Product system
Start with the workflow agencies can sell: prove which brand AI prefers, isolate LLM Perception by target, track Share of Voice by market, and convert Brand Hub knowledge-graph findings into fixable client actions.
This page explains the new agency workflow: AI Preference for recommendation wins, ultra-targeted LLM Perception for buyer context, target-aware Share of Voice, and Brand Hub actions from the knowledge graph.
Agency workflow
01 / AI Preference + LLM Perception
VectorGap runs targeted and ultra-targeted checks across major AI providers so agencies can see whether a client is preferred, accurately described, recommended, or displaced by competitors in a real buyer context.
02 / Web perception
The web perception layer reviews signals from search results, third-party pages, review surfaces, and social mentions so the team can see which sources reinforce or weaken AI trust.
03 / Brand Hub actions
Brand Hub actions translate entity gaps, missing proof, weak source coverage, and unclear topical clusters into practical content, source, and SEO/GEO remediation work.
Why the features exist
Agency Unlimited gives SEO agencies recurring AI Preference, LLM Perception, Share of Voice, Brand Hub actions, reporting, and client proof for ongoing GEO delivery.
Run a client baseline across buying, comparison, and category prompts.
Segment LLM Perception and AI Preference by persona, local market, prompt language, and competitor set.
Use target-aware Share of Voice to show where competitors win mentions, citations, or recommendations.
Map the loss to Brand Hub actions: missing facts, weak sources, entity confusion, technical extractability, or thin proof pages.
Ship the next fixes and show trend movement in repeat audits.
New agency advantage
This is the commercial wedge for agencies: not a generic AI visibility score, but a target-specific AI-market diagnosis. You can show how AI Preference, LLM Perception, and Share of Voice change for a CEO in Belgium, a procurement buyer in France, or an English-speaking analyst comparing the same client.
AI Preference
Tie it to the selected target, not a generic global average.
Targeted LLM Perception
Tie it to the selected target, not a generic global average.
Share of Voice
Tie it to the selected target, not a generic global average.
Feature map
Each module answers a specific agency question: does AI prefer the client, how does each LLM describe them, what share of the answer do they own, and which Brand Hub action should the client fix first?
See which brand AI prefers in a targeted buying context, why it wins, and what evidence would change the answer.
Check whether your content is easy for AI systems to cite and reuse.
Track visibility against competitors for the selected market, language, persona, and audit target.
Turn brand-knowledge graph findings into source, entity, proof, and content actions that improve LLM visibility and SEO stature.
Review how major providers describe the brand for a specific buyer, market, prompt language, and competitor set.
Compare positioning, recommendation behavior, and narrative gaps.
Spot bad facts, stale claims, and mismatches against your source of truth.
Share progress with stakeholders using audit history and trend summaries.
VectorGap is not a bag of disconnected feature cards. The practical agency sequence is: diagnose AI Preference and LLM Perception, benchmark Share of Voice for the selected target, convert Brand Hub graph findings into actions, then repeat the audit to prove movement.
Run AI Preference and LLM Perception checks to find recommendation, accuracy, sentiment, and competitor gaps.
Use Brand Hub actions to separate source weakness, entity confusion, public truth gaps, extractability issues, and competitor narrative pressure.
Use target-aware Share of Voice, history, trend summaries, and exportable reports to show the client what changed after the fix cycle.
The founding cohort is €445/mo before the standard €997/mo rate returns. Start now if your agency is ready to package AI visibility diagnostics into client delivery.
