CASE STUDY 6 · CONTENT INTELLIGENCE
HealthTech SaaS Platform: How Structured Content Moved the Needle From 0 to 3 of 10 AI Prompts in 45 Days
A Pune-based HealthTech SaaS company providing hospital management software was invisible in AI responses for every buyer query their sales team tracked. Competitors with smaller product footprints were being recommended because their content was structured for AI extraction. Forty-five days of content intelligence work produced citations in 3 of 10 buyer prompts.
| Sector | HealthTech SaaS / Hospital Management |
|---|---|
| Location | Pune, India |
| Anonymised | Yes — client name withheld |
The Challenge
The company served over 200 hospital and clinic networks across India with a full-stack HMS platform covering OPD management, EMR, pharmacy, lab, and billing. Their product was genuinely comprehensive. But "best hospital management software India" on ChatGPT returned three competitors consistently, and the company did not appear in any response.
Their content marketing team had published 40 blog posts over 18 months. None were being cited by AI. The team had been optimising for Google but had never considered how AI platforms read and extract from their content.
What the Diagnostic Found
All 10 buyer prompts showed zero citations for the company across ChatGPT, Perplexity, Gemini, and Claude.
The citation gap map showed that competitors were being cited from three source types: G2 comparison pages, a Healthcare IT India editorial article on top HMS vendors, and a Quora thread answering "what HMS software do Indian private hospitals use?"
The company had a G2 listing with 3 reviews in the wrong subcategory (Practice Management instead of Hospital Management). The Healthcare IT article listed 8 vendors — the company was not one of them. The Quora thread had 12 answers — the company was not mentioned in any.
Of their 40 blog posts, none opened with a direct answer to a buyer question. Every post started with context: "In today's fast-paced healthcare environment..." AI platforms skipped all 40 and cited cleaner sources.
What Dattva Did
Dattva's Content Intelligence service had identified which of the 40 existing posts were closest to earning citations and needed only structural restructuring. The 45-day sprint focused on three actions.
First: three existing blog posts were identified as having the right topic but wrong structure. "Choosing an HMS for a Multi-Speciality Hospital," "Understanding EMR Implementation in Indian Hospitals," and "OPD Management Software: What Hospitals Need in 2026" were each restructured — opening paragraph replaced with a direct 50-word answer, one data point with a named source added in paragraph two, FAQPage schema added with five questions each.
Second: the G2 listing was recategorised under Healthcare Management Software and Hospital Information Systems. The description was rewritten to answer "what does this software do" in the first two sentences. Four review requests were sent to operational clients specifically mentioning the platform's EMR and billing modules.
Third: a Quora answer was posted to the hospital management thread with a genuine practitioner response citing the company's 200+ hospital client base and one specific implementation outcome. The answer was written by the company's implementation lead with their genuine expertise, reviewed by Dattva for AI extraction structure.
The Lesson
Publishing 40 blog posts for keyword ranking and publishing 3 blog posts for AI extraction are different activities producing different results. The company did not need more content — they needed their existing content restructured. Three posts with direct answer openings, named data points, and FAQPage schema produced more AI citations in 45 days than 40 posts optimised for Google produced in 18 months.
See Content Intelligence and the GEO Content Engine.
FAQ
Why were 40 existing blog posts producing zero AI citations?
Every post started with a context paragraph before answering the buyer's question. AI platforms extract the first complete self-contained answer to an implied query. If that answer is not in the first 60 words, AI either skips the page or extracts a low-confidence paraphrase without attribution. The posts were well-written for human readers scanning a list of search results — not for AI systems looking for the direct answer.
What makes a data point citation-worthy in AI?
Named source, specific number, and verifiable publication year. "Hospitals using digital EMR saw 23% reduction in prescription errors (FICCI Health Report 2025)" is citation-worthy. "Studies show significant improvements in healthcare outcomes" is not. AI platforms extract attributable facts — the more specific and sourced, the more likely the extraction.
Why does G2 category placement matter for AI visibility?
AI platforms read G2 as an authoritative third-party source for B2B software recommendations. A listing in the wrong G2 category means the product appears in AI responses for the wrong buyer queries. Recategorising to the correct sub-category (Hospital Management Software vs Practice Management) determines which buyer prompts the company can appear in.
What is the right mix of content types for a B2B SaaS company doing GEO?
Based on Dattva's diagnostic data: structured articles on the company's own domain produce the most durable citations. Review platform presence (G2, Capterra) produces citations for comparison queries. Community platform responses (Quora, Reddit for B2B) produce citations for informational queries. The mix should be 40% structured articles, 40% review platform, 20% community — weighted toward whichever source type the citation gap analysis shows competitors winning from.
See where your brand stands in AI answers today
Run a free AI Visibility diagnostic across ChatGPT, Perplexity, Gemini, and Claude — with prioritised, copy-paste fixes at no cost.