AEO Website Strategy

AI Loves A Good Story.

A residential focused roadmap for creating loved-by-AI websites
in the new era of LLM search and citation.

SCROLL
0%

OF PROPERTIES ARE INVISIBLE IN AI SEARCH

0+%

OF AI RECS GO TO JUST THE TOP 5% OF COMMUNITIES

In traditional search you slip.
In AI search you fall off a cliff.

Why?

Because we spent the last 15 years building sites around reductive SEO constraints that rewarded singular, repetitive language with no regard for imagery, style or story.

That script just flipped and AI now wants a conversational, dialogue-based copy with unique photos and visuals to match – while turning its cheek to botspeak, stock and templates.

THE NEW LANDSCAPE

Search has fundamentally changed.
Most sites haven't adapted.

37%
of people search in an LLM not Google
30%
of renters in your markets now use AI to find a home
26
average words in an LLM search vs. 6 in Google
93%
of AI searches end without a single click

SEARCH EVOLUTION

Google Search — Then
|
6 avg. words per query
Returns 10 blue links. User clicks through to compare. Ranking depends on keywords and backlinks.
AI Search — Now
26 avg. words per prompt
Returns one synthesized answer. User gets a direct recommendation. Being cited requires structured data and authority signals.
0 clicks
view property → 93% of AI conversations end without a click. All of the info is neatly summarized and delivered — no searching around, no clicking on result links, no scrolling, no page 2. Which means being recommended is the conversion event. If your property isn't in the answer, it simply doesn't exist for that prospect.

THE CRITICAL INSIGHT

Answer engines are looking for more than answers.

Traditional search and SEO were based on mechanically narrowing and repeating. AI is looking for a human narrative, verbally and visually. It's reading — and looking at — your website and asking with human sensibilities and instant discernment:

Are you unique? Are you authentic? Are you quotable? Do you have taste? Do you have a perspective? Do you have a story to tell? Do you have something I can share with my client that will make me look good? Or are you just another forgettable, templated, bot-tailored commodity that's not just to be ignored but steered away from?

OPTIMIZING FOR HOW LLMS ACTUALLY THINK

AI wants to be your personal assistant, not a statistician.

An LLM revels in finding on-point info, tailored recs, unique finds and sought-after spots. It gets scolded when it delivers underrated, uninspired, same-as, simple ideas.

Its job is to research, curate, and recommend — to take a conversational question and come back with a confident answer. It was trained on how people actually talk, and to choreograph a response in kind.

That's a fundamentally different approach than what drove search engines and SEO.

Google rewarded keywords because Google was a librarian — efficient, indexed, scanning your page for terms that matched a short, specific query. So you seeded your site with "new luxury apartments in downtown LA now leasing" and trusted the algorithm to do the rest. AEO rewrites that contract. When an LLM lands on a page that says "luxury apartments now leasing," it sees a thousand other property sites saying the exact same thing — decides you have nothing distinctive to offer — and routes around you.

VERBAL AUTHORING SAMPLE

Story-led language, with a distinct point of view.

Imagine telling your PA: Can you find me a new apartment near DTLA – you know my style, high-end, refined, luxurious but niche and personal. If they came back with Park La Brea, you'd fire them.

What you actually want is the assistant who comes back with something like:

You |

AI
I found a new boutique residence in Silver Lake — close enough to DTLA but without the madness. Highly curated, intentional design, Japandi-inspired interiors. The kind of quiet luxury that feels like a refuge, not a scene. You'd be a short walk from Aesop, Sunspel, Erewhon, and some of your other regs. The community skews younger — creatives, low-key entertainment types — but the social spaces, organic retreats and outdoor amenities make for a calm scene. The architecture leans Scandi and reads more refuge than be-seen. It's called Mysa, there are some nice 1-beds available and I can set up a tour or pull rents if you like.

This is the ideal LLM response. And notice what made it into the recommendation. It wasn't a single SEO keyword. It wasn't repetition. It wasn't drawing from a bullet list of finishes. It was a vibe. A story and perspective that spoke to a narrowed taste, distinct design aesthetics, a neighborhood texture, a peek into fellow residents, local retail that aligns to known likes, and a sense of "I found your perfect match."


VISUAL AUTHORING SAMPLE

Original photography. True-to-life scenes.

SEO is word-centric. AI sees the whole picture. And it carries an originality bias that rewards unique photography and true context — while actively ignoring or punishing overused stock and generic scenes.

Stock

"Modern luxury kitchen with white countertops, stainless steel appliances and tile backsplash."

Original

"An East-facing kitchen at Mysa with morning light over a honed marble island. The courtyard's fig tree through the window. A vintage Danish table lamp adding a mid-century accent. Fresh groceries from Erewhon calling."

What AI ignores
  • Stock libraries. Same fingerprint as everyone else.
  • Generic luxury scenes. Rooftop pools, white kitchens, empty lobbies.
  • Scripted staging. Identical, too-perfect looks from the same design catalog.
  • Polish without proof. Beautiful images of anywhere.
What AI rewards
  • Specificity. Named places, real details.
  • Originality. Imagery, angles or scenes the AI hasn't seen before.
  • Cross-modal consistency. Photos that sync with brand copy that narrates the scene.
  • Proof of place. Real residents in real spaces.

