Four disciplined pillars — Citation, Authority, Retrieval, Entity — that compound a brand's citation share across AI Overviews, ChatGPT, Perplexity, Gemini, and Claude. Built from real client work, not invented for marketing.
Free 30-min scoping call · No obligation
In late 2024 we started tracking a pattern across our retainer book. Clients ranking well on Google were getting bypassed inside AI answers. A fintech holding position #2 on a comparison query was absent from the Perplexity response on the same intent. A D2C brand owning a category SERP was being cited by Google's AI Overview as "according to" — but with the answer pulled from a 2019 blog post on a third-party site.
Generic "AEO" advice did not close those gaps. The bulk-schema-everywhere approach added noise without lifting citation share. The Reddit-spamming approach got accounts banned. The "just write better content" approach ignored that the engines were not refusing to cite us — they were refusing to find us in the first place.
CARE is what fell out of two quarters of working backwards from per-engine citation gaps to the specific levers that closed them. Four pillars, named because each addresses a different failure mode. A deliberate sequence, because the foundation pillars (Citation and Retrieval) have to ship before the compounding pillars (Authority and Entity) have anything to compound on.
CARE is the methodology behind every Nico Digital AEO and GEO retainer. It is the working name for what we already do, not a separate product to upsell. The rest of this page lays out the pillars, the sequence, the measurement model, and — as importantly — the things CARE explicitly refuses to do.
The four letters map to four distinct reasons a brand fails to be cited.
| Pillar | Failure mode it fixes | Symptom in a citation audit |
|---|---|---|
| Citation | Engine can't extract a clean answer from your page even if it lands there | Page appears in retrieval but not in the cited answer |
| Authority | Engine doesn't trust your source enough to cite it over a competitor | Page extracted, but only as a secondary source, not the lead citation |
| Retrieval | Engine can't find your page in its retrieval layer in the first place | Page exists but never surfaces in any prompt audit |
| Entity | Engine doesn't recognize your brand as a distinct entity worth grounding to | Brand mentioned by generic descriptor only; named competitors get cited instead |
Each pillar carries its own deliverables and its own measurement model.
Citation
Engineering the surfaces AI engines actually pull from when answering category queries.
Deliverables
Authority
Building the E-E-A-T signals that decide whether an engine cites your brand or a competitor.
Deliverables
Retrieval
Making the brand findable inside the retrieval layers each engine uses to ground its answers.
Deliverables
Entity
Establishing the brand and its founders as distinct entities in the Knowledge Graph and LLM training data.
Deliverables
The 90-day rhythm that drives the first measurable citation movement.
Weeks 1–2
Diagnose
Five-engine citation audit (ChatGPT, Perplexity, Gemini, Claude, Copilot) on the brand's top 30 priority queries. Output: a real gap report tied to each CARE pillar, not a generic AEO deck.
Weeks 3–6
Foundation
Citation + Retrieval pillars first. Schema shipped to priority pages, short-answer blocks deployed, Bing Webmaster + IndexNow wired up, llms.txt routes built. Why first: nothing else compounds without these basics in place.
Weeks 7–10
Authority + Entity
Author E-E-A-T hub builds, Wikidata work, editorial PR pipeline started. Original-research data study scoped if the engagement includes one.
Weeks 11+
Compound
Monthly per-engine reporting against citation share, AIO appearance rate, branded-search lift. Schema and content layer extended page by page. Authority + entity work is the slowest to show — most clients see meaningful movement at month 4–6.
What an engagement looks like in practice. Anonymised — D2C beauty brand, ₹35Cr revenue, six months in.
Starting state
Strong SEO foundation: ranking top-3 on 40+ category queries. Branded search healthy. ChatGPT, Perplexity, and Gemini citation share on category queries: under 5%. AI Overviews citing them on 2 of 25 priority queries. Knowledge Graph entry: none. Wikidata: empty. Founder Person schema: none. The brand was visible to Google and invisible to every other engine.
Weeks 1–2 · Diagnose
Five-engine prompt audit on the top 30 priority queries. Output identified the failure mode per query — 18 were Retrieval failures (pages existed but never surfaced), 8 were Citation failures (pages surfaced but were not extracted from), 4 were Authority failures (extracted as secondary, not lead). Entity failure was system-wide — the brand was not recognised as a distinct entity in any engine's grounding layer.
Weeks 3–6 · Citation + Retrieval foundation
Schema deployment scoped to the 18 Retrieval-failure pages first (FAQPage, HowTo, Article, ItemList where relevant). Short-answer blocks inserted at the top of the 8 Citation-failure pages with definition-first paragraphs and structured comparison tables. Bing Webmaster + IndexNow integration shipped to fix the underlying retrieval gap — ChatGPT Search pulls from Bing's index, and the brand's Bing index health was thin. llms.txt + llms-full.txt routes deployed.
