How to Run a GEO Audit for a Prospective Client in Under an Hour (With a Free Pitch Deck Template)

Key takeaways

• A credible GEO audit does not require a week of research. You can produce a defensible, data-backed snapshot in under an hour by timeboxing five checks: crawler access, citation baseline, chunk/content readiness, entity consistency, and a competitor comparison.
• The audit's real job is not to be thorough. It's to be specific enough that the prospect recognizes their own gaps and wants the full engagement.
• "Money prompts" (10-15 buyer-style questions run across ChatGPT, Perplexity, and Google AI Overviews) replace keywords as your core measurement unit.
• Pitch decks built from live audit data close better than generic capability decks. Lead with the prospect's own visibility gaps, not your agency's service menu.
• Tools like Promptwatch can compress the baseline and competitor steps from 20 minutes to about 5, which is where most of your hour gets saved.

Why the one-hour constraint actually matters

I've sat through agency sales calls where the "AI audit" took three weeks to produce and arrived after the prospect had already signed with someone else. That's the real risk here, not inaccuracy. Speed is the pitch. If you can show a CMO their brand is invisible in ChatGPT for the exact questions their buyers ask, in the same meeting where they raised the topic, you've won most of the deal before you've priced anything.

The constraint also forces discipline. A full GEO audit, the kind Winston Digital Marketing lays out in its seven-step methodology, involves a real crawl, passage-level chunk scoring, schema review, entity checks, a citation map, and a competitive delta. That's a paid engagement, not a pitch asset. Your one-hour version is a compressed, defensible preview of that same methodology, not a watered-down guess.

The five checks that fit in 60 minutes

Minute 0-10: crawler access and robots.txt

Start here because it's a pass/fail gate. If a prospect's site blocks GPTBot, Google-Extended, PerplexityBot, or ClaudeBot, nothing else you find matters until that's fixed. Demand Local's agency walkthrough flags this as one of the most common and most embarrassing findings: sites that blocked AI crawlers in 2023-2024 to avoid being used as training data, and are now simply absent from AI answers as a result.

Check both the robots.txt file and actual server behavior. Don't just read the file. CDNs, WAFs, and bot-management tools sometimes reject AI user-agents even when robots.txt looks open, so a quick manual user-agent test (curl with a spoofed GPTBot header, or a free checker like Rank Prompt's robots.txt tool) catches cases the file alone would miss.

Minute 10-30: the money-prompt baseline

This is the core of the audit. Write 10-15 "money prompts": the actual conversational questions a buyer would type into ChatGPT or Perplexity when shopping in this category. Not keywords. Questions. "What's the best [category] for [use case]" or "Is [competitor] better than [prospect]" type phrasing.

Run those prompts across ChatGPT, Perplexity, and Google AI Overviews, and note three things for each response: whether the prospect's brand appears at all, whether they're cited with a link, and who is cited instead. This is where a dedicated tracking tool saves the most time. Doing this manually across three engines and 15 prompts is 45 minutes of copy-pasting. A platform like Promptwatch runs the same baseline in a few minutes and gives you citation share, prompt volume, and difficulty scoring you can screenshot directly into the deck.

Promptwatch

Track and optimize your brand's visibility in AI search engines

View more

Keep the prompt phrasing tight. Promptwatch's own data on query fan-outs shows ChatGPT's average query length has dropped from roughly 117 characters to about 53 over the past several months, meaning the model is searching more like a keyword engine now than a conversational one. Write your money prompts the way a person actually talks, but don't be surprised if the retrieval behind the scenes is shorter and blunter than you'd expect (ChatGPT Query Fanouts).

Minute 30-40: content readiness spot check

You don't have time to chunk-score the whole site, so pick the three pages most likely to be cited: the homepage, the top comparison or pricing page, and whatever ranks for the category's main keyword in Google. For each, ask whether the first sentence of each section actually answers the heading above it, without relying on "as mentioned above" or a pronoun that only makes sense in context. Winston Digital calls this chunk citability, and it's the single fastest proxy for whether a passage can be lifted cleanly into an AI answer.

Also glance at structured data. FAQ, Product, and HowTo schema correlate with higher citation odds, though that's a vendor claim worth treating as directional rather than proven. What's not in dispute is Promptwatch's citation-type data for August 2026, showing how-to content nearly doubled its share of ChatGPT citations (4.3% to 9.1%) and documentation nearly tripled (3.3% to 8.2%) within the month, while product pages still led overall at around 29% (ChatGPT Citation Types Over Time, August 2026). If the prospect has zero how-to or documentation content, that's a genuine, fixable gap worth naming out loud in the pitch.

Minute 40-50: entity consistency check

Open the prospect's website, LinkedIn page, and Google Business Profile side by side. Do the descriptions match? Does the site call them a "platform" while LinkedIn calls them an "agency"? AI models lean on consistent entity signals to decide what a brand actually does, and conflicting descriptions push them toward citing a competitor with cleaner data instead. This check takes ten minutes and produces one of the most quotable lines in the pitch: "Google calls you X, your own homepage calls you Y."

Minute 50-60: competitor snapshot

Run two or three of your money prompts again, this time scoring which competitors get cited instead of the prospect. You don't need a full share-of-voice report here, just enough to show a side-by-side: "For 'best [category] for small teams,' ChatGPT cited these three competitors and not you." That single slide usually does more work than any slide about your agency's methodology.