EXECUTION

So, how do we get there?

THE STRATEGY

Three key pillars, three key questions.

1
Architecture
Can the AI read you?
The technical foundation. A site AI bots can actually crawl and read — no JavaScript walls, no content buried behind accordions or modals. Schema markup and structured data that tell AI exactly what your building is and offers. A lateral structure with real pages at real URLs, each one substantive enough to be worth citing. Without this, your building isn't on the LLM's initial reading list.
  • Crawlability for AI agents
  • Schema markup, structured data, technical SEO
  • Lateral navigation, non-interactive design
  • Purposeful pages with depth of content
  • No JavaScript, accordions, modals, hovers
2
Authoring
Can the AI quote you?
The brand layer. Good architecture gets you vetted, good branding gets you cited. This is where your story, voice and narrative come in — having a unique perspective, a design voice, an aesthetic point of view, a neighborhood texture — the language and descriptions that make your building feel like a specific place rather than a bullet list of features. This is what gets you cited and served.
  • Clear, unique story and perspective
  • Personal, conversational, anti-bot/bullet
  • Consistent tone and point of view
  • Design-led, neighborhood specific
  • Expressive, distinguishing details
3
Authority
Can the AI verify you?
The supporting layer. Third-party corroboration — GBP, press, interviews, neighborhood blogs, review surfaces, listing presence. Authoring is what gets you cited; authority is the supporting evidence that gives AI the confidence to do it.
  • Google Business Profile optimization
  • Third-party listing presence
  • Micro-editorial, Reddit, Substack, blog recs
  • Team press and interviews
  • Quality outbound citations from Journal to authoritative sources
  • Cross-source consistency

REFERENCE WIREFRAME

A complete site, page by page.

A best-practices blueprint applying every principle from this deck. Use it as a reference when scoping, designing, or auditing a multifamily site for the AEO era.

SITEMAP
HOME
Residences Floor Plans Amenities Neighborhood Journal Team FAQ Contact

Each top-level item resolves to its own substantive page at a real URL. Deeper pages live one level down where it makes sense (e.g. /floor-plans/the-larkin, /journal/[slug]).

01

Home

/

First impression and AI's first read. Establishes voice, perspective, and the primary entity schema for the building.

Page Structure

  • Hero: original full-bleed photography + narrative headline (no stock, no generic taglines)
  • Building summary — 2–3 paragraphs in established voice, surfacing what makes this place specific
  • Featured Interiors strip — 3 hero images of signature spaces, linking to /residences
  • Neighborhood teaser with named texture (named retail, culture, walkability)
  • Journal preview — 3 latest entries with thumbnails and real excerpts (no "Read more" walls)
  • Team intro snippet — faces and a sentence about who's behind the building
  • FAQ block — top 5 Q&A's fully expanded inline, not in accordions
  • Footer: full NAP, hours, social, sister-property links

Schema & Tech

  • Organization, ResidentialBuilding, Place at page level
  • FAQPage on the inline FAQ block
  • BreadcrumbList across the site
  • All content rendered server-side — no JS hydration walls
  • Hero photography with descriptive alt text + ImageObject schema
02

Residences

/residences

The lived experience. Interiors, finishes, materials, design — where prospects come to feel the place and AI infers a distinct aesthetic and identity.

Page Structure

  • Hero: design philosophy narrative — point of view, partners, what the look and feel means
  • Materials and finishes section — detail photography of specific surfaces, hardware, lighting, stone
  • Interior tour gallery — kitchens, bathrooms, living rooms, bedrooms in real-life context
  • Design partner credits — architect, interior designer, landscape architect, lighting consultant
  • Editorial on signature choices — why this hardware, this paint, this stone, this tile
  • Cross-links to Journal interviews with the architect, interior designer, and landscape team
  • Original photography only — no stock interiors, no staging-shop generics

Schema & Tech

  • ResidentialBuilding / Place at page level
  • ImageObject for each detail photo with descriptive captions
  • Person / Organization for design partners with sameAs links
  • CreativeWork for awards, design press mentions
  • All images rendered on load — no carousels, no click-to-reveal galleries
03

Floor Plans

/floor-plans

The unit-finding layer. Plan types and current availability live together — citable plan types on top, dynamic inventory inside each.

Page Structure

  • Index: every floor plan type listed laterally (typically 10–20 in a high-rise) — thumbnail, name, beds/baths/sqft, starting price
  • Each plan type = its own page at /floor-plans/[name]
  • Per-plan page: specs in HTML, lifestyle narrative, floor plan diagram (image, not interactive)
  • Available units listed inline on the plan page — Unit 1204 · $4,200 · SW corner · available 6/1
  • Variant notes by stack or orientation — corner units, view bands, floor groups
  • Unit inventory rows stay inline; no 200 individual unit pages
  • No PDF-only floor plans, no zoom-trapped modal galleries

Schema & Tech

  • Apartment / Accommodation per plan type
  • numberOfBedrooms, numberOfBathrooms, floorSize, leaseLength
  • Available unit inventory as a structured table with pricing and availability
  • Crawl-de-prioritize unit-level detail views in robots.txt if exposed
  • Plan-level pages become the citable entity, not the inventory rows
04

Amenities

/amenities

Show the specifics, not the list. Each amenity gets descriptive prose AI can quote, not a tag in a grid.

