GEO Case Study: How a SaaS Brand Went from 0 to 40% AI Citation Coverage in 90 Days

Author: Onxeera Editorial Team | Last Updated: August 2026 | Reading Time: 11 min

TL;DR: This case study documents how a mid-size B2B SaaS brand in the project management category went from zero measurable AI citations to 40% citation coverage across its target query set in 90 days — using a structured GEO implementation program. The brand had strong SEO performance and good content quality, but zero AI citation presence. The intervention combined a technical schema audit, FAQPage schema implementation across 18 pages, Organization entity optimization, answer-first content rewrites on 12 key pages, and systematic citation tracking. Results: 40% citation coverage at 90 days, 58% at 120 days, and a 23% increase in direct brand search volume attributable to AI-driven awareness. The tactics, timeline, and specific interventions are documented in full.

Table of Contents

1. Brand Background
2. Starting Position: The Zero Citation Problem
3. GEO Audit Findings
4. The 90-Day Intervention Plan
5. Month 1: Technical Foundation
6. Month 2: Content Optimization
7. Month 3: Entity and Authority Building
8. Results at 90 Days
9. Results at 120 Days
10. What Worked — and What Didn’t
11. Key Lessons
12. The Replication Framework
13. FAQs
14. Related Articles

Brand Background

The brand in this case study — referred to as “ProjectFlow” throughout — is a mid-size B2B SaaS company offering project management and team collaboration software. Founded in 2019, ProjectFlow serves primarily mid-market companies (100 to 2,000 employees) across professional services, marketing agencies, and technology sectors. At the time of the GEO intervention, ProjectFlow had approximately 3,200 paying customers, $8.4M ARR, and a marketing team of six people including one dedicated content marketer.

ProjectFlow’s SEO position was solid: 47 blog posts indexed, ranking on page one for 23 target keywords, and 38,000 monthly organic visits from Google. The brand had a domain rating of 52 (Ahrefs) and strong on-page content quality across its core product pages and blog. Despite this SEO health, ProjectFlow had essentially no AI search presence — a gap that became apparent when the marketing team began manually testing queries in ChatGPT and Perplexity in early 2026 and found their brand absent from answers to queries they ranked well for in traditional search.

Related: GEO for SaaS Brands | GEO Optimization: The Complete Guide

Starting Position: The Zero Citation Problem

Baseline Citation Audit

ProjectFlow’s marketing team established a citation baseline by testing 35 target queries across four AI platforms — ChatGPT (Browse mode), Gemini, Perplexity, and Google AI Overviews. The query set covered five categories: brand identity queries (“what is ProjectFlow?”), category recommendation queries (“best project management software for agencies”), feature comparison queries (“which project management tool has the best time tracking?”), use case queries (“project management software for remote teams”), and FAQ queries (“how much does project management software cost?”).

Baseline results across all 35 queries and 4 platforms (140 query-platform combinations total): ProjectFlow cited in 4 of 140 combinations — a 2.9% citation rate. The 4 citations were all on Perplexity for branded queries (“what is ProjectFlow?”) and were brief, incomplete descriptions with no mention of key features or differentiators. On ChatGPT, Gemini, and Google AI Overviews, ProjectFlow had zero citations across all 35 queries — including queries for which ProjectFlow ranked on page one of Google’s traditional search results.

The SEO-GEO Gap

The SEO-GEO gap was striking: ProjectFlow ranked position 3 for “project management software for marketing agencies” in Google traditional search — but was not cited once for this query across any AI platform. It ranked position 7 for “best project management tools for remote teams” — and again, zero AI citations. This disconnect confirmed a pattern observed across the industry: traditional SEO performance does not automatically translate to AI citation performance. The signals that earn traditional search rankings (backlinks, on-page keyword optimization, click-through rates) differ significantly from the signals that earn AI citations (structured schema, answer-first content format, entity clarity, FAQPage implementation, content comprehensiveness).

GEO Audit Findings

A comprehensive GEO audit identified the specific gaps responsible for ProjectFlow’s zero citation position. The audit covered four areas: schema implementation, content format, entity signals, and crawler access.

