Resource
This page describes the AI visibility service. Engagements start with a 300-prompt visibility battery across ChatGPT, Google AI Overview and Google AI Mode, then move into structural remediation against the specific extraction failures the battery surfaces. The output a client buys is a baseline mention-rate number, a delta target, and the engineering tickets that close the gap. No prediction. No “best practice.” Just the measured rate before and after.
Brands that rank well on the classical 10-blue-links SERP are discovering they do not surface in conversational AI answers for the same commercial intent. The mechanics are different. ChatGPT, AI Overview and AI Mode retrieve, summarise and cite based on entity-extraction logic, structured-data presence and citation density across the wider web. A page that ranks #1 on Google for “best instant business loan” can show a single-digit mention rate inside ChatGPT for the same query.
The number that proves this is not a hypothetical. On a 25,000-page NBFC site that pulled 2 million organic visits per month, the 300-prompt visibility battery returned 8 percent on ChatGPT, 15.6 percent on AI Overview and 19 percent on AI Mode. Translated: for every 100 commercial questions a buyer asked the assistant, the brand surfaced fewer than 20 times. The competitor with one-third the organic footprint surfaced more often, because its product pages carried cleaner schema and its third-party citations were denser.
The harder problem is that most analytics stacks do not measure this. GA4 will not tell a brand whether ChatGPT is recommending it. GSC will not surface AI Mode citations. Without an external prompt battery, the AI visibility gap is invisible to the marketing team until a sales rep mentions “our customers said ChatGPT recommended Competitor X.”
The engagement runs in four phases, each with a defined deliverable.
Phase 1: Prompt battery design and baseline measurement. A 300-prompt set is constructed from three inputs: the client’s paid keyword universe (DataForSEO + ahrefs), the top-converting commercial intents from GA4, and 50 prompts written against jobs-to-be-done that buyers articulate during sales calls. The battery is fired through ChatGPT (API), Google AI Overview (live SERP capture) and Google AI Mode (live SERP capture). Output is a CSV with prompt-by-model brand mention status, competitor mention status, citation source for each mention, and a fabrication flag where the model invented a feature.
Phase 2: Citation and schema gap analysis. Where a competitor wins a citation and the client does not, the source page is examined. Three patterns recur. Pattern A: competitor page carries Product, FAQPage, HowTo or Service schema with named entities; client page carries WebPage only. Pattern B: competitor is cited by 8 to 15 third-party publications with the exact entity name in title and H1; client is cited by 1 to 3 with inconsistent entity naming. Pattern C: competitor’s product page reads as a spec sheet (numbers, ranges, terms); client’s reads as a marketing brochure. The gap analysis ships as a per-prompt remediation ticket.
Phase 3: On-site structural fix. Schema is written page-by-page (not via a sitewide plugin) so entity boundaries are explicit. Product specifications are rewritten as quantified tables. FAQPage blocks are derived from the prompt battery itself, not from PAA scraping. Internal linking is restructured so the highest-citation pages on the site point to the entity-poor pages that need lift. The roadmap is sprint-phased and handed off as developer tickets. See the technical SEO page for the audit specification this layer plugs into.
Phase 4: Re-measurement and citation acquisition. The same 300-prompt battery is re-fired at 30, 60 and 90 days. The delta is the deliverable. In parallel, the off-site team works on third-party citation acquisition (industry publications, expert directories, structured data submissions) targeted at the specific entities the battery showed as weak.
A major BFSI lender came in with 2M organic visits per month, 94.1K ranking keywords and 578K backlinks, but an opaque sense of how the brand fared inside the new AI surfaces. The 300-prompt battery returned the 8 / 15.6 / 19 numbers above. The schema audit found 81 percent of pages with no canonical, 224 invalid structured-data items, and a 78 percent hreflang error rate. The roadmap shipped as a Drupal-and-Akamai-specific 16-sheet workbook with drush commands, theme paths and Akamai cache rules. The engineering team executed against the audit ticket-by-ticket because the deliverable spoke their language. The point of the engagement: the AI visibility gap was a downstream symptom of structured-data debt, and fixing the debt fixed the citation rate.
