From 0 to 50 AI Mentions/Month: A 6-Month Journey

Key Results

| Metric | Month 0 | Month 3 | Month 6 | Change | |---|---|---|---|---| | AI mentions/month | 0 | 14 | 53 | From invisible to 53 mentions | | AI Share of Voice | 0% | 4.2% | 11.8% | Top 3 in local category | | AI referral traffic | 0 visits | 87 visits | 410 visits/month | New acquisition channel | | AI Visibility Score | 8/100 | 42/100 | 68/100 | +750% improvement |

Company Overview

KsiEgowi Pro (name changed for confidentiality) is a regional accounting and tax advisory firm based in Krakow, Poland, serving 340 small and mid-sized business clients. With a team of 18 accountants and tax advisors, the firm had operated successfully for 12 years, building its client base almost entirely through referrals and local Google search.

In September 2025, the firm's managing partner heard from three separate prospective clients that they had "asked ChatGPT for an accountant recommendation in Krakow" before making contact. None of them had found KsiEgowi Pro through AI. They had found the firm through Google afterward -- but the initial discovery channel was AI.

When the team tested this themselves, asking ChatGPT, Gemini, and Perplexity to recommend accounting firms in Krakow, KsiEgowi Pro appeared in zero responses. The firm was completely invisible to AI despite being well-established locally and ranking on page 1 of Google for "biuro rachunkowe Krakow" (accounting firm Krakow).

This prompted a decision to invest in AI SEO from scratch -- a 6-month commitment to becoming visible in AI recommendations. This case study documents every month of that journey.

The Challenge

The initial audit (AI Visibility Score: 8/100) revealed a starting position typical of small service businesses:

Technical barriers. robots.txt blocked all bots except Googlebot. The site loaded in 3.2 seconds (well above the 0.4-second threshold where AI citation rates peak). No llms.txt file existed.

Zero structured data. No schema markup of any kind. AI models had no structured way to understand what the firm did, where it was located, or what services it offered.

Thin, unstructured content. The website had 8 service pages and 4 blog posts, all written in a promotional tone without specific, citable information. Pages averaged 200 words -- far too short for AI to consider as an authoritative source.

No third-party signals. No Google Business Profile optimization for AI, zero industry directory presence, no reviews on platforms AI models reference, and no mentions on forums or discussion sites.

No monitoring. The firm had no way to track whether or when AI models mentioned them. They would need to set up AI visibility monitoring from the ground up.

Month-by-Month Implementation

Month 1: Technical Foundation

Week 1-2: Fix technical access.

• Reconfigured robots.txt to allow AI search crawlers (OAI-SearchBot, ChatGPT-User, PerplexityBot) while blocking training bots
• Implemented server-side caching that reduced page load time from 3.2s to 0.6s
• Created an llms.txt file describing the firm's services and expertise areas

Week 3-4: Deploy foundational schema.

• Added Organization schema site-wide with business name, address, founding year, service area, and social profiles
• Added LocalBusiness schema with geo-coordinates, opening hours, and service categories
• Implemented BreadcrumbList schema for navigation context

Month 1 result: AI Visibility Score rose from 8 to 19. No mentions yet, but the foundation was set. AI crawlers could now access and understand the basic entity.

Month 2: Content Foundation

Following the 90-day AI SEO roadmap, Month 2 focused entirely on building a content base that AI models would consider worth citing:

Content created:

• 6 comprehensive service pages (800-1,200 words each) covering: tax advisory, bookkeeping, payroll, audit preparation, startup accounting, and international tax
• Each page followed BLUF format with a clear service description in the first paragraph
• Every page included an FAQ section (5-7 questions) with FAQPage schema
• All content was written in a factual, educational tone rather than promotional

Key content principle: Instead of "We offer the best tax advisory services in Krakow," the team wrote content like "Small businesses in Krakow typically need tax advisory support for three key areas: quarterly VAT filings, annual CIT declarations, and transfer pricing documentation for companies with international transactions."

The second version is what AI models cite -- specific, factual, educational content that answers real questions.

Month 2 result: AI Visibility Score rose to 31. First AI mention detected: Perplexity cited the firm's startup accounting page when a user asked about "setting up a company in Krakow."

Month 3: Content Expansion and Local Signals

Content expansion:

• Published 4 blog articles targeting common questions: "How to register a company in Poland," "VAT rates in Poland 2026," "Accounting deadlines for Polish businesses," and "How to choose an accounting firm in Krakow"
• Each article was 1,500-2,000 words with BLUF formatting, FAQ sections, and Article schema
• Content focused on information gain -- original data from the firm's experience with 340 clients

Local signals:

• Optimized Google Business Profile with complete category information, service descriptions, and regular posts
• Created profiles on 5 industry directories with consistent NAP (Name, Address, Phone) data
• Solicited 12 Google reviews from satisfied clients, focusing on detailed reviews that mentioned specific services

Month 3 result: AI Visibility Score reached 42. AI mentions hit 14/month. AI Share of Voice for "accounting firms in Krakow" queries reached 4.2% -- the firm appeared in roughly 1 in 25 AI responses for relevant queries.

