Define the exact user intent your content answers - think in terms of the questions someone might ask ChatGPT, Perplexity, or Google AI Overviews.

Avoid heavy industry jargon - use simple, real-world language that AI (and people) can understand without extra context.

Place your main keywords or phrases prominently in the title, opening paragraph, and headings so AI can quickly identify the topic.

Present core information in plain text, not embedded solely in images, PDFs, or JavaScript-rendered components.

Use a clear heading hierarchy (H1 → H2 → H3), bullet points, and numbered lists so AI can parse topic relationships.

Keep sentences under ~25 words for better AI parsing.

Break paragraphs every 2–4 sentences to avoid “text wall” effects.

Include a short, AI-readable summary of who you are, what you do, and who you serve - ideally in the intro or footer so it’s on every page.

State product/service differentiators explicitly (e.g., “Ally Bank offers…”) rather than implying them.

Avoid overly clever headlines without plain-language explanations - AI often takes copy literally unless context is provided on-page.

Context tab — Standardize personas, competitors, and key topics.

AXP (Agent Experience Platform) — Serve structured AI context invisibly to bots while keeping user-facing design intact.

Ensure the page renders all key text without JavaScript (use server-side rendering or pre-rendering for SPAs).

Allowlist AI user agents like:

Configure your firewall/CDN (e.g., Cloudflare, AWS WAF) to let these user agents through.

Provide alt text for images, transcripts for videos, and captions for diagrams.

Site Audit → Access Controls — Checks robots.txt/firewall rules.

AI User Agents Guide — Step-by-step allowlisting instructions.

Write title tags in under 60 characters and meta descriptions in under 160 characters, summarizing (not selling) the page in plain language.

Use short, readable URLs with relevant keywords (e.g., /best-online-banks instead of /article?id=1234).

Include schema markup if applicable, but ensure all key facts also exist in on-page text.

Site Audit → Metadata checks — Flags missing/overlong metadata.

Insights → Optimizer — Metadata playground environment.

Keep primary content under ~10,000 tokens (~7,500 words) to fit comfortably in most AI context windows without truncation.

Avoid “fluff” - focus on relevant, topic-specific detail that supports the main query.

Cover related subtopics so AI sees the page as authoritative, but avoid burying the main answer deep in the content.

Site Audit → Content Length check — Flags pages that are too short/too long.

Use the Citations tab in Scrunch as your primary tool for spotting high-authority third-party domains AI relies on for your target topics.

Apply platform and topic filters to see which external sites have the highest Influence Scores and appear most often in AI responses for the platforms that matter to you.

Target these domains for earned mentions - whether through outreach, partnerships, or contributing content - to boost your authority signals and increase the likelihood of AI citing your brand.

Citations tab → Influence Score + Consistency % — See which sites matter most for your topics.

Run the page through Scrunch Site Audit to check for:

Test retrieval in Scrunch Sources to see how models cite and summarize your page over time.

Monitor Presence and citation trends in Scrunch for 2–3 weeks post-launch; adjust structure, language, or metadata based on AI response patterns.

Ensure your brand appears in multiple independent sources.

Participate in industry forums, Reddit, review sites, comparison platforms.

Keep facts consistent across all listings.

Citations tab → Citation Consistency — See if AIs repeatedly reference your brand across different sources.

Make each 40–50 word block self-contained.

Lead with the answer under each H2, then explain.

Remove filler words and transitions.

Use bullets and tables for scannability.

Site Audit → Content Clarity — Flags buried/fragmented content.

AI Simulator (Scrunch Labs) — Test if answers are extracted cleanly.

Explicitly link your brand to industry leaders.

Add comparisons with competitors (neutral framing helps).

Seek inclusion in category roundups and tool lists.

Seed natural co-mentions on sites LLMs scrape.

Prompts tab → Competitor Presence filters — Track your mentions alongside theirs.

Citations tab — Identify roundup/list sites AI pulls from.

Include numbers, dates, metrics, specifications.

Cite primary sources and studies (go 1–2 layers deeper than competitors).

Update stats quarterly.

Use verifiable claims AI can cross-reference.

Citations tab → Domain frequency — Spot if AI is citing outdated competitor data over yours.

Insights tab — Flags metadata/content gaps (e.g., missing specs).

Use clean HTML (avoid over-complicated builders).

Maintain consistent formatting across page types.

Add explicit Q&A sections (helpful for Perplexity & Google AI Overviews).

Ensure schema complements text, not replaces it.

Site Audit → Content Clarity + Dynamic Content checks — Highlights hidden or mangled structures.

AI Preview (Audit) — See what bots actually parse.