AEO Case Studies: How Real Teams Win Citations, Traffic, and Trust
Read these for the pattern, not the specifics. Then ask: which situation looks most like mine — and what’s the first step I’d borrow? Seven composite case studies across SaaS, law, DTC, consulting, e-commerce, healthcare publishing, and a cautionary tale. The cross-cutting pattern: winners lead with the answer, invest in first-party data, and decouple clicks from visibility. Six of seven cases saw flat or declining traffic alongside growing citation share and brand outcomes — plan for both curves.
Key Takeaways
- Every winning case led with the answer, added first-party data, and layered in credentials.
- Plan for the decoupled traffic + visibility curves — six of seven saw flat/declining traffic alongside rising citations.
- Freshness compounds. Truth Review beats Truth Sprint.
- Genuine assets win. Scaled mediocrity loses, fast and publicly.
- Don’t copy all seven. Pick one pattern. Steal one tactic. Run a 30-day experiment.
Case 1 — The SaaS Knowledge Base That Became a Citation Magnet
Profile: mid-market B2B SaaS in HR analytics, ~80 employees, strong brand in HR circles but invisible in AI answers.
What they did:
- Audited 50 “how do I calculate” queries from AlsoAsked and support tickets; competitors had cleaner answer modules.
- Rewrote top 40 KB articles to lead with a 3-sentence Answer Module: definition, formula, worked example.
- Added a Proof Layer: one data point from their customer base (anonymized) plus one external source.
- Built a calculator widget for five of the most-searched metrics, linked from relevant articles.
Results at 90 days:
- Citation share across 50 tracked queries: 3% → 28%.
- Direct traffic unchanged; ChatGPT-referred sessions grew from near-zero to ~1,200/month.
- Sales attributed 6 pipeline-qualified opportunities in Q2 to “cited in ChatGPT.”
Smart Tip: Knowledge bases are AEO gold if you rewrite them to be citable. Answer Module + first-party data point is 80% of the win.
Case 2 — The Law Firm That Won Local AEO Without Writing Anything New
Profile: 12-attorney personal injury firm serving three mid-sized US cities. Heavy GBP competitor density.
What they did:
- Optimized GBPs across all three offices with consistent NAP, detailed categories, weekly Posts answering one common question each.
- Added Organization + Person schema with credentials, bar admissions, case outcomes (where legally permitted).
- No new content. Added Answer Modules to seven existing practice-area pages, pulled from a paralegal’s intake notes.
Results at 120 days:
- AIO citation share on 30 tracked local queries: 0% → 22%.
- GBP calls up 31%; website contact-form submissions up 19%.
- Managing partner reported being “named specifically” in ChatGPT answers to three consultation leads in one quarter.
Case 3 — The DTC Beauty Brand That Recovered From AIO Traffic Loss
Profile: ~$40M revenue DTC skincare brand, 70% of traffic from organic blog posts on ingredient education.
What they did:
- Rebuilt top 25 ingredient pages with a citation-friendly rewrite: direct-answer module + named expert byline (board-certified dermatologist).
- Added a proprietary Ingredient Safety Scorecard built from FDA and CIR data, displayed as a structured table.
- Published a quarterly “Skin Research Report” with original survey data from 5,000 customers (gated only behind email).
Results at 6 months:
- Organic traffic: still down 35% vs pre-AIO baseline (the structural loss is real).
- AIO citations grew to appear on 60%+ of tracked ingredient queries, up from near-zero.
- Branded search volume rose 28% — citation-first visibility generated the brand awareness previously won by clicks.
Smart Tip: Traffic and visibility have decoupled. The brand lost clicks but won mindshare — and eventually, branded demand.
Case 4 — The Enterprise Services Firm That Discovered ChatGPT Was Their Top-of-Funnel
Profile: global consulting firm, 3,000 employees, selling $500K+ engagements to Fortune 500 CFOs.
What they did:
- Ran monthly prompt audits across ChatGPT, Perplexity, Claude, Gemini on 40 buyer-intent queries.
- Found they were cited in 18% of responses; competitors in 45–60%.
- Reverse-engineered: cited competitors had firm-branded research, partner bios with credentials, Wikipedia-adjacent third-party coverage.
- Commissioned two pieces of proprietary research, pushed partner thought leadership into HBR and MIT Sloan, built Wikidata entries for top 20 partners.
Results at 12 months:
- Citation share across 40 queries: 18% → 49%.
- Two Q4 engagements (~$2.4M combined) had first-touch attribution to “ChatGPT recommended you” per the client’s disclosure.
Case 5 — The E-commerce Category That Lost 30% of Clicks and Still Grew Revenue
Profile: home fitness retailer, 40,000 SKUs, $120M annual revenue.
What they did:
- Accepted the click loss; stopped trying to optimize for CTR recovery on AIO-heavy queries.
- Rebuilt top 30 buyer’s-guide pages to be the source AI cited: comparison tables, “Best For” rules, first-party testing data.
- Added ChatGPT-specific landing pages for products most asked about, with UTM tags.
- Made product feed available via public API and structured data.
Results at 6 months:
- Organic clicks: still down 30%.
- AIO citation share on tracked queries: 11% → 54%.
- Revenue from affected categories up 12% YoY — fewer clicks, higher-intent clicks, meaningful share of sales from AI-started journeys.
Smart Tip: Don’t chase clicks you can’t recover. Invest in being the cited source; the traffic you get will convert better.
Case 6 — The Healthcare Publisher That Rebuilt Editorial for AEO
Profile: independent medical information site, 2M monthly readers, 8-person editorial team with medical reviewers.
What they did:
- Added visible Medical Reviewer byline + credentials block to every article, linked to Person schema page.
