Service · Technical SEO

Make your site readable to crawlers and to AI.

Googlebot still matters. So does GPTBot, ClaudeBot, PerplexityBot, and the citation parsers downstream. We fix the crawl, render, and schema layer for both — and add the AI readability signals (llms.txt, structured citations, parser-friendly markup) most technical programs skip entirely.

Book a technical review →See the GEO program

One URL · four bots · two failuresDIAGNOSED

URL

/pricing

§01 · Crawl

200 OK

§02 · Render

JS-blocked

§03 · Parse

No schema

§04 · Surface

Not cited

Crawl OK ≠ CitedRender + Parse decide it

62%

of B2B SaaS sites we audit have JS-rendering issues

Last 30 audits

8/10

ship with at least one critical schema bug

First-pass crawls

0.41s

median LCP improvement, post-engagement

Field data

94%

of pages cited by AI also pass our render audit

Citation correlation

§01 — Where technical SEO breaks now

Three new failure modes
that didn’t exist in 2022.

§01.01

Your JS framework hides the page from AI parsers.

Most LLM crawlers don’t execute JavaScript the way Googlebot eventually does. Your hero copy, schema, and even the H1 can render perfectly in a browser — and look like an empty shell to GPTBot.

§01.02

Your schema is technically valid and semantically wrong.

Product schema on a comparison page. FAQ schema on a landing page with no FAQs. Article schema with no author entity. Validators pass; Google’s rich results don’t fire; AI parsers ignore the markup.

§01.03

Your site is crawled fine and never cited anyway.

Crawl budget is healthy, status codes are clean, sitemap submits — and you’re still missing from AI answers. The bottleneck isn’t crawl. It’s the citation-friendly markup AI looks for and you haven’t shipped.

§02 — The technical stack we run

Six layers,
audited and shipped.

Crawl

Crawl & index health

Crawl budget, status-code anomalies, redirect chains, soft-404s, robots.txt and meta-robots conflicts, sitemap hygiene, log-file analysis where the volume justifies it.

Render

JS rendering parity

Server-side render audit across the four bots that matter (Googlebot, GPTBot, ClaudeBot, PerplexityBot). Hydration timing, lazy-load conflicts, client-only critical content moved server-side.

Schema

Structured data, AI-aware

Organization, Product, Article, FAQ, Author, Breadcrumb. Entity consistency across pages. JSON-LD for AI parsers, microdata for rich results, OpenGraph for social.

AI readability layer

llms.txt setup with section-level granularity, citation-friendly markup (clear quotes, bylines, dates), source-page format that LLMs prefer to lift from. The piece nobody else is shipping yet.

Perf

Core Web Vitals, on real devices

Field LCP/INP/CLS, render-blocking diagnostics, edge-cache and image-format improvements. Tied to revenue impact, not synthetic Lighthouse scores.

Arch

Information architecture

Internal link graph, topic clusters, orphan pages, faceted-nav explosions, canonical and hreflang sanity, multi-language and multi-region site structure.

§03 — The bots that decide whether you’re cited

Five crawlers,
five quirks.

None of them behave like Googlebot used to. We tune for all five — because the ones you ignore decide what AI cites.

Googlebot

JS execution. Render queue. Long-standing playbook.

GPTBot

OpenAI’s crawler. No JS. Aggressive markup parsing.

ClaudeBot

Anthropic’s crawler. Respects robots. Markup-first.

PerplexityBot

Citation parser. Reads structured quotes and bylines.

Google-Extended

AI Overviews + Gemini training. Separate UA.

§04 — Why the 2022 technical playbook leaks

Technical SEO was built for one bot.
Now there are five.

2022 technical SEO

Geology technical SEO + GEO

Goal

Crawled and indexed

Crawled, parsed, and cited

Render

Optional — Google handles it

Mandatory SSR for the AI bots that don’t

Schema

Validator-passing

AI-parser-friendly + entity-consistent

llms.txt

Doesn’t exist in the playbook

First-class deliverable

CWV

Lighthouse green

Field LCP, INP, CLS tied to revenue

Audit cadence

Annual

Quarterly with bot-traffic monitoring

§05 — Common questions

What technical leads
actually ask.

Our site is on Next.js / React / Vue. Are we already fine?

Maybe — depends on whether you’re actually rendering server-side and how. Next.js with SSG or SSR, configured properly, is usually safe. App Router with default RSC is solid for Googlebot and decent for GPTBot. SPA-style React with client-only data fetching is almost always a problem for the non-Google bots. We test it directly rather than guess.

Do we really need llms.txt?

It’s not yet a hard ranking factor and won’t be the difference on its own. But it’s a low-cost signal that meaningfully helps AI crawlers index the right sections of your site, and we’ve seen citation lift correlated with shipping it well. We treat it as cheap, low-risk, and on the standard stack.

Can you fix Core Web Vitals on a JS-heavy site?

Yes — and it’s usually less invasive than the dev team fears. Most of our CWV wins come from edge caching, image-format swaps, render-path cleanup, and decoupling above-the-fold critical content from the framework hydration cycle. Full SSR migrations are sometimes the right call but they’re the exception.

How does this overlap with the GEO Optimization service?

Technical SEO is the technical layer of GEO. If you’re engaging on GEO Optimization, the technical work is included. This page is for clients who want only the technical scope — typically because their content and off-site work is already in good hands and they want a focused engagement on the crawl/render/schema/AI-readability layer.

What does the deliverable look like?

Same format as our SEO Audit deliverable: a prioritized fix list by week, by owner, by expected impact. Pre-written tickets for your tracker. A 90-minute hand-off call. We stay on Slack for 30 days for blockers.

Will fixing technical SEO move pipeline on its own?

Sometimes. More often it removes a ceiling. Most sites we audit have content and brand that should be ranking and being cited far more than they are — and the technical layer is what’s holding it back. Fix the technical layer and the existing content compounds.

Get started

Crawl OK isn't cited.

Book a 30-minute technical review. We'll run your URLs through the four AI bots that matter, surface the render and schema gaps, and tell you what we'd ship in the first 30 days.