Weeks 7–10 · Authority + Entity layer
Wikidata entry created for the brand (founders, founding year, headquarters, sector). ProfilePage schema deployed for founder + lead PR voice. Original-research data study scoped — a 1,200-respondent consumer survey on category buying behaviour. Author byline programme started: senior team members started publishing under their own bylines in trade press where editorial relationships already existed. Reddit + Quora seeding limited to genuinely useful contributions on three high-leverage subreddits — no spam, no scripted activity.
Weeks 11–24 · Compound
Schema + content work extended page by page. Research study published with 14 outbound trade-press placements. Monthly per-engine reporting cadence established. At month 6: ChatGPT citation share on category queries up to 28%, Perplexity up to 34%, Gemini up to 19%. AI Overviews citing on 11 of 25 priority queries. Branded search lifted 18% over the period — the indirect read on AI citation share working.
Lessons from this engagement
The differences that actually matter when judging methodologies.
| Dimension | Generic AEO playbook | CARE |
|---|---|---|
| Diagnosis | Generic schema audit + content review | Per-engine prompt audit on top 30 queries, classifying each gap by failure mode |
| Sequencing | Everything in parallel, no priority | Citation + Retrieval first, Authority + Entity second — because the latter compound on the former |
| Retrieval focus | Rarely discussed — usually skipped | Treated as a distinct pillar; Bing index health, crawl budget, RAG-readiness explicitly audited |
| Schema philosophy | Add FAQPage + HowTo everywhere | Schema scoped to query intent — Speakable, ClaimReview, DefinedTermSet, ItemList where the engine actually needs them |
| Entity work | Mentioned, rarely executed | Wikidata + Knowledge Graph + ProfilePage shipped as a defined deliverable, not aspiration |
| Measurement | Rankings, "media impressions," screenshots | Per-engine citation share, AIO appearance rate, branded-search lift, citation-to-pipeline attribution where data allows |
| Exclusions | Rarely declared | Explicit list of refused tactics published on this page |
Five metrics. Reported monthly. Each chosen because it is harder to game than the metric it replaces.
Per-engine citation share
What percentage of priority queries cite the brand inside ChatGPT, Perplexity, Gemini, Claude, Copilot. Measured via systematic prompt audits, not screenshots.
AI Overview appearance rate
Across a defined query set, what percentage trigger an AIO that cites the brand. Separated from rankings because the two move independently.
Branded-search lift
The leading indirect indicator that AI citations are working. When AI engines cite a brand, branded search rises before direct traffic does.
Retrieval coverage
Percentage of priority pages indexed in Bing, surfaced in Perplexity's retrieval layer, and reachable inside ChatGPT Search. Often the biggest hidden gap.
Citation-to-pipeline attribution
Where analytics + CRM allow, attempted attribution from AI citation sessions to qualified pipeline. Imperfect (AI referrers are often stripped), but worth modelling. We disclose the modelling assumptions on every report.
The exclusions matter as much as the inclusions.
Citation, Authority, Retrieval, Entity. The four pillars an Indian brand has to work on to move citation share inside AI Overviews, ChatGPT, Perplexity, Gemini, and Claude. It is the working name for the methodology we already use inside Nico Digital AEO retainers — not a separate product.
Because 'AEO' has been diluted to mean almost anything. CARE is specific — four pillars, defined deliverables under each, a sequence (Citation + Retrieval first, Authority + Entity second), and a 90-day diagnostic-then-foundation rhythm. Naming the methodology forces honesty about what is in scope and what is not.
No. CARE is the methodology behind every AEO and GEO retainer we run. It is not an additional invoice. Pricing follows the published bands on our AEO and GEO pricing pages.
Generic checklists list every possible AEO tactic without sequencing. CARE forces sequence (Citation + Retrieval before Authority + Entity), forces measurement (per-engine citation share, not rankings), and explicitly excludes the things that do not work — bulk schema dumps without strategy, Reddit-spam-as-a-service, paid Wikipedia placements. The exclusions matter as much as the inclusions.
Depends on starting state. A brand with strong SEO but weak schema starts at Citation. A brand with great content but no Bing index health starts at Retrieval. A brand with neither but a recognisable founder may get more leverage from Entity work first. The diagnostic in weeks 1–2 sets the order — we do not run a generic sequence regardless of starting state.
Yes — the pillars are engine-shaped, not market-shaped. The local-press piece of Authority shifts (Indian tier-1 publications for India briefs, US trade press for US briefs), and the Entity work needs market-specific Wikidata coverage. The Citation, Retrieval, and core Authority work translates directly.
Conservative: 4–8 weeks for first AIO + ChatGPT citations on long-tail queries where competition is thin. 3–6 months for category-level citation share on competitive answer-intent queries. Authority + Entity work is the slowest layer — most clients see meaningful compounding at month 4–6.
Yes, if you have a senior schema engineer, a content strategist comfortable with AI extraction formatting, someone who can navigate Wikidata + Wikipedia policy without getting pages deleted, and a person to run per-engine citation audits monthly. That is roughly the headcount of a 4–5 person in-house team. Below that, an agency engagement is usually cheaper.
30-minute call. We will audit your current AEO surface across five engines and scope a realistic CARE engagement. No deck, no pressure.