Schema Audit Findings

• No Organization schema on the homepage — the most fundamental entity signal was entirely absent
• No FAQPage schema on any of the 12 blog posts that contained FAQ sections — the FAQ content existed but was invisible to AI citation systems as structured data
• Article schema present on blog posts (via Yoast SEO) but missing dateModified on 41 of 47 posts — all showing the original publication date despite many posts having been updated
• SoftwareApplication schema absent on all product pages — product features, pricing, and compatibility were described in HTML text with no structured data
• No BreadcrumbList schema on any page

Content Format Audit Findings

• 0 of 47 blog posts used answer-first paragraph structure — all opened with contextual introductions rather than direct answers to the primary query
• 12 blog posts had FAQ sections but no FAQPage schema — the highest-priority content gap given FAQPage schema’s direct citation impact
• Heading structure was keyword-optimized but not question-formatted — H2s like “Time Tracking Features” rather than “Does ProjectFlow have time tracking?”
• Product pages were feature-list heavy with minimal FAQ or answer-first content

Entity Signal Audit Findings

• No Wikidata entry for ProjectFlow
• LinkedIn company page present but generic description (“ProjectFlow helps teams work better together”) with no founding year, no specific product category, and no industry classification beyond “Software Development”
• G2 profile present with 47 reviews (3.8 average) but not linked from the website via sameAs
• Crunchbase profile missing founding date and had incorrect HQ city (an old office location)
• No sameAs array in any schema (since Organization schema was absent entirely)

Crawler Access Audit Findings

The crawler access audit was clean — no robots.txt blocks for AI crawlers, no noindex tags on content pages, and HTTP 200 responses confirmed for GPTBot, PerplexityBot, and Googlebot. This was good news: the zero citation problem was not a crawler access issue but a content and schema quality issue. All pages were accessible — they just lacked the structured signals that AI citation systems needed to select them as citation sources.

The 90-Day Intervention Plan

Based on the audit findings, a prioritized 90-day intervention plan was developed. The plan organized work into three monthly phases: Month 1 (technical schema foundation), Month 2 (content format optimization), and Month 3 (entity authority building). Each phase was designed to produce measurable citation improvement before the next phase began — allowing the team to validate the approach and build confidence before investing in the more time-intensive content optimization work.

Priority Rationale

Schema fixes were prioritized in Month 1 because they require no new content creation, can be implemented quickly, and have the most predictable citation impact timeline (4 to 6 weeks for FAQPage schema effects to register). Content format optimization was placed in Month 2 because it requires more effort per page and has a longer citation impact timeline (6 to 8 weeks). Entity authority building was placed in Month 3 because it has the longest impact timeline (8 to 12 weeks) and benefits most from the schema and content foundation established in Months 1 and 2.

Month 1: Technical Foundation (Days 1-30)

Week 1: Organization Schema Implementation

Organization schema was implemented on the homepage with: canonical brand name, legal name (“ProjectFlow Inc.”), founding date (2019-03-15), headquarters address, specific description (“B2B project management and team collaboration SaaS platform for mid-market professional services, marketing agencies, and technology companies”), sameAs array linking to LinkedIn, Twitter/X, G2, Capterra, Crunchbase, and ProductHunt profiles, and knowsAbout populated with 12 specific terms including “project management software,” “team collaboration tools,” “agile project tracking,” “resource allocation software,” “client project management,” and “remote team coordination.” Implementation time: 3 hours including validation with Google Rich Results Test.

Week 1-2: FAQPage Schema — Batch Implementation

FAQPage schema was implemented on all 12 blog posts with existing FAQ sections simultaneously. The schema was added via a custom HTML block in WordPress (Yoast SEO did not auto-generate FAQPage schema for the existing FAQ format used on ProjectFlow’s blog). Each FAQ section had 4 to 8 questions — the schema implementation for each post took approximately 30 to 45 minutes including content matching (ensuring schema question and answer text matched the visible page content exactly) and validation. Total implementation: 12 pages, approximately 8 hours of work. All 12 pages were submitted to Google Search Console URL Inspection → Request Indexing on the day of implementation.

