The AEO Agency
Playbook
Everything you need to build, price, and sell AEO/GEO services: positioning templates, market pricing benchmarks, a scope-of-work framework, and a 30-60-90 day delivery roadmap. Grounded in 2026 market data.
Agency Client AI Search Auditor
Pitching AEO to a client starts with proving their current invisibility. Input a client's URL below to run a real-time synthetic simulation across ChatGPT, Perplexity, and Claude, extracting crawler readiness metrics.
Five things every agency owner should know before selling AEO
AI search is now the first touch between the buyer and your client's brand
AI assistants generated an estimated 45 billion sessions worldwide, roughly 56% of global search-engine volume [4], and ChatGPT alone handles ~2.5 billion prompts a day [5]. Which means the first impression of your client's brand is now routinely formed inside an answer that nobody at the company has ever read.
Analytics see everything.
zero click. Impression formed, no visit logged
the rest click pre-sold, and convert
The zero-click case
Most first touches now end where they start: inside the answer. When an AI Overview was shown in a randomized field experiment, zero-click searches jumped from 54% to 72% of queries [8]; Pew found only 1% of visits clicked a link inside an AI summary, and 26% of users ended their session entirely after reading one [8]. In Google's AI Mode, 75% of sessions end without any external visit [1]. The buyer got their answer; the client's site simply never saw the session. Once that's true, session counts stop measuring influence. What you can still measure is whether the brand shows up in the answer. Citation presence is the first-touch metric now, and showing up pays: 85% of B2B buyers think more highly of vendors an AI names in its answer [1].
Organic clicks are being repriced
Independent studies disagree on magnitude but not direction: when Google shows an AI Overview, clicks to websites fall.
One nuance worth using in a pitch: brands cited inside the AI Overview earn roughly 35% more organic clicks on those same queries [7]. The clicks didn't vanish. They moved to whoever the answer names.
Small traffic, outsized revenue
A visitor who arrives from an AI answer has already compared their options and read a recommendation. Most of the selling happened before the click, and it shows in the conversion data:
| Study | Segment | AI-referred vs. organic | Multiple |
|---|---|---|---|
| Opollo, 2026 [6] | 312 B2B tech firms | 14.2% vs. 2.8% | ~5× |
| Seer Interactive [6] | ChatGPT referrals, B2B | 15.9% vs. 1.76% | ~9× |
| Ahrefs (own site) [6] | B2B SaaS signups | 0.5% of traffic → 12.1% of signups | 23× |
| Semrush, 2026 [6] | Cross-industry | blended average | 4.4× |
| Adobe Analytics, Q1 2026 [10] | US retail, 1T+ visits | +42% conversion lift vs. non-AI | 1.4× |
For B2B tech clients specifically, Opollo tracked AI referrals growing from under 1% of traffic in January 2025 to an average of 6.4% by January 2026, up 975% year over year [6]. The channel is small but compounding, and hardly anyone reports on it; only about 14% of marketers track AI search as its own channel [6]. That blind spot is where an agency earns its keep.
Positioning templates: pick one, commit
A client asking about AEO in 2026 usually isn't shopping on price. They have a problem they don't fully understand, it feels urgent, and they want someone who obviously does understand it [11]. That impression starts with a sharp position. Here are three that work, with fill-in-the-blank language:
A · The SEO Expansion
Upsell existing retainers"We protect and grow [client]'s organic revenue across both Google and AI answers: your existing [SEO program], plus tracking and fixing how ChatGPT, Perplexity, and Gemini recommend you against [competitor]."
Lead with: the client's own declining CTR despite stable rankings. Watch out: priced as a small add-on, it anchors AEO as "extra SEO". Bundle it as a program upgrade, not a line item.
B · The AEO Specialist
Win new logos"We get [ICP, e.g. B2B SaaS companies] cited by the AI engines their buyers trust. We audit your AI visibility across six platforms, fix the signals keeping you invisible, and track your share of voice against [named competitors] every week."
Lead with: a live audit of their brand; most B2B SaaS brands are invisible in 88–96% of relevant prompts [12]. Watch out: specialists get commoditized by dashboards; your moat is the fixing, not the monitoring.
C · The Revenue-Attribution Leader
Highest retainers"We turn AI answers into attributable pipeline for [client]. Every citation is tracked from prompt to site visit to signup to closed-won, so your [CMO/CFO] sees exactly what AI visibility is worth in revenue."
