The AI Buyer Journey Matrix.

Most AI visibility tools report one blended score. Monroya separates LLM visibility into the three stages buyers actually move through — Discovery, Evaluation, and Decision — so you can tell whether you have a discoverability problem, a comparison problem, or a vendor-selection problem. The fix is different for each, and only a stage-by-stage view shows which one is costing you pipeline.

A fourth category, Retention / churn risk, sits outside the funnel and is tracked for every account: what AI tells your existing customers when they search for alternatives, complaints, or a way to cancel.

Three numbers from Monroya's aggregated scans across ChatGPT, Claude, Gemini, and Perplexity explain why stage segmentation matters:

• Average discovery-stage inclusion is 7.6%, versus 37.7% at decision stage — brands are roughly 5× less visible when buyers are still framing the problem than when they're comparing named vendors.
• Evaluation-stage inclusion averages 23.3%, sitting much closer to discovery than decision — the middle of the funnel is where most brands quietly lose the buyer.
• 78.9% of the brands we track have never been cited in a tracked answer: they get mentioned, but no source link points back to their site. Citation gaps are far more common than mention gaps.

Aggregated across Monroya customer scans, ChatGPT / Claude / Gemini / Perplexity, 90-day rolling window. Refreshed with each scan cycle.

• AI Buyer Journey Intelligence is the measurement of how a brand appears across the full sequence of questions a buyer asks a generative AI assistant, from first problem framing through final vendor selection, segmented by funnel stage and persona.
• The AI Buyer Journey Matrix is the canonical instrument: journey stages on one axis, buyer personas on the other.
• Each cell reports three things: its own AI Share of Voice, its top competing brand, and the specific prompts driving the result across ChatGPT, Claude, Gemini, and Perplexity.
• Why one blended score misleads: a 42% average can mean dominance in discovery and invisibility in decision (a deal-loss pattern), or the reverse (a brand-strength pattern with a discoverability gap). The fix differs for each, and only a stage-by-persona view shows which one you have.

• Discovery.
  • Prompt type: problem framing, before the buyer knows a category exists.
  • Example: "why are our AI answers inconsistent across providers".
  • Winning move: category-defining content on third-party sources AI assistants pull from.
• Evaluation.
  • Prompt type: options and approaches, once the category is understood.
  • Example: "best tools to monitor brand visibility in ChatGPT".
  • Winning move: comparison and alternative pages, plus structured feature breakdowns on your own domain.
• Decision.
  • Prompt type: vendor selection and head-to-head.
  • Example: "monroya vs otterly pricing".
  • Winning move: fact sheets, reviewer evidence, current pricing, direct competitor pages.
• Retention / churn risk.
  • Tracked separately because the asker is already your customer.
  • Examples: "better alternative to monroya", "why are people leaving monroya", "how do I cancel monroya".
  • Winning move: make sure AI does not hand your renewal to a competitor. Every account gets these prompts tracked automatically.

• G2's 2026 AI Search Insight Report describes AI search as a third compression of the B2B software buying journey, where buyers act on a synthesized answer rather than a ranked list of vendors. G2, The Answer Economy (2026)
• An independent 2026 audit of 3,200 queries across ChatGPT, Perplexity, and Google AI Overviews puts AI citation share at roughly 17% of branded discovery for B2B SaaS, up from about 4% a year earlier, and finds cited brands are not simply the highest-ranked ones. The 2026 State of AI Search
• Exposure Ninja's 2026 B2B and B2C buyer-journey report tracks AI assistant usage across separate stages of the purchase journey rather than as a single funnel-wide behaviour. AI Search Buyer Journey, 2026

External sources are cited for market context only. Every stage figure on this page comes from Monroya's own scan data.

Generative engine optimization does not move on an SEO clock. Once you ship a fix, here is the sequence we observe across Monroya customer scans:

• Days 0-3: crawl and re-index. Assistants with live retrieval hit the changed page first. Nothing visible changes in answers yet; you are waiting on the crawl.
• Days 3-7: technical changes surface. Technical indexing changes often reflect in AI responses within 3-7 days after schema deployment — structured data, FAQ and product markup, canonical and robots fixes are the fastest-moving levers because they change how the page is parsed, not how authoritative it is judged to be.
• Weeks 2-6: early visibility indicators. The first durable movement shows up as mention rate and citation rate on discovery and evaluation prompts. Expect direction, not a finished number: a stage moving from 0% to single digits is the signal to keep going.
• Weeks 6-12: decision-stage and share-of-voice shift. Comparison, alternatives, and vendor-selection prompts are last to move because they depend on third-party sources and reviewer evidence being recrawled, not just your own domain.

Timings reflect Monroya's own re-scan data on customer sites after a tracked fix ships. Sites with slower crawl budgets or blocked AI user agents sit at the long end of every band.

We generate a draft prompt set from your category and ICP, organized by stage and persona. You review and edit. Prompts are tagged automatically, then scanned daily across every major AI assistant so each cell is auditable back to the raw responses. Each cell carries its own share of voice, its own top competitor, and its own opportunity queue.

Three patterns recur. If you're weak in discovery, invest in category-defining content on third-party sources where AI answers pull from. If you're weak in evaluation, ship comparison and alternative pages on your own domain. If you're weak in decision, publish structured fact sheets, reviewer evidence, and direct competitor pages with current pricing.

Most AI visibility tools report one number. We separated the journey because operators kept telling us the average was useless — they needed to know which specific stage was bleeding pipeline. The matrix is the first thing every Monroya customer opens.

• Evaluation stage first: what 19 tracked companies show
• How do I know if AI is sending customers to my competitors?
• How long until a new company shows up in search?