SOURCE

Local SEO in the AI Era: What’s Working, What’s Dying & What to Do Now (Joy Hawkins & Caleb Ulku)

Properties1
tagsseo, local-seo, geo-aeo, gbp, reviews, citations, awards, core-30, duplicate-content, joy-hawkins, caleb-ulku, edward-sturm
AUTHOR
Joy Hawkins & Caleb Ulku (guests); Edward Sturm (host)
PUBLISHED
2026-08-06
ORIGINAL
view original

7 min read

Local SEO in the AI Era: What’s Working, What’s Dying & What to Do Now

Guests: Joy Hawkins / Sterling Sky · Caleb Ulku · Host: Edward Sturm (NEW) | Published: 2026-08-06 | Source: transcript (72:24, 10.0K views)

Summary

Edward Sturm (episode 1,128 of his daily show) puts Joy Hawkins and Caleb Ulku on the same call for the first time. The value is comparative: they agree that local SEO is shifting from GBP-review volume toward LLM-citable proof, and they disagree on whether location pages should be unique.

On GEO / AEO (Getting Recommended by AI), Caleb says humans bounce before they read, so his agency historically optimized for trust; LLMs actually read the site, so he added a fifth content type written for citation (a Gary, IN plumber publishing water-heater pricing so Gemini/ChatGPT cite the local shop over Angie’s List). Joy’s equivalent move is awards: Best Law Firms and niche badge sites (Homestars, Consumer’s Choice) are surging while generic directories crash; a Canadian lawn-care client with a decade of badges “dominates” LLMs once the badges are placed prominently on-site. Winning is not enough — you have to blast it. She also tracks Lily Ray: local businesses that self-publish too many “we’re the best” listicles are getting nailed, which is why paying a third party to say it has become the workaround.

On Review & Reputation Management, they split the pack from the AI pack. Joy: volume + frequency still gate the classic 3-pack, but reviews have taken a hit in AI local packs (~10% of US results; not fully rolled out). Caleb: less GBP-only reviewing, more a diversified portfolio — Gemini “just freaking loves Yelp”; OpenAI has a Yelp data deal; coach clients to ask for experience + outcome, not a generic “great service.” Both treat Yelp/Reddit as scarce human-content reservoirs; Joy says Yelp’s fake-review detection and “badge of shame” are harsher than Google’s 30-day slap.

Tactics they both want dying: guest posts, scaled exact-match anchors (Joy: don’t chase more than ~5), 2017 spray-and-pray citations, CTR/direction-click manipulation (works, reverses when you stop, can suspend GBPs), and thin city-gate / service-area farms. Joy now calls service-area pages a liability (~25% of recent audit leads hit by a core update; Google labeled city-gate spam last summer) and says her test of making rewritten multi-location pages identical performed better — a live corroboration of the Duplicate Content (for Service-Area Pages) claim she already backed with James Dooley. Caleb still ships a unique Core 30 per location (Willis Tower vs. Wrigleyville). Joy goes further on same-city brands: one Chicago page listing all locations beat per-agent pages, with no local-pack penalty.

Press releases, a Joy staple a year ago, have “gone downhill” in the last six months (she thinks Google is hitting the PR networks). Keyword-stuffed GBP names still work in the classic pack after 20 years, but AI Overviews are surfacing firms that don’t do it. Low-hanging fruit they do agree on: fix the GBP landing page (list every service, match intent, no origin-story lead), verification-gated citations over MapQuest-style volume, local sponsorships / award listings over guest posts, short nav, keyword in the URL slug for pack rankings, and AI used as a researcher/interviewer rather than a spinner. 30-day actions: Joy — get an award and tell everyone; Caleb — get categories/services right, don’t keyword-stuff the name, put secondary categories on the landing page.

Key Claims

  • LLM-citable “fifth content type” (Caleb): pages that answer local how-much / how-it-works questions so models cite the business over national directories. (creator-opinion, mechanism plausible)
  • Awards are the new citation layer (Joy): niche award/UGC sites surge; generic directories crash; badges must be on-site for LLMs to find them. Best Law Firms ~60k clicks/mo and tripled prices. (practitioner-observation)
  • Reviews weaker in AI local packs than in the 3-pack (Joy); Yelp overweight for Gemini / AI packs (both; OpenAI–Yelp deal). Ask for specific outcome-reviews, diversify off GBP. (emerging)
  • Keyword-stuffed GBP names still work in the classic pack, may not in AI packs (Joy). AI local packs ~10% of US results. (observation, “no hard data”)
  • Duplicate/identical multi-location pages outperformed unique rewrites (Joy, from testing); service-area / city-gate pages are now a liability. Direct tension with Caleb’s unique-per-location Core 30.
  • Same-city multi-location: one city page listing all locations beat per-location pages, no local-pack harm (Joy, State Farm / Allstate-style test).
  • Press releases no longer seed LLMs the way they did ~2025; guest posts and scaled exact-match anchors are liabilities (Joy: ≤~5 exact-match; June core update hit a site doing only ~5 links/month).
  • GBP “diversity filter”: ranking in the pack and linking the GBP to the same URL that ranks organically suppresses that URL; swapping the GBP to a service page can lift homepage organic (Joy).
  • CTR / driving-directions manipulation works but reverses when stopped and can suspend GBPs; fake reviews also work until stripped. Both refuse to sell it.
  • Verification-hard citations > volume (Caleb, aligns with Super Citations); Joy: only list on sites that actually rank, MapQuest-style LLM citations will die.