Lead with: the conversion-multiple data (Part 01) plus their own GA4 AI-referral numbers. Watch out: this position requires real attribution tooling. Promising it without the stack to prove it is how agencies lose accounts (see Part 03).
Selling AEO
How to frame AI search for a client who has never heard of it, and how to close. Most prospects have never met the acronym. What they do know is that their organic traffic is falling and a competitor keeps coming up "everywhere." Start from what they already feel, not from what you want to teach them.
- Lead with jargon. "GEO," "LLMO," "RAG," and "embeddings" mean nothing to a CMO.
- Sell citations as the outcome. A citation is a vanity metric until it's tied to pipeline.
- Pitch it as a small SEO add-on; that anchors the price and hides the new work.
- Guarantee placement in AI answers. Nobody controls the models, and clients remember the promise long after you regret making it.
- Lead with buyer behavior: "Your customers already ask ChatGPT what to buy. Are you in the answer?"
- Use the analogy: AI search is word-of-mouth at scale.
- Bridge from SEO: "SEO got you on page one. This gets you into the answer."
- Anchor on revenue math: at their deal size, one AI-sourced deal usually pays the retainer.
The single most effective sales tool in AI search is showing, not telling. On the call, share your screen on a live AI-visibility dashboard and walk through category prompts, share of voice, and citation gaps.
See a sample dashboard with demo data →Discovery questions that open the deal
- When a buyer asks ChatGPT "best [category] for [ICP]", do you know if you're named? Is your competitor?
- Has your organic traffic declined over the last 12 months while your rankings held?
- What share of your current traffic is AI-referred, and how does it convert vs. organic?
- Is GPTBot allowed in your robots.txt? Do you have an llms.txt?
- When did you last check how AI describes your brand: enterprise option, budget pick, or afterthought?
- Where does your entity data live (LinkedIn, Crunchbase, G2, Wikipedia), and does it agree with your website?
- Can your team deploy schema without an engineering sprint?
- If AI visibility doubled your highest-intent traffic next quarter, who would need to see that number to fund it?
The three objections — and the answers
"AI traffic is less than 1% of our sessions."
True, and that 1% converts at 4–23× organic and is growing triple digits. Ahrefs' 0.5% of AI traffic drove 12.1% of signups [6]. You buy channels before they're expensive, not after.
"Isn't this just SEO with a new name?"
Ranking and citation are different systems: pages ranking #1 often earn zero AI citations, brands are 6.5× more likely to be cited via third-party sources than their own domain [5], and the work itself (schema engineering, entity consistency, prompt testing, consensus building) is different work.
"How will we measure ROI?"
Three layers: citation rate on your buyers' prompts (leading), AI-referred sessions and conversion rate in GA4 (current), and CRM-tagged AI-origin pipeline (lagging). Defined in the contract on day one.
Change the report, keep the client
How to price it
Across published 2026 pricing research, credible AEO/GEO engagements cluster into consistent bands [9][11][13]:
| Engagement | 2026 market range | Notes |
|---|---|---|
| One-time AI visibility audit | $1,500 – $5,000 | The proven tripwire into a retainer |
| Foundations retainer | $2,500 – $3,500 / mo | 1 brand, core engines, schema + answer-first fixes |
| Growth retainer | $4,500 – $7,500 / mo | + content production, digital PR, weekly tracking |
| Authority / enterprise | $8,000 – $15,000+ / mo | Heavy off-site authority; enterprise scopes start ~$15k |
| Implementation project | $5,000 – $15,000 | Schema deployment, content restructuring; enterprise overhauls to $100k [13] |
| Hourly consulting | $100 – $250 / hr | Strategy sessions, audits, team training |
| Tooling (pass-through) | $200 – $500 / mo | State up front whether it's included — hidden tool costs are a trust issue [11] |
The four pricing models that work
The dominant model: citation gains compound and AI models keep changing, so the work never ends. Scope monthly deliverables clearly.
A $1.5–5k paid audit de-risks the decision for clients who want proof first, and converts naturally into the retainer that fixes what it found.
Fold AEO into an existing SEO retainer at a higher blended rate. Fastest path for agencies with a book of business.
Base retainer + bonus tied to citation rate or AI-sourced pipeline. Only credible if you can actually attribute revenue, otherwise it invites disputes [11].
- Never price below ~$1,500/mo; the market reads it as rebranded SEO [9].
- Never price by deliverable count (pages optimized, schemas shipped), which invites scope arguments. Price the outcome: visibility and attributed pipeline [11].
- Charge a premium over your SEO rate: the work (schema engineering, multi-engine prompt testing, entity consistency, off-site consensus building) sits outside a legacy SEO skill set [14].