Week 2: Article Schema dateModified Fix

Yoast SEO was configured to automatically update Article schema dateModified whenever a post was saved. For the 41 posts with stale dateModified values, each post was opened, a minor content update was made (adding one current statistic or updating one outdated reference), and the post was saved — triggering a dateModified update to the current date. The content updates were substantive enough to justify the dateModified change and were not cosmetic-only edits. This process took approximately 20 minutes per post — 41 posts total, completed over 3 days with 2 to 3 posts per hour.

Week 3-4: SoftwareApplication Schema on Product Pages

SoftwareApplication schema was implemented on ProjectFlow’s 4 primary product pages (core platform, time tracking add-on, client portal add-on, and reporting dashboard). Each schema block included: applicationCategory (“ProjectManagementApplication”), featureList (8 to 12 specific features per page), operatingSystem, offers with current pricing and priceCurrency, aggregateRating (populated from G2 data — 3.8 rating, 47 reviews), and softwareRequirements. The aggregateRating addition was particularly significant — it converted ProjectFlow’s existing G2 reviews into structured data that AI systems could reference for quality signals without having to parse the G2 platform directly.

Month 2: Content Optimization (Days 31-60)

Answer-First Rewrites on 12 Priority Pages

The 12 blog posts with the highest commercial query value (the pages targeting the most valuable queries in the citation target set) were rewritten to use answer-first paragraph structure. The rewrite approach: identify the primary query the page targets, write a 60 to 80 word opening paragraph that directly answers that query (including the key features, use cases, and differentiators relevant to the query), and move the original contextual introduction to after the first paragraph or integrate it naturally into the second paragraph. Average rewrite time per post: 45 minutes. Total: approximately 9 hours of content work across the month, prioritized 3 to 4 posts per week.

FAQ Section Addition on 6 High-Value Pages

Six blog posts that targeted high-value queries but had no FAQ section received new FAQ sections — 6 to 8 questions each, with FAQPage schema implemented simultaneously. The questions for each FAQ were selected by testing actual AI queries for the page’s topic and identifying which question formulations were being submitted to AI engines but not answered directly on the page. Each FAQ section was added to the bottom of the post, before the conclusion, with the same formatting as existing FAQ sections on other posts to maintain consistency. Average time per post: 90 minutes (question selection, answer writing, schema implementation, validation). Total: approximately 9 hours.

Heading Structure Optimization

H2 and H3 headings on the 12 priority pages were updated from keyword-label format to question format where appropriate. Examples of specific changes: “Time Tracking Features” → “Does ProjectFlow include time tracking?”, “Integrations” → “What tools does ProjectFlow integrate with?”, “Pricing” → “How much does ProjectFlow cost?”, “Security” → “Is ProjectFlow secure and GDPR compliant?” This change was made carefully — not every heading was converted (some structural headings work better as labels), but every heading that addressed a specific question a buyer would ask was converted to question format. dateModified was updated on each page after the heading changes.

Month 3: Entity and Authority Building (Days 61-90)

Wikidata Entity Creation

A Wikidata entity was created for ProjectFlow with: instance of (software company, Q4830453), name, inception date (2019), headquarters location, official website, industry (project management software, enterprise software), founded by (founders’ names, linked to their individual Wikidata entities where they existed), official logo, and external ID cross-references — LinkedIn organization ID, Crunchbase organization ID, G2 company slug, and ProductHunt slug. The Wikidata Q-number was then added to the Organization schema identifier property on the homepage. Total time: approximately 90 minutes.

Platform Profile Optimization

Four platform profiles were updated with specific entity signal improvements: LinkedIn company page (description rewritten with founding year, specific product category, target customer profile, and key differentiators — replacing the generic “ProjectFlow helps teams work better together”), G2 profile (description updated, all product categories correctly selected, and a review generation campaign launched to the customer success team’s active customer list), Crunchbase (founding date corrected, HQ city corrected, description updated, and employee count range updated), and Capterra (full profile review — description, categories, features checklist updated to match SoftwareApplication schema featureList). Each profile update also updated the sameAs URL in Organization schema to confirm the correct profile URL was linked.