Month 4: Third-Party Authority

Month 4 focused on building the external signals that AI models weight heavily when making recommendations:

Forum engagement:

• The managing partner began answering tax and accounting questions on Polish business forums (JustJoinIT, Wykop, and dedicated accounting groups on Facebook)
• Average of 8 substantive responses per week, providing genuine expertise without promotional content
• Within 4 weeks, other users began recommending KsiEgowi Pro in response to "accounting firm recommendation" threads

Review acceleration:

• Launched a systematic review collection program
• Grew from 12 to 38 Google reviews (average 4.8 stars)
• Secured 6 reviews on Clutch.co (relevant for B2B service firms)

Content partnerships:

• Contributed a guest article to a Polish startup publication about tax optimization for new businesses
• The firm was cited as a source in 2 local business articles

Month 4 result: AI Visibility Score reached 52. Mentions hit 28/month. ChatGPT started including the firm in responses to Polish-language accounting queries.

Month 5: Optimization and Expansion

Content refinement:

• Updated all 10 service/blog pages with fresh data from Q1 2026
• Added comparison content: "In-house accountant vs accounting firm for Polish SMEs"
• Created a comprehensive "Accounting costs in Poland" page with specific pricing ranges -- exactly the kind of concrete data AI models prefer to cite

Schema expansion:

• Added Person schema for the managing partner and two senior accountants, establishing individual expert entities
• Linked Person schema to published articles (author attribution)
• Added Review schema pulling in Google Business Profile ratings

Month 5 result: AI Visibility Score reached 61. Mentions hit 41/month. The firm's managing partner was now mentioned by name in some AI responses, not just the firm -- a sign that individual entity recognition was working.

Month 6: Consolidation and Measurement

Activities:

• Continued weekly content updates and forum engagement
• Published 2 new articles on emerging tax topics (global minimum tax impact on Polish businesses, cryptocurrency tax reporting)
• Launched AI visibility monitoring dashboard with weekly automated reports
• Conducted full before/after analysis

Month 6 result: AI Visibility Score reached 68. Mentions hit 53/month. AI Share of Voice for local accounting queries reached 11.8%, making KsiEgowi Pro one of the top 3 most-mentioned firms in Krakow for accounting-related AI queries.

Cumulative Results

The 6-month scorecard

| Metric | Month 0 | Month 1 | Month 2 | Month 3 | Month 4 | Month 5 | Month 6 | |---|---|---|---|---|---|---|---| | AI Score | 8 | 19 | 31 | 42 | 52 | 61 | 68 | | AI mentions/month | 0 | 0 | 3 | 14 | 28 | 41 | 53 | | AI Share of Voice | 0% | 0% | 0.8% | 4.2% | 6.7% | 9.1% | 11.8% | | AI referral visits/month | 0 | 0 | 12 | 87 | 195 | 310 | 410 |

Business impact

The AI visibility investment translated into concrete business results:

• 17 new clients acquired through AI referrals over the 6-month period (clients who reported finding the firm through an AI assistant)
• Average client lifetime value of 2,400 EUR/year, representing 40,800 EUR in annual recurring revenue from the AI channel alone
• Cost per acquisition from AI was approximately 60% lower than from Google Ads, making it the most efficient new client acquisition channel
• Client quality was higher -- AI-referred clients arrived with specific questions about services they had already read about, reducing the average consultation-to-close time by 35%

Platform breakdown

| Platform | Month 6 Mentions | Share | Trend | |---|---|---|---| | ChatGPT | 22/month | 41.5% | Fastest growing | | Perplexity | 16/month | 30.2% | Most referral traffic per mention | | Gemini | 11/month | 20.8% | Steady growth | | Copilot | 4/month | 7.5% | Emerging |

Key Takeaways

1. Starting from zero is an advantage, not a limitation. Low competition in local AI search means that even basic optimization produces visible results quickly. The first mention came within 8 weeks. For a complete framework, see AI SEO strategy from scratch.

2. Follow the 90-day roadmap. Months 1-3 followed a structured sequence: technical access, then schema, then content, then third-party signals. Skipping steps or doing them out of order reduces effectiveness. Our 90-day AI SEO roadmap details the exact sequence.

3. Educational content beats promotional content. AI models cite factual, specific information -- not marketing copy. "VAT rates in Poland for 2026" gets cited; "We are the best accounting firm" does not.

4. Third-party signals create the tipping point. Months 4-5 saw the steepest growth because third-party signals (reviews, forum mentions, directory listings) compound with on-site optimization. AI models trust recommendations that are corroborated across multiple sources.

5. Measure weekly with AI Share of Voice. Monthly AI mention counts are the headline metric, but Share of Voice -- your percentage of AI responses in your category -- is the metric that shows competitive position. The firm went from 0% to 11.8% in 6 months.