- Define success in the contract before work starts — citation rate, share of voice, AI-referred conversions.
How to pitch with a F.A.C.T. diagnosis
The live demo opens the deal; the diagnosis closes it. F.A.C.T. gives your pitch a repeatable answer to the client's first question, "why doesn't AI recommend us?", by breaking AI invisibility into four diagnosable layers. Run them in order, because each depends on the one before it [16].
Findable: can AI crawlers reach the site at all?
The most common failure is the cheapest to fix: AI bots blocked or starved before they ever read a word.
Check: GPTBot, PerplexityBot, ClaudeBot, Google-Extended in robots.txt · llms.txt present · XML sitemap current · no bot-hostile CDN/WAF rules · crawlable URL structure.
Agent Accessible: can a machine parse what it finds?
Agents don't execute your JavaScript app shell and don't guess at ambiguous entities. Content that only exists after client-side rendering is invisible to most retrieval pipelines.
Check: server-side / pre-rendered HTML for key pages · Organization, Product, FAQPage schema · entity consistency — the same company name, description, and category on the site, LinkedIn, Crunchbase, and G2 · clean pre-click metadata.
Citable: is there anything worth quoting?
Engines cite fresh, extractable, answer-shaped content: 44.2% of LLM citations pull from the first 30% of a page [15], and pages updated within two months earn measurably more citations.
Check: 40–60-word answer blocks under each H2 · answers in the top third of the page · comparison tables and source-attributed statistics · original data the brand owns · update recency on money pages.
Trustable: does the rest of the web agree?
AI engines build answers from consensus. Brands are 6.5× more likely to be cited via third-party sources than their own domain [5]. A perfect website with no external validation still loses.
Check: presence on review sites (G2, Capterra), Wikipedia/Wikidata, Reddit and community threads · inclusion in the listicles LLMs already cite for the category · analyst and press mentions · consistent sentiment across sources.
Score each layer 0–25 in your audit and you have a defensible 100-point diagnosis, plus a built-in narrative for the SOW: F and A failures are projects (fast, technical, high-visibility wins); C and T failures are retainers (ongoing content and authority work).
Scope of work, by tier
Structure the SOW around the four-stage loop, Audit → Fix → Track → Attribute, so every deliverable maps to a stage the client can understand, and every tier is an obvious upgrade of the one below.
| Deliverable | Foundations $2.5–3.5k | Growth $4.5–7.5k | Authority $8–15k |
|---|---|---|---|
| Audit | |||
| AI crawlability & Client-Side Ghosting audit (robots.txt, GPTBot access, llms.txt, sitemap, raw HTML verification) | ✓ | ✓ | ✓ |
| Structured-data & schema gap analysis | ✓ | ✓ | ✓ |
| Baseline citation rate across 30–50 buyer prompts | ✓ | ✓ | ✓ |
| Entity consistency audit (site, LinkedIn, Crunchbase, G2, directories) | — | ✓ | ✓ |
| Fix | |||
| Pre-rendering / SSR HTML & Schema deployment (eliminate Client-Side Ghosting for React/SPA apps) | top 10 pages | sitewide | sitewide |
| Answer-first restructuring (answers in top 30%) [15] | 10 pages/mo | 20 pages/mo | 40 pages/mo |
| Net-new answer-ready content (comparison pages, FAQs, research assets) | — | 4–6 /mo | 8–12 /mo |
| Off-site authority & digital PR (~85% from third-party) [9] | — | 2 placements/mo | 4+ /mo |
| Track | |||
| Prompt tracking cadence across major engines | monthly | weekly | weekly–daily |
| AI share of voice vs. named competitors | ✓ | ✓ | ✓ |
| Prompt visibility mapped across TOFU / MOFU / BOFU | — | ✓ | ✓ |
| Attribute | |||
| GA4 AI channel group + AI-referred conversion reporting | ✓ | ✓ | ✓ |
| CRM tagging of AI-origin leads (citation → signup → closed-won) | — | ✓ | ✓ |
| Quarterly business review with pipeline-level ROI | — | — | ✓ |
Volumes are starting points; scale them to the client's page count and competitive density, not the other way around.
The 30-60-90 day delivery roadmap
Set expectations in the contract: with the technical quick wins shipped early, first new AI citations can show up within 2 weeks of starting; full ROI takes three to six months depending on existing authority [13]. The roadmap below gives the client something visible every 30 days.