- Instituted a quarterly Truth Review: every article over 10K pageviews re-checked against current guidelines, confirmed/updated/archived.
- Built a Condition Hub for 50 common conditions with a consistent template (each section a standalone Answer Module).
Results at 9 months:
- AIO citation share on tracked medical queries: 8% → 31%. Did not catch Mayo Clinic but became the clear #4–5 cited source.
- Truth Review caught 14 articles out of date; two corrections became the AI-cited version within 30 days.
- Traffic flat YoY — a win in a vertical where most publishers lost 20%+.
Case 7 — Cautionary Tale: The Startup That Tried to Shortcut AEO With AI Content
Profile: seed-stage fintech-adjacent startup, 12 employees, trying to build organic visibility cheaply.
What happened:
- Founders decided AEO was “just writing AI likes” and used GPT-4 to generate 200 articles in two months.
- Google’s March 2025 helpful-content update de-indexed 60% of the content within one cycle.
- Citation share in AI answers: zero. AI systems cited the original sources their generators were trained on, not the derivative site.
- Domain authority dropped 14 points. Six months and ~$80K lost.
Recovery:
- Deleted 180 of 200 articles.
- Commissioned 12 pieces of original research from their own anonymized transaction data, published with named bylines.
- Partnered with two industry associations for co-authored content.
- 12-month recovery: citation share 0% → 9% on a narrower set of 20 queries. Slow, but real.
Myth Buster — Myth: AI-generated content at scale is a shortcut to AEO.
Reality: It’s an anti-pattern. Originality, first-party data, and named expertise are the floor — not the ceiling.
Cross-Cutting Patterns
- Winners lead with the answer. Every case that moved the needle rewrote pages to open with a 2–3 sentence direct answer.
- First-party data is the sharpest lever. Proprietary benchmarks, customer data, original surveys, testing results compound.
- Schema and credentials shift the trust needle. Named experts with Person schema win against better content lacking markup.
- Clicks and visibility decoupled. Six of seven saw flat/falling traffic alongside rising citations and brand outcomes.
- Freshness discipline compounds. Standing Truth Review cadences win long-term share.
- Shortcuts lose. Every case that invested in narrower, genuinely cite-worthy assets beat the instinct to publish more.
How to Apply These to Your Program
Don’t copy all seven. Pick the case most like yours, steal one tactic, run a 30-day experiment.
- Content-heavy and losing traffic (DTC, publishers): rewrite top 20 pages with answer modules + proof layer (Cases 1, 3, 6).
- Credentialed vertical (law, medicine, finance, consulting): add visible expert bylines, Person schema, one proprietary research asset (Cases 2, 4, 6).
- E-commerce: build the buyer’s guide you’d want AI to cite — comparison tables, Best For rules, testing data (Case 5).
- And the rule from Case 7: don’t shortcut. AEO rewards genuine assets; it punishes scaled mediocrity.
Common Mistakes
- Trying to copy all seven cases at once — Pick one. Steal one tactic. Run a 30-day experiment.
- Panicking when traffic falls — Six of seven cases saw flat/declining traffic with rising citations.
- Investing in volume to recover from AIO loss — Recovery comes from narrower cite-worthy assets, not more pages.
- Skipping first-party data because it feels expensive — Even a small benchmark or 50-respondent survey lifts citations.
- Treating Truth Review as one-time cleanup — Make it quarterly and standing.
- Confusing AI hallucinations with bad SEO — Improve the source pages AI is summarizing from.
Action Checklist
- Identify which case most resembles your situation.
- Pick one specific tactic from that case.
- Run it as a 30-day experiment with clear before/after metrics.
- Track citation share, branded search lift, and conversion quality — not just traffic.
- After 30 days, double down or move to the next pattern.
- Avoid the Case 7 trap: never scale AI-generated content as a shortcut.
Frequently Asked Questions
What’s the cross-cutting pattern across the winning cases?
Lead with the answer. Add first-party data. Layer in credentials. All seven winning cases combined these three. The losing case (Case 7) skipped them and tried to scale AI-generated content instead.
Why did most cases see traffic decline alongside citation growth?
The decoupling is structural — AI Overviews answer many queries on the SERP, so clicks drop even as citation share grows. Brands that planned for both curves grew revenue and brand strength even with smaller traffic.
Should I worry if my AEO traffic is down 30%?
Not if your citation share is up. Case 5 lost 30% of clicks and grew revenue 12% YoY because the remaining clicks were higher-intent. Track conversion quality, not just volume.
Can a small business compete in AEO?
Yes — Case 2 (a 12-attorney law firm) and Case 7’s recovery (a 12-employee startup) show that focus and authenticity matter more than scale. Pick a narrow niche, lead with the answer, add first-party data, and ship credentials.
What’s the lesson from the AI-generated content cautionary tale?
Scaled AI-generated content is an anti-pattern. AI engines cite the original sources their generators were trained on — not the derivative site. Originality and named expertise are the floor for AEO, not the ceiling.
How long until AEO investment shows results?
Citations and mentions: 30–90 days. Conversions and pipeline: 60–180 days. Brand-search lift: 90–180 days. Plan for the visibility curve to lead the revenue curve by one to two quarters.
Sources & Further Reading
- Pew Research — Google AI summaries
- Conductor — AI Overviews analysis
- SearchPilot — GEO A/B testing
Work With Riman Agency
Riman Agency runs AEO programs across SaaS, services, e-commerce, and publishing. Get in touch to identify which case pattern fits your business — and ship the first 30-day experiment.
Part 28 of our 29-part AEO series. Previous: E-commerce AEO. Up next: The AEO Toolkit Appendix.