How we took IT law firm to #1 in AI answers across 3 service clusters in 3 months

Most LLMO engagements concentrate on a single niche. The firm needed visibility across three at the same time: IT lawyers and Diia.City residency in Ukraine, and cross-border IT contracts in Romania. The three clusters were at very different stages.
The IT lawyers cluster was already close to full visibility, Diia.City was visible but sitting in second place behind a direct competitor, and the Romania cluster was barely surfacing in English at all. Treating all three the same would have wasted effort on the cluster that needed it least.

We measured each cluster separately and treated the program as a portfolio, not a single push:

– Set up tracking for 34 commercial-intent prompts split across the three clusters (11 IT lawyers, 13 Diia.City, 10 cross-border Romania)
– Established a per-cluster baseline for every KPI, so each cluster’s progress could be judged on its own curve
– Reallocated effort month to month: as the IT lawyers cluster hit and held 100% visibility, support shifted toward the two clusters with more room to grow
– Flagged the Romania cluster early as the weakest link, with a clear root cause — not enough authoritative English-language content for AI systems to draw on

Result: All three clusters moved up together, and by May, the firm held #1 LLM position in each of them. The total trajectory — 159 → 472 → 1,060 mentions — was driven by deliberate reallocation, not by spreading effort evenly.

The firm had relevant service and landing pages, but the backlink and third-party trust signals behind them were thin, especially in the Romania cluster, where almost nothing existed in English. Structural AI-readiness work alone wasn’t going to move authority without external signals that LLMs already trust.

We ran link-building in-house to control quality and target it cluster by cluster:

– Focused placements on the priority service pages in each of the three clusters
– Prioritized listicle-format placements, which produce both faster SEO returns and the strongest AI-citation signals
– Conducted outreach to existing high-ranking listicles with proposals to add the firm vetting each host on two criteria: 1) it already ranks in Google for the target queries, and 2) it already appears as a cited source in AI answers for related prompts
– Concentrated English-language placements on the Romania cluster to close the gap AI systems were drawing attention to

Result: Mentions climbed sharply across all three clusters — IT lawyers 64 → 431, Diia.City 66 → 502, Romania 29 → 127 — with the same listicle placements feeding both SEO and AI citation over time.

Some of the comany’s most important Diia.City content lived inside accordion / tab blocks that only rendered when a user clicked, meaning the underlying detail was loaded via JavaScript. Most AI crawlers don’t execute JS, so that content was effectively invisible to them, even though it was perfectly readable for a human visitor.

We rebuilt the affected content with both Google and LLM extraction in mind:

– Converted the JS-rendered accordion content on the Diia.City pages into static HTML, so AI crawlers could actually read it
– Restructured priority content using AI-ready patterns: answer-first paragraphs, defined entity references, structured headings
– Optimized existing pages across the three clusters rather than rewriting from scratch, preserving the rankings already in place
– Flagged the constraint to the client’s dev team as a standing rule: important content must live in HTML, not load via JavaScript

Result: The Diia.City cluster, previously held back by hidden content, gained a full prompt in visibility (10/13 → 11/13) and improved its average position from 1.73 to 1.51 — the clearest single jump in the program.

A crawlability check found that several AI crawlers — including GPTBot, OAI-SearchBot, PerplexityBot, ImagesoftBot, and Bytespider — weren’t explicitly allowed in the site’s robots.txt. Even with strong content, a brand can’t be recommended by an AI system that isn’t permitted to read the site.

We brought the robots.txt in line with what AI systems need to read and cite the site:

– Explicitly allowed the major AI crawlers (OpenAI’s GPTBot and OAI-SearchBot, Anthropic’s ClaudeBot, PerplexityBot, ImagesiftBot, Bytespider)
– Added an llm.txt allowance and confirmed the sitemap reference
– Re-ran the crawlability check the following month to confirm the bots were being admitted

Result: With access opened up, more AI systems could read and reference the site, which broadened where the law firm could be picked up across ChatGPT, Perplexity, Gemini, and Claude.