Original Data Publication

One original data piece was published in Month 3: a survey of 312 project managers across ProjectFlow’s customer base measuring productivity outcomes from implementing structured project management processes — “The 2026 Project Management Productivity Report.” The report was published as a full HTML page (not PDF-only) with Article schema, a dedicated data section with specific statistics, and a summary infographic. The report was promoted through ProjectFlow’s newsletter and LinkedIn — earning 23 backlinks in the first 30 days and coverage in 3 project management industry newsletters. This piece was designed as both a citation asset (original data that AI engines must cite to use the statistics) and an external authority signal (press coverage and backlinks from industry sources).

Results at 90 Days

Citation Rate by Platform

The 90-day citation measurement tested the same 35 queries across the same 4 platforms (140 combinations) as the baseline. Results:

Citation Rate by Query Category

Named vs Anonymous Citations

Of the 56 citations at Day 90: 41 were named citations (ProjectFlow mentioned by name in the AI answer) and 15 were anonymous citations (ProjectFlow’s content used as a source but not named in the visible answer text). Named citation rate: 73% of total citations. The high named citation rate was attributed to the strong Organization schema and Wikidata entity implementation — AI systems with clear entity data are more likely to name the source in the citation text rather than paraphrasing anonymously.

Results at 120 Days

A 120-day measurement was conducted to track continued citation growth without any additional interventions beyond routine content maintenance (two new blog posts, routine dateModified updates on refreshed content). Results at 120 days: 81 citations of 140 combinations — a 58% overall citation rate. The additional 25 citations between Day 90 and Day 120 came primarily from: the original data publication gaining more indexed coverage and citations (contributing citations on 8 new query-platform combinations), G2 review count growing from 47 to 71 reviews (average improving to 4.1 — improving aggregateRating citation signals), and the Wikidata entity becoming more deeply integrated into AI knowledge graphs (improving brand identity query performance further).

Business Metrics at 120 Days

Beyond citation rate, three business metrics showed measurable changes attributed to improved AI visibility: direct brand search volume (Google Search Console) increased 23% compared to the equivalent period in the prior year — indicating that AI citations were driving brand awareness and subsequent direct searches. Trial signups from “how did you hear about us?” responses showing “AI search” or naming specific AI platforms (ChatGPT, Perplexity) increased from 2 attributed trials in the prior quarter to 19 in the 120-day measurement period. Website referral traffic from Perplexity (tracked via UTM parameters in Perplexity citations) increased from near-zero to 847 sessions over the 120-day period — confirming that Perplexity citations were driving measurable direct traffic.

What Worked — and What Didn’t

What Worked Best

FAQPage schema (highest citation impact per hour invested): The 12 FAQPage schema implementations produced the most immediate and most significant citation rate improvement — citation rates for FAQ-format queries went from 0% to over 60% within 6 weeks of implementation. At approximately 30 to 45 minutes per implementation, FAQPage schema produced the highest citation improvement per hour of any intervention in the program. This validated the consistent finding across GEO research that FAQPage schema is the single highest-ROI GEO technical investment.

Organization schema with knowsAbout (brand identity citation improvement): Brand identity query citation rates went from 11% (Perplexity only, incomplete descriptions) to 82% across all platforms within 45 days of Organization schema implementation. The quality of brand descriptions in AI answers also improved dramatically — AI engines shifted from vague descriptions (“a project management tool”) to specific, accurate descriptions that included ProjectFlow’s target customer, key use cases, and founding context. The knowsAbout property specifically improved citation rates for category queries where ProjectFlow was not previously associated with specific product category terms.

Wikidata entity + Organization schema identifier linkage: The Wikidata entity creation produced a step-change improvement in Gemini citation rates specifically — Gemini’s Knowledge Graph integration means Wikidata entities directly influence its entity recognition. Gemini citation rates went from 0% at baseline to 31% at Day 90, with much of the improvement attributable to the Wikidata entity enabling Gemini to confidently identify and describe ProjectFlow as a distinct, well-defined entity.