Page Structure

  • Hero: amenity philosophy in voice — why these and not those
  • Each amenity = its own section (not a card grid) with original photography
  • Descriptive copy: not "rooftop pool" but "South-facing rooftop pool with hand-poured concrete decks and skyline views"
  • Quotes from the team or residents where authentic
  • Cross-references to specific neighborhoods or building features

Schema & Tech

  • LocationFeatureSpecification for each amenity
  • amenityFeature array on the building-level schema
  • All amenity names and descriptions in HTML — no image-only cards
  • Alt text on every amenity photo, descriptive not generic
05

Neighborhood

/neighborhood

Texture, not a map. This is the page that paints AI a true understanding of the location and how it lives, not just geo data points.

Page Structure

  • Hero: neighborhood essay in voice — the character of the place
  • Named retail, restaurants, cultural touchstones — each with a mini story, not a logo
  • Walkability data in prose AND HTML (distances, walk times to named places)
  • Original photography of specific places (the corner café, the park, the bookstore)
  • Cultural context: who lives here, what they care about, what's nearby that matters
  • Linked Journal posts on neighborhood discoveries, retail finds, and local culture

Schema & Tech

  • Place / AdministrativeArea for the neighborhood itself
  • Each named business as a linked LocalBusiness entity
  • geo + distanceTo schema between residence and named entities
  • Map embed as static image with aria-label, link to Google Maps for interactive
06

Journal

/journal

The editorial heart. The building's own publication — interviews with the architect and interior designer, notes from the landscape team, nearby retail discoveries, the Tokyo vinyl bar that inspired the lounge, a sit-down with the developer on their vision. Each post adds named voices, narrative texture, and the kind of authoritative cross-references AI trusts.

Page Structure

  • Index: chronological + categorized listing with rich excerpts visible — no "Read more" walls
  • Each post = its own page at /journal/[slug]
  • Categories: Architect & designer interviews, Design diary, Landscape notes, Neighborhood discoveries, Inspiration & influences, Developer point of view, Resident stories
  • Outbound links to authoritative sources — publications, designers' portfolios, referenced places and works
  • Long-form (800+ words), original photography, named authors
  • Cross-linking to Residences, Neighborhood, Team, and Floor Plans pages
  • Monthly cadence minimum — signals an active editorial voice

Schema & Tech

  • BlogPosting / Article per post
  • author as Person with sameAs links
  • datePublished + dateModified
  • image as ImageObject with caption
  • Referenced people, places, and works marked up with sameAs to canonical sources
  • RSS feed available at /journal/feed.xml
07

Team

/team

Faces and verifiable bona fides. The "who's behind this" page that gives AI the trust signals to cite confidently.

Page Structure

  • Hero: collective intro — mission, point of view, what binds the team
  • Per-member: photo, name, role, bio paragraph, press links, social links
  • "In the press" / "Selected interviews" section with external links
  • Linked Journal interviews with team members, design partners, and the developer
  • Origin story or background on ownership / operations

Schema & Tech

  • Person schema per team member
  • sameAs links to LinkedIn, Twitter, Substack, personal sites
  • Press mentions as CreativeWork with author + publisher
  • Team member affiliation to building's Organization
08

FAQ

/faq

Direct, citable Q&A. The single most AI-friendly page format — and the one most often broken with accordions.

Page Structure

  • Categories: Leasing, Pets, Parking, Building, Neighborhood, Move-in
  • Every answer visible on load — no accordions, no tabs, no "click to expand"
  • Direct, conversational answers (not lawyer-speak)
  • 30–60 questions covering everything a prospect or resident actually asks
  • Anchored links (/faq#pets) so categories are linkable

Schema & Tech

  • FAQPage with Question / Answer pairs
  • Every Q&A pair in raw HTML — visible before JS executes
  • Question text in <h3> or <dt>, answer in <p> or <dd>
  • Anchor IDs on each question for direct linking
09

Contact

/contact

The verification layer. Where AI confirms the basics it'll quote — and where every fact must match GBP exactly.

Page Structure

  • Full address, phone, email, leasing hours — all in HTML, all visible
  • Embedded map (static image + link to Google Maps)
  • Tour booking link or simple form (server-side accessible, no required JS)
  • Direct lines to leasing, maintenance, press, and partnerships if relevant
  • Building access hours, package handling, after-hours protocols

Schema & Tech

  • LocalBusiness with PostalAddress, OpeningHours, telephone, email
  • GeoCoordinates matching map embed
  • NAP (Name / Address / Phone) identical here, on GBP, and across all listings
  • ContactPoint for each role (leasing, maintenance, press)