- Kickoff: define ICP, name 3–5 competitors, agree success metrics in writing.
- Build the prompt library — 30–50 real buyer queries across TOFU/MOFU/BOFU; run the baseline across all six engines and record citation rate + share of voice.
- Technical audit: bot access (GPTBot, PerplexityBot, ClaudeBot), llms.txt, sitemap, schema coverage, entity consistency.
- Ship the quick wins: unblock crawlers, publish llms.txt, deploy schema on the top 10 pages, set up the GA4 AI channel group.
Client sees: audit report with visibility score, competitor benchmark, and a prioritized fix roadmap.
- Restructure the top 20 revenue pages answer-first: 40–60-word answer blocks under each H2, comparison tables, source-attributed statistics.
- Extend schema sitewide; fix entity inconsistencies across LinkedIn, Crunchbase, G2, and key directories.
- Publish the first 4–6 answer-ready assets (comparison pages, FAQ hubs, original data).
- Open the off-site track: pitch 2–3 third-party placements on pages LLMs already cite for the category.
Client sees: fix log, before/after page structure, and first movement in tracked prompts.
- Scale off-site authority: listicle and review-site placements, analyst mentions: the third-party consensus AI engines reward.
- Move tracking to weekly cadence; iterate on prompts where competitors still win.
- Turn on attribution reporting: AI-referred sessions, conversion rate vs. organic, CRM-tagged AI-origin pipeline.
- Run the 90-day review and present the expansion path (more engines, more prompts, more markets).
Client sees: the 90-day impact report — citation rate vs. baseline, share of voice trend, and AI-sourced pipeline in dollars.
Standard monthly scope (Growth tier)
Once the 90-day setup lands, the retainer settles into a steady monthly rhythm. Paste this into your SOW as the recurring scope. It keeps deliverables capped while still covering every stage of the loop.
- 30–5,000+ prompts tracked weekly across the buyer's target markets, engines, and languages
- AI share of voice vs. named competitors, with movement alerts
- Cited vs. not-cited source breakdown per platform
- AI-referred sessions, conversions, and pipeline in the monthly report
- 5–10 commercial page updates (pricing, product, services first)
- 4–8 new answer-ready assets targeting zero-visibility prompts
- Schema expansion + content freshness passes on money pages
- Internal links routing supporting content to commercial pages
- Third-party citation outreach: review sites, industry listicles, analyst mentions
- Community presence where engines listen (Reddit, Quora, niche forums)
- Entity upkeep: Wikipedia/Wikidata, Crunchbase, G2, knowledge-graph consistency
- Sponsored placements in publications the engines already cite
- Paid inclusion in credible category directories
- Promotion that speeds up earned pickup for new assets
The three media columns are the POEM model in delivery form; Part 09 explains how to sell that framing to the client.
Shaping market position with the POEM model
AI engines don't take a brand's word for its market position; they synthesize it from every channel at once. That makes the classic POEM lens (Paid, Owned, Earned media) the cleanest way to show a client how an agency moves the answer from "afterthought" to "recommended option." Each channel trades reach against control, and AEO changes what each one is for.
where AI consensus lives
the accelerant
your site & content
As control rises, reach and AI-citation weight fall. ~85% of AI citations come from third-party sources [9].
| Channel | What the agency does for AI visibility | What it moves in the answer |
|---|---|---|
| Owned | Answer-first pages, schema, llms.txt, comparison/FAQ hubs, original research the brand publishes: the F, A, and C layers of the F.A.C.T. diagnosis. | Whether AI can describe the brand accurately, in its own words: the positioning language engines quote. |
| Earned | Digital PR, review-site presence (G2, Capterra), analyst and press mentions, community threads, placement in the listicles LLMs already cite. This is the T layer and the biggest lever: engines weight Wikipedia, Reddit, and review sites heavily [5]. | Whether AI recommends the brand at all, and against whom: category membership and share of voice. |
| Paid | Sponsored placements and advertorials in publications AI engines crawl, paid inclusion in credible directories, promotion that accelerates earned pickup. Paid doesn't buy citations directly; it buys the surfaces earned media grows on. | Speed: how fast the consensus footprint builds versus waiting for organic pickup. |
The pitch this enables: "Your competitors own the earned column — that's why AI calls them the category leader and you the alternative. Here is the owned work we control this quarter, the earned placements we'll win next quarter, and the paid spend that shortens the gap." Positioning stops being a brand-deck abstraction and becomes a channel-by-channel work plan with a budget attached.