What Worked Moderately

Answer-first content rewrites: The answer-first rewrites on 12 pages produced moderate citation improvement for category recommendation and use case queries — citation rates for these query categories improved from near-zero to 22 to 34% over 90 days. The improvement was real but slower to materialize than schema changes (citation improvement was first detectable at 7 weeks rather than the 4 to 5 weeks for FAQPage schema). The answer-first format also required the most editorial judgment to implement well — not every rewrite was equally effective, and some pages required a second iteration after the initial rewrite did not produce expected citation improvement.

dateModified updates with substantive content refreshes: The dateModified updates with substantive content changes produced measurable improvement on Perplexity (the platform most sensitive to freshness signals) — Perplexity citation rates were the highest of any platform at Day 90 (60%), and the freshness signal from current dateModified values was identified as a significant contributing factor. The impact on ChatGPT and Gemini was less direct but contributed to overall content quality signals that improved those platforms’ citation rates over the 90-day period.

What Underperformed

Heading structure optimization: Converting H2/H3 headings from keyword-label to question format produced the smallest measurable citation improvement of any intervention — citation rate changes on the specifically tested pages with heading changes were within the measurement noise margin. The heading format change may have contributed to citation improvements in combination with other changes on the same pages, but as an isolated variable it was the least impactful intervention. This does not mean it was wrong — question-format headings may contribute to longer-term citation performance through improved content parsing — but the 90-day impact was not distinguishable from background noise.

Original data publication (within the 90-day window): The “2026 Project Management Productivity Report” earned backlinks and newsletter coverage but produced limited measurable citation improvement within the 90-day measurement window. The 120-day measurement showed more significant impact — original data publications have longer citation ramp-up times than schema changes, as the data must be indexed, cross-referenced by other sources, and incorporated into AI knowledge systems before citation volume grows. The publication was the right investment but required patience beyond the initial 90-day window to show full impact.

Key Lessons

• Lesson 1: Good SEO does not automatically produce good GEO. ProjectFlow’s solid SEO performance — page one rankings, strong domain rating, quality content — provided zero AI citation advantage at baseline. The signals that earn traditional search rankings differ fundamentally from the signals that earn AI citations. Every organization with good SEO should assume their GEO position is poor until tested — because the test results will almost certainly be surprising.
• Lesson 2: FAQPage schema is the fastest, highest-ROI first GEO investment. The 12 FAQPage schema implementations took approximately 8 hours total and produced the largest citation rate improvement of any intervention in the program. No other investment of equivalent time produced equivalent citation results. Start with FAQPage schema on every page that has existing FAQ sections.
• Lesson 3: Organization schema with complete knowsAbout and sameAs dramatically improves brand identity citation accuracy. ProjectFlow’s brand identity citations went from vague to specific and from 11% to 82% coverage almost entirely through Organization schema implementation. The knowsAbout array was critical — it directly mapped the brand to the specific product category terms that AI engines use to classify it for recommendation queries.
• Lesson 4: Measure citation rates, not just traffic or rankings. Traditional marketing metrics — organic traffic, keyword rankings, domain rating — showed no change during the 90-day program (they were not expected to). Only direct citation measurement revealed the dramatic improvement in AI search presence. Brands without a citation tracking system cannot assess their GEO performance — and cannot know whether their investments are working.
• Lesson 5: AI citation improvements drive measurable downstream business outcomes within 120 days. The 23% increase in direct brand search volume and 19 new AI-attributed trial signups at 120 days confirmed that AI citation performance translates to real business outcomes — not just abstract citation metrics. The referral traffic from Perplexity further confirmed that AI citations drive direct, trackable website visits.

