Orbitas Digital, a 12-person web agency based in Ghent offering web design, UX, and digital strategy to Belgian SMEs, had spent three years building a solid reputation through word of mouth and a modest SEO presence. Their pipeline was healthy but unpredictable. In early 2026, a prospective client mentioned they had found Orbitas by asking ChatGPT to recommend a web agency in Belgium that specialized in service businesses. This surprised the founders: they had never consciously optimized for AI visibility.
When they tested the query themselves, Orbitas appeared in only 1 of 12 variations of the question. Four competitors appeared in 8 or more. The gap between their actual reputation and their AI presence was stark. The agency decided to close it systematically rather than rely on serendipity. They engaged AISOS for a focused 75-day implementation. For context on the landscape, see our guide to AI visibility for digital agencies and our overview of what AI visibility means in practice.
This case study documents the approach taken, the specific challenges of positioning a service business in AI responses, and the concrete results achieved within the engagement window. It is directly applicable to any professional services agency operating in Belgium or the broader Benelux market. For city-level context, see also our Ghent AI visibility overview and our page on working with AISOS.
The Challenge
Orbitas Digital faced a challenge common to service businesses: AI assistants find it harder to recommend agencies than products, because agency quality is contextual, relationship-dependent, and harder to verify from public data. When a user asks ChatGPT for a web agency in Belgium, the model draws on its training data and, for browsing-enabled models, real-time web access. Agencies that appear in structured directories, have machine-readable case studies, and are mentioned in credible third-party sources get recommended. Agencies with beautiful websites but thin public presence do not.
Orbitas had a well-designed website and positive client testimonials, but almost nothing that AI systems could use to evaluate and recommend them. No schema markup beyond basic Organization data. No structured case studies in machine-readable format. No presence in the B2B agency directories that LLMs regularly sample. No llms.txt file to guide model crawlers. Their Google Business Profile was minimal. Understanding Answer Engine Optimization for service businesses was the conceptual starting point for the engagement.
Competitive analysis revealed that the agencies consistently recommended by AI tools had three things in common: detailed, factual case studies published in HTML with schema markup; presence in at least two industry directories specifically indexed by AI crawlers; and a clear, consistent service definition that matched the way prospective clients phrased their queries. Orbitas had none of these three. The implementation plan addressed each one in sequence.
The AISOS Strategy
The engagement opened with an AI visibility audit across ChatGPT, Perplexity, Claude, and Gemini using 42 queries relevant to Orbitas' target clients: SME owners and marketing managers in Ghent, Brussels, and Antwerp looking for web design or digital strategy support. Orbitas appeared in 4 of the 42 responses (9.5%), always as a brief mention without a recommendation context. The top three competitors appeared between 28 and 35 times each.
Phase one addressed the content foundation. AISOS worked with Orbitas to publish six structured case studies, each following a format purpose-built for AI readability: a clear client profile in the opening paragraph, a specific problem statement, a documented methodology, and quantified outcomes. These were published as standalone HTML pages with Service, LocalBusiness, and CaseStudy schema markup. Every case study linked to the relevant industry page and included geographic signals matching the client location.
Phase two targeted signal injection. AISOS placed Orbitas in four high-authority Belgian business directories and two pan-European agency listings known to be sampled by AI platforms. A Wikidata entity record was created for the agency. An llms.txt file was deployed providing model crawlers with a structured summary of services, target clients, and geographic scope. Phase three optimized the Google Business Profile with structured service categories and a complete Q&A section answering the top 15 questions AI assistants receive about web agencies in Belgium. Internal links from these case studies pointed to the resources section and to relevant glossary terms to reinforce topical authority.
The Results
By day 75, Orbitas' AI mention rate on the original 42-query audit had increased from 9.5% (4 mentions) to 61% (26 mentions). The improvement was strongest on Perplexity, where the agency went from 1 to 11 appearances out of 14 relevant queries. ChatGPT with browsing showed 9 of 14, up from 2. Claude showed 6 of 14, up from 1. Gemini, reflecting its slower corpus update cycle, showed 4 of 14, up from 0.
The pipeline impact was concrete. Over the 75-day period following the first content publication, Orbitas received 28 inbound inquiries where the prospective client mentioned discovering them through an AI assistant. Of these, 11 converted to discovery calls. Seven progressed to proposal stage. Three signed contracts within the engagement window, with a combined project value of 47,000 euros. This return exceeded the full cost of the AISOS engagement more than four times over within the first quarter.
A secondary benefit emerged from the structured case studies: they began ranking organically for long-tail queries that had never previously generated traffic. The page targeting "web agency for medical practices Ghent" reached position 4 in Google within 45 days of publication. The dual benefit of AI visibility and traditional SEO improvement from the same content asset is consistent with what AISOS observes across service business engagements in Belgium and the Netherlands.
Key Success Factors
The structured case study format was the single highest-leverage action in this engagement. Orbitas had compelling client stories but had never documented them in a format that AI systems could process. Once published in structured HTML with rich schema markup, these pages became the primary source AI models used to evaluate and recommend the agency. The lesson for any professional services firm is that undocumented reputation is invisible reputation in the AI layer.
Geographic signal precision mattered significantly. Belgian AI queries often include city names or regional references. Content that explicitly mentioned Ghent, Brussels, and Antwerp in a meaningful, contextual way, rather than as a footer address, performed notably better in location-specific queries. This aligns with what AISOS documents in its Brussels AI visibility guide and reflects the broader principle that geographic entities must be woven into content to function as AI signals.
The founders' decision to participate actively in the engagement rather than delegate it entirely also contributed to outcome quality. They reviewed and approved case study drafts within 24 hours, provided specific client outcome data, and made schema deployment decisions quickly. Agencies that treat AI visibility as a passive vendor deliverable rather than a strategic initiative consistently underperform those where leadership is directly engaged. See our contact page to discuss how engagement structure affects outcomes for your specific situation.
Lessons Learned
The most important lesson from the Orbitas engagement is that for service businesses, AI visibility is primarily a trust signal problem, not a content volume problem. The agency did not need more content. It needed the right content in the right format, verifiable by AI systems through third-party corroboration. One well-structured case study with real metrics and a named client (with permission) outperforms ten generic service pages with no factual specificity.
The second lesson concerns the timing of results. Perplexity, which queries the live web in real time, responded to the case study publication within two weeks. ChatGPT's browsing mode followed within three to four weeks. Base model responses (the version of ChatGPT without live web access) reflected the changes more slowly, as these depend on training data updates. Agencies planning an AI visibility implementation should set expectations accordingly: early results are real and measurable, but the full effect accumulates over three to six months.
Finally, the engagement demonstrated that the Belgian market is still early in AI visibility adoption among professional services firms. The gap between Orbitas and its top competitors was large at the start, but those competitors were themselves still far from saturation. First movers in any local professional services market gain a compounding advantage that becomes harder for competitors to overcome as AI models reinforce their recommendation patterns over time. The window for building this advantage in Belgium is open now. Contact AISOS to assess your current position before your main competitors do.