A specialty hospital chain entering Chennai assumed it was invisible against the legacy networks (Apollo, Kauvery, Rela, MIOT). The brief asked for a paid-media plan to “build awareness from zero.” The team here ran a multi-LLM verification pipeline (four parallel WebSearch agents plus Gemini CLI plus DataForSEO, with a Claude supervisor that rejected five fabricated Gemini claims) across 30 priority Chennai kidney and urology queries. The data inverted the brief. The client was already #1 among branded Chennai hospitals on kidney-care visibility: 11 top-3 ranks across 30 priority queries, 25 top-10 ranks, 14 of 30 local-pack appearances (most-frequent single brand), outranking competitors 4 to 33 times larger by domain index. The strategic recommendation that came out of this was a budget split (Search 50 percent, Meta 16 percent, SEO 14 percent, YouTube 10 percent, CRO 10 percent) anchored to the seven specific cluster gaps the audit found (dialysis-centre, geo-locality dialysis, robotic urology, laser kidney stone, paediatric nephrology, 24-hour dialysis, Tamil informational queries) rather than a generic “build awareness” envelope. The visibility data turned a defensive paid plan into a category-leadership plan.
Week 0: 60-minute scoping call. Read-only access requested to GA4, GSC, the live sitemap and the existing schema. No NDA at this stage.
Week 1 to 2: 300-prompt battery is constructed, reviewed with the client, and fired. The CSV ships at end of Week 2 with a recorded walk-through.
Week 3 to 4: schema and citation gap analysis. Per-prompt remediation tickets are written into the client’s existing ticketing system (Jira, Linear, Asana) or delivered as a Notion / DOCX bundle if there is no ticket system yet.
Week 5 to 10: structural fix sprint. The client’s engineering team executes against the tickets; the ScaleGrowth Digital lead runs daily standup attendance for blocker resolution. Parallel: off-site citation acquisition team works the third-party publication list.
Week 11 onwards: monthly re-measurement, retainer phase. Read the content service for the schema-first content pipeline that feeds this phase.
300-prompt baseline battery + schema and citation gap analysis: fixed at $6,500 / ₹5.4L. Two-week turnaround.
Structural fix sprint (6-week engagement): $14,000 to $28,000 / ₹11.5L to ₹23L depending on site size, framework complexity and number of priority entities.
Retainer (monthly re-measurement + ongoing citation work): $4,500 / month flat. 90-day rolling notice.
Pricing does not include third-party publication placement fees where required (industry directories with paid listing, expert databases). These are passed through at cost with prior approval.
The 300-prompt battery is fired through the OpenAI API using the model variants that match the client’s likely buyer surface (ChatGPT 4o for free-tier buyers, ChatGPT 4o-mini for high-volume cases). Each response is parsed for exact-match brand name, entity variants (parent company, product names) and known competitor mentions. Citation URLs are extracted and logged. The output is a per-prompt mention vector. Aggregate rate is the simple average across the 300 prompts.
Schema improves entity disambiguation at retrieval time for systems that use live web retrieval (AI Overview, AI Mode, ChatGPT with browse, Perplexity). It does not change the weights of the base model. Engagements that measured before-and-after across three months saw double-digit absolute mention-rate gains on AI Overview and AI Mode, mid-single-digit gains on ChatGPT, and no measurable change in queries that did not trigger retrieval at all.
The battery flags fabrications explicitly. Each fabrication is traced to a likely source (often a low-quality directory page or a competitor blog that misrepresented the product). Two parallel fixes run: a clarification post on the client domain with the correct specifications and tight schema, and an outreach pass to the source page asking for a correction. Most fabrications clear within 60 to 90 days as retrieval picks up the corrected source.
Yes. The work overlaps on technical SEO and on-page content but adds a measurement layer the standard retainer does not own. The cleanest split is to keep the existing retainer on classical ranking and backlinks, and add this engagement as a separate AI-visibility track with its own KPIs (mention rate, citation rate, fabrication rate).
Google has retired AI Overview triggers on specific YMYL categories before. The measurement framework is model-agnostic: if AI Overview goes dark for a category, the battery shifts weight to AI Mode, Perplexity and ChatGPT, and the underlying schema and citation work stays valid because the same signals are read by every retrieval system. No engagement has had to be paused because a single model surface changed.
Two weeks, fixed fee, one CSV deliverable. The number a brand actually surfaces at inside ChatGPT, AI Overview and AI Mode for its top 300 commercial prompts. Plus the per-prompt remediation tickets.
Technical SEO Schema Markup That AI Actually Uses vs Schema Markup That Doesn’t Matter We…
Read more →AI Agents WhatsApp AI Agents: Beyond Auto-Replies to Actual Business Outcomes Most WhatsApp “automation” is…
Read more →AI Visibility How to Test Whether Your Content Gets Cited by ChatGPT, Gemini, and Perplexity…
Read more →AI Visibility llm.txt, ai-plugin.json, and WebMCP: The Three Files That Make Your Site AI-Ready AI…
Read more →