, AI information, and contact pages, weighted into a 0–100 score: one H1, heading hierarchy, title and meta description length, canonical URL, structured data, and the city you serve in the title. Any site can be scored side by side against another of your sites, check by check.

Answer-engine readiness

Checks for what AI assistants actually quote: a visible FAQ with conversational answers, marked up as FAQPage schema, and an AI information page with one H1, a visible last-updated date, and facts that haven't gone stale.

Name, address, and phone consistency

The address and phone on every page compared against your office records and your Google listing, including the easy-to-miss case of a site publishing a sister office's address as its own.

Drafted fixes

FAQPage and LocalBusiness JSON-LD, rewritten titles and meta descriptions, an AI information page, and social posts, generated from the page itself and your office record. Each one is a draft a person edits and approves.

Listings and reviews

Duplicate Google listings, a 'permanently closed' listing still carrying your live number, and offices with no active listing, flagged per office. New reviews are triaged by rating, and negative reviews with feedback reach the local office with a recommended reply and a deadline.

Outside-firm visibility

A short survey of which outside firm each office uses, what it costs each month, and how the office rates it, rolled up centrally alongside the offices that have no professional help at all.

HOW WE ENGAGE

Four phases. The audit comes before any rewrite.

1–2 weeks
Audit
A scored audit of each site, every finding tied to the page and the text it came from, compared against your strongest site
1–2 weeks
Draft Fixes
- Structured data, FAQ, title and meta, and AI information page drafts for each site
- Listing problems and review backlog flagged per office
2–4 weeks
Approve & Hand Over
Your team edits and approves each draft; approved assets are frozen and handed to whoever runs your site
ongoing
Re-audit & Operate
Scheduled re-audits, review triage, and the latest score for every location in one place

TECHNOLOGY

The stack, plainly: the checks are code, not opinions.

An agent fetches pages and narrates what it found. It never grades. Every finding and the score come from deterministic, tested checks.

Agents & models: Claude on AWS Bedrock, driven through the Vercel AI SDK. Every page fetch and check is recorded as a step with its source, so a run can be watched live or replayed later.
Page analysis: Live fetches of public pages, parsed into headings, title, meta description, canonical URL, and JSON-LD, then scored by checks written and unit-tested as ordinary code.
Structured data: schema.org JSON-LD: FAQPage built from the questions already on the page, and LocalBusiness with full street address, postal code, geo coordinates, service area, and hours.
Listings & reviews: Your listing and review records checked against your own office records, with review tiers and listing rules written as explicit, testable code.

GOVERNANCE & RISK

Nothing publishes itself.

These fixes change your public face, so the controls are strict:

Every generated asset is a draft. A person approves it, and once approved it is frozen rather than quietly edited.
Generated content uses only facts from your own pages and office records. A rating appears in schema only when recorded listing data holds one.
Every check shows its evidence: the page, the quoted text, and why it passed, warned, or failed.
Review responses never promise a refund or an outcome, never admit fault, and never name staff or reviewers. Unsafe wording is rejected and the approved template is used instead.
Reviews alleging misconduct or threatening legal action are held for a person. No automatic response is posted.

QUESTIONS

What to ask before you hire anyone for SEO.

RELATED SERVICES

Being found is half of it.

Search gets you considered. Three services usually show up alongside this one:

GET IN TOUCH

Show us the site. We'll show you what the assistants see.