The Replication Framework

Based on ProjectFlow’s experience, this replication framework applies to any B2B SaaS brand starting from zero or near-zero AI citation coverage:

Week 1 (First 5 business days)

• Run 35-query citation baseline across ChatGPT, Gemini, Perplexity, Google AI Overviews
• Implement Organization schema with full knowsAbout and sameAs on homepage
• Identify all pages with FAQ sections lacking FAQPage schema

Weeks 2-3

• Implement FAQPage schema on all pages with existing FAQ sections
• Fix Article schema dateModified on all stale posts (with substantive content updates)
• Submit all updated URLs to Google Search Console for indexing

Weeks 4-6

• Implement SoftwareApplication schema on all product pages
• Rewrite introductory paragraphs on 10 to 15 priority pages to answer-first format
• Create Wikidata entity with external ID cross-references

Weeks 7-12

• Add FAQ sections to 5 to 8 high-value pages without existing FAQ content
• Update all platform profiles (G2, Crunchbase, LinkedIn, Capterra) for entity consistency
• Launch G2 review generation campaign to active customer base
• Publish one original data piece relevant to your product category
• Measure citation rates at Day 45 and Day 90 against baseline

FAQs

How long does it take to see AI citation improvement after implementing GEO changes?

Schema changes (FAQPage schema, Organization schema) typically produce measurable citation improvement within 4 to 6 weeks — the time needed for AI crawlers to re-crawl, index the schema, and update citation selection systems. Content format changes (answer-first rewrites, FAQ additions) typically take 6 to 8 weeks. Entity signal changes (Wikidata creation, platform profile updates) take 8 to 12 weeks. In ProjectFlow’s case, the first measurable citation improvement appeared at 5 weeks (FAQPage schema taking effect on Perplexity), with the full 90-day result representing the cumulative effect of all interventions.

Is 40% citation coverage in 90 days a realistic expectation for all SaaS brands?

ProjectFlow’s 40% result at 90 days was achieved from a zero baseline with a strong pre-existing content library (47 indexed blog posts) and good domain authority (DR 52). Brands starting from zero with fewer than 20 indexed articles or lower domain authority may see slower initial progress — 20 to 30% citation coverage at 90 days is more realistic. Brands in highly competitive categories (CRM, marketing automation, security) with many established players may face a slower citation ramp-up as AI engines have more established citation sources to draw from. Brands in less competitive SaaS niches may achieve 40%+ faster than 90 days.

What was the total time investment for the 90-day program?

The 90-day program required approximately 120 to 130 hours of total work across the marketing team: schema implementation and validation (approximately 30 hours), content rewrites and FAQ additions (approximately 45 hours), entity and platform profile work (approximately 15 hours), original data research and publication (approximately 25 hours), and citation tracking and measurement (approximately 10 hours). This equates to approximately 1.5 full-time weeks of equivalent effort spread across 90 calendar days — or approximately 10 to 15 hours per week from the content marketer and 2 to 3 hours per week from the technical team member who implemented schema.

Which AI platform responded fastest to GEO improvements?

Perplexity responded fastest — citation improvement was first measurable on Perplexity at approximately 5 weeks after FAQPage schema implementation, and Perplexity reached the highest citation rate (60%) at Day 90. This is consistent with Perplexity’s real-time web crawl architecture, which updates citation selection faster than ChatGPT or Gemini (which have longer knowledge update cycles). ChatGPT Browse responded second-fastest, followed by Gemini, with Google AI Overviews showing the slowest but most sustained improvement pattern.

Apply These Lessons to Your Brand

ProjectFlow’s experience is not unique — the pattern of strong SEO performance combined with zero AI citation presence is common across B2B SaaS brands that have not yet implemented GEO. The interventions that produced ProjectFlow’s results are available to any SaaS brand: FAQPage schema, Organization schema, Wikidata entity creation, answer-first content, and platform profile optimization. The 90-day replication framework provides a starting point — adapt it to your specific content library, query set, and competitive landscape.

→ Run your free AI Visibility Audit at Onxeera — establish your citation baseline in under 5 minutes

Related Articles

• GEO for SaaS Brands
• AI Citation Tracking System
• GEO Schema Audit
• FAQ Schema for GEO
• GEO A/B Testing
• GEO ROI Measurement
• GEO Optimization: The Complete Guide