The stack: run the whole loop on one platform
Most AEO tools stop at monitoring: a dashboard that tells your client they're invisible, then leaves the fixing, tracking, and proving to you. The margin in this business is in running all four stages without stitching five tools together.
ThriveStack citedby is built for exactly that: it audits a brand against 73+ structured signals across ChatGPT, Perplexity, Gemini, Claude, Copilot, and Grok; generates copy-paste schema files and answer-ready briefs your team can deploy without a developer [16]; tracks prompts, share of voice, and buyer-stage visibility weekly; and, uniquely in the category, attributes AI citations through to signups and closed-won revenue [16]. Agency mode manages prompts across multiple client brands from one account, with transparent usage-based pricing (plans from $28/mo, free audit to start) [17], so tooling stays a predictable pass-through line on every retainer.
Try Agency Mode on your first client, free.
Run a free AI visibility audit on any client's domain, get the full report in about ten minutes, and walk into your next pitch with their score, their gaps, and their competitors' share of voice.
Frequently asked questions
Sources & Academic References
- G2 Buyer Behavior Research, April 2026 — 74% of B2B software buyers start research in an AI chatbot; 85% think more highly of vendors AI includes; via Position Digital, AI SEO Statistics 2026[Source ↗]
- OpenAI via Reuters, February 2026 — 900M weekly active users; via Omnibound, ChatGPT User Statistics[Source ↗]
- Ahrefs, December 2025 — AI Overviews Reduce Clicks by 58% (300,000 keywords, GSC data)[Source ↗]
- Graphite.io / SparkToro, 2026 — AI assistants ≈ 45B sessions, ~56% of global search volume; via Mentionova, AI Search Adoption Statistics[Source ↗]
- Backlinko / Previsible / Position Digital compilations, 2026 — 2.5B daily prompts; +527% LLM referral traffic; G2 AEO category +2,000%; 6.5× third-party citation likelihood; via Cintra, AI Search Statistics 2026[Source ↗]
- Conversion studies: Opollo 2026 AI Search Benchmark (14.2% vs 2.8%, 312 B2B firms; 6.4% of traffic by Jan 2026, +975% YoY); Ahrefs internal data, June 2025 (23×); Seer Interactive case study (ChatGPT 15.9% vs 1.76%); Semrush 2026 (4.4×); tracking gap (14%) via AirOps[Source ↗]
- Seer Interactive, September 2025 — AIO Impact on Google CTR (organic CTR 1.76%→0.61% with AIO; cited brands +35% organic CTR)[Source ↗]
- Pew Research, July 2025 (8% vs 15% click rate); ISB/Carnegie Mellon randomized field experiment, 2026 (−38%); via Search Engine Journal[Source ↗]
- Pricing benchmarks: The Remarkable Agency; The Digital Elevator AEO/GEO Pricing Guide (retainers $3–15k/mo, audits $1.5–5k, sub-$1,500 = rebranded SEO; ~85% of AI citations from third-party sources)[Source ↗]
- Adobe Digital Insights, Q1 2026 — AI referral traffic converted 42% better than non-AI traffic in March 2026; via Digital Applied[Source ↗]
- Pierview, 2026 — How to Price AEO/GEO Services (pricing models, tier structure, tooling pass-through, performance-pricing caution)[Source ↗]
- QuickSEO, 2026 — AI Search Adoption by Industry 2026 (88–96% B2B SaaS AI-invisibility figure)[Source ↗]
- Stackmatix, 2026 — AEO Services Pricing (hourly rates, enterprise projects to $100k, first citations in 60–90 days, ROI 3–6 months)[Source ↗]
- Humanswith.ai, 2026 — What AEO and GEO Actually Cost (premium over $500–1,000 common SEO retainers; tool tiers $29–489/mo)[Source ↗]
- ThriveStack Research, July 2026 — Best AEO Tool 2026 (44.2% of LLM citations pull from the first 30% of the page; category landscape)[Source ↗]
- ThriveStack citedby — product page (73+ signals, six engines, schema generation, revenue attribution)[Source ↗]
- ThriveStack — AI Visibility pricing ($28/mo billed annually, $49 base + usage, agency mode)[Source ↗]
Published by ThriveStack GTM Research, July 2026. Third-party figures are reported as published by their sources; ranges reflect disagreement between studies rather than a single consensus number. ThriveStack citedby is a ThriveStack product; Part 07 describes our own offering.
Get the latest SaaS growth signals
Join SaaS Founders and Growth Leaders
Join 200,000+ growth leaders. No spam, just high-signal insights.