Notable quotes

“If you’d asked me a couple years ago, [I] wouldn’t [have] cared about [awards] whatsoever. We are strongly pitching this to lots of our clients.” — Joy Hawkins

Why citable: Marks a real strategy shift from a practitioner who previously dismissed awards; pairs with her on-site “blast the badge” requirement.

“We did some testing around this where we actually made them the same… and they performed better because we kept the words the same.” — Joy Hawkins on multi-location pages

Why citable: Independent, on-air duplicate-content test — the same position she took with James Dooley, now stated in front of Caleb.

“I can’t remember the last time we posted human written content on a client’s website.” — Caleb Ulku

Why citable: States the production reality behind Core 30 / GEO / AEO (Getting Recommended by AI) content: AI-written, but grounded in Census/Reddit/GSC/client interviews, capped at ~10–15 pages/month.

Connections

Entities mentioned: Joy Hawkins, Sterling Sky, Caleb Ulku, Google, Google Business Profile, Claude, ChatGPT, Reddit · Edward Sturm (NEW), Gemini, Yelp, OpenAI, Best Law Firms, Homestars, Consumer’s Choice, Lily Ray, Noah Learner, Jake Hundley, Ian Evans / Agency Assassins, Ben Fischer, WordPress, Ahrefs (NEW as entity) Concepts referenced: Local SEO, GEO / AEO (Getting Recommended by AI), Entity-Based SEO, Core 30, GBP–Website Alignment (Consistency Signals), Duplicate Content (for Service-Area Pages), Review & Reputation Management, Super Citations, Local Link Building & Authority, Local Ranking Factors (Relevance, Proximity, Prominence/Authority), Brand & User Signals as Ranking Drivers, AI-Assisted Development

Contradictions / Tensions

  • Duplicate Content (for Service-Area Pages) vs Core 30 uniqueness (on-air, same episode): Joy = identical location copy won in tests; service-area farms are getting hit. Caleb = unique Core 30 per location, neighborhood-specific (Willis vs. Wrigleyville). Edward asked “how unique?”; they gave opposite answers. This is the wiki’s strongest live disagreement on the topic.
  • Same-city architecture: Joy collapses many locations onto one city page; Caleb would replicate the Core 30 per GBP.
  • Awards vs 250 Authority Protocol / Super Citations: Joy’s 30-day move is “get a badge and blast it.” Caleb’s 30-day move is GBP hygiene + landing-page services. Awards overlap Dooley’s “reviews/awards replace links” and Caleb’s third-party-validation bucket, but Joy is more award-centric and more skeptical of commodity LLM citations (MapQuest).
  • Reviews: Joy says review importance dropped in AI packs; Caleb still invests in review quality and platform diversity. Compatible if you split classic-pack vs AI-pack, but they weight the lever differently.
  • Press releases: Joy previously promoted them (including on this show); she now walks it back. Treat 2025 PR-seeding claims as stale.
  • Keyword stuffing the GBP name: Joy says it still works (20 years); Caleb’s closing advice is “don’t keyword stuff.” Tension inside “what’s working vs. what we recommend.”
  • vs. Nico (AI Ranking) / volume citations: both guests reject 2017 directory blasting; aligns with quality/verification over volume.
  • Commercial bias: both guests sell retainers; host sells Compact Keywords (mid-roll, ignore as evidence). Claims are practitioner-observation / testing-without-published-methods unless noted.

Notes

Highest-leverage ingest in this batch: two named authorities, already in the wiki, arguing in real time. Use it to move Duplicate Content (for Service-Area Pages) and GEO / AEO (Getting Recommended by AI) debates, and to record awards + Yelp-for-Gemini as independent (non-Caleb-branded) tactics. Do not average the uniqueness disagreement. Edward Sturm flagged NEW (host/channel only; not an SEO authority in this episode).