What this looks like as a repeatable checklist

Run it in this order. The crawler check can end the audit early (in a good way) if it surfaces a blocking issue you can fix before the pitch meeting even happens.

Picking a tool for the baseline step

The money-prompt baseline is where speed actually gets won or lost, and the tool landscape here is crowded. A few practical notes from comparing entry pricing and feature depth across the category.

Promptwatch

Track and optimize your brand's visibility in AI search engines

View more

If you're auditing one-off prospects before they've signed anything, the free tiers on Promptwatch, Otterly.AI, or AthenaHQ cover the baseline step fine. Once you're running this audit as standard prospecting across a client base, the Business tier on Promptwatch (5 projects, 42,000 responses a month) is built for exactly that volume.

For technical crawl checks on larger sites, Screaming Frog still handles anything under 10,000 URLs quickly; for JS-heavy sites you'll need something that renders pages, where Firecrawl or a DataForSEO On-Page API call does the job faster than a desktop crawler.

Screaming Frog

Industry-leading website crawler for technical SEO audits

View more

Turning the audit into a pitch deck

The audit is the content. The deck is just the container. Structure it in this order, and keep it to roughly 10-12 slides:

1. One sentence framing the shift. Not a market-size slide. Something concrete: "For [category] questions, ChatGPT cites 5 sources per response versus Google's roughly 10, which makes each citation slot worth fighting for" (Average Sources Per Response).
2. The prospect's own invisibility, shown live. Screenshot the actual ChatGPT or Perplexity response where their brand is absent and a named competitor is cited. This is the slide people lean forward for.
3. Why it's happening. One slide per finding: crawler block, thin content readiness, entity inconsistency. Keep each to a single specific example, not a general statement.
4. Who's winning instead, and why. The competitor snapshot. Resist the urge to pad this with more than 2-3 competitors.
5. What changes, and what it's worth. Translate citation counts into the business language the prospect actually cares about: pipeline, not citation volume.
6. The engagement, in tiers. Offer an entry option (a focused audit expansion, commonly priced $1,500-$3,000 for a 10-prompt deep dive) alongside a fuller multi-platform engagement ($5,000-$7,500), so the prospect has a low-friction way to say yes without committing to the largest package first.
7. Cadence and proof. Note that GEO visibility shifts month to month, not quarter to quarter, which is why this isn't a one-time deliverable. Citation share data changes often enough that a single snapshot goes stale within weeks.

A few structural notes that apply to any agency pitch deck, not just GEO ones: tie every claim to a metric the client already tracks (leads, demos booked, not just "visibility score"), and name the risks you've already mitigated rather than waiting for the prospect to raise them. Both points show up consistently in agency pitch-deck guidance from Awesomic and general advice on structuring client-facing decks.

Common mistakes that undercut the pitch

A few patterns show up often enough to call out directly.

Treating the audit as a one-time snapshot instead of flagging it as ongoing monitoring. Citation rates genuinely move month to month, and if you present a single static number without acknowledging that, a sharp prospect will ask why they'd pay you again in Q2.

Optimizing the pitch around blog content alone. Service pages, comparison pages, and case studies are frequently stronger citation assets than blog posts, and ignoring them in the audit signals a shallow understanding of how these engines actually retrieve content.

Over-indexing on Reddit as a GEO channel. Promptwatch's data shows Reddit's share of ChatGPT Search citations collapsed from roughly 3.8% to just 0.5% in a single day in mid-August 2026, an 86% relative drop, while Google AI Overviews and AI Mode showed a much more gradual decline over the same window (Reddit Citations Are Dropping in ChatGPT). If a prospect's existing GEO strategy leans heavily on UGC and forum seeding, that's worth raising as a risk, not a strength.

Assuming domain authority is the deciding factor. It isn't, at least not the way most SEO teams assume. In August 2026, domains in the DR 46-75 range captured almost half of all ChatGPT citations, while the very largest sites (DR 91-100) fell to around 3% (ChatGPT Citation Share by Domain Rank, August 2026). If the prospect assumes they're locked out because a competitor has a bigger domain, that assumption is worth correcting in the pitch, because it changes what you'd actually recommend fixing first.

Making this repeatable across your client base

Once the audit template works for one prospect, the real value is running it proactively against your existing SEO clients before they ask about GEO at all. If you already manage SEO retainers for a roster of accounts, build this into your quarterly business review process rather than waiting for a client to bring up ChatGPT in a meeting. Expansion revenue from existing accounts tends to carry a far better return than new client acquisition, simply because the relationship and billing infrastructure already exist.

Agencies that need to scale this across dozens of accounts usually end up on a tool with white-label reporting and client-workspace features rather than running everything manually. If you want to see how the broader GEO and AI-visibility tool category compares beyond the handful mentioned here, the directory at bestgeosoftware.com breaks down pricing and feature sets across a much wider set of platforms. And if GEO isn't a fit for a given prospect but they're clearly behind on technical SEO fundamentals, 1001 SEO Media runs full audits and ongoing technical, content, and digital PR programs that complement exactly the gaps this hour-long audit surfaces.

The audit itself doesn't need to be perfect. It needs to be specific, fast, and tied to something the prospect can verify themselves in the room. That's what gets the signature, not the polish on the deck.