Across 212 AI visibility audits, ReSO found a consistent pattern: brand visibility declines as buyers move closer to a purchasing decision. Average share of voice falls from 23.4% on Discovery prompts to 15.7% on Pricing prompts. The average, however, hides a more important finding. Half of the audited brands receive zero visibility when buyers ask pricing-related questions.
The drop is not evenly distributed across the buyer journey. Some brands remain visible during category discovery, then disappear during comparisons, feature evaluation, persona-fit questions, or pricing. Measuring average visibility alone misses where those gaps occur.
Key findings
- AI visibility declines as buying intent increases: Across 212 audits, the average share of voice drops from 23.4% on Discovery prompts to 15.7% on Pricing prompts.
- Pricing is the biggest visibility blind spot: More than half of the audited brands receive zero visibility when buyers ask pricing-related questions.
- Average share of voice tells only part of the story: Measuring visibility by buyer-journey stage reveals the gaps that a single overall score cannot.
- The funnel drop varies by industry: SaaS and Services show clear declines, while the size of the drop differs significantly across verticals.
- AI engines behave differently across the funnel: ChatGPT, Perplexity, and Google AI Overviews each have stronger visibility at different buyer-journey stages.
- The key question is where your brand disappears: Knowing the stage where buyers stop seeing your brand is more valuable than knowing your overall share of voice.
The five buckets a buyer’s question passes through
Each prompt in the audit is classified into one of five buyer-journey buckets:
- Discovery: Category awareness queries such as “who’s in this category?” or “best category“.
- Comparisons: Active evaluation queries such as “X vs Y” or “top alternatives to X”.
- Feature: Feature fit queries such as “which category supports capability?”
- Persona: Segment fit queries such as “which category is best for role?”
- Pricing: Final consideration queries such as “how much does X cost?” or “X pricing”.
Each audit runs prompts across all five buckets, allowing visibility to be measured at every stage of the buyer journey. Together, these buckets map the progression from early discovery through to purchase consideration.
The funnel drops, bucket by bucket
Aggregate share of voice by bucket, across all 212 audits (mean and median reported separately because the distributions are skewed):
| Bucket | Mean share | Median share | Missed-prompt rate |
|---|---|---|---|
| Discovery | 23.4% | 18.35% | 57.7% |
| Comparisons | 21.0% | 6.15% | 66.7% |
| Feature | 20.7% | 11.60% | 58.1% |
| Persona | 17.4% | 12.50% | 67.7% |
| Pricing | 15.7% | 0.00% | 71.2% |
The table highlights two consistent patterns. First, the average share of voice declines steadily as buyers move from Discovery to Pricing. Second, the miss rate, the percentage of prompts where a brand is absent from all three AI engines, increases at each stage of the journey.
Together, these trends show that brands become less visible as buying intent increases. Pricing records the highest miss rate (71.2%) and the lowest median share of voice (0%), meaning the typical audited brand is not mentioned at all when buyers ask pricing-related questions. Comparisons and Persona sit between Discovery and Pricing, where visibility continues to decline but remains measurable for the median brand.
Looking only at the average share of voice hides these differences. Measuring visibility bucket by bucket reveals where brands disappear from AI-generated answers, making it easier to identify the stages of the buyer journey that need the most attention.
The median zero pricing finding, unpacked
The median-zero Pricing statistic is the sharpest finding in the dataset. It means the typical audited brand is not mentioned in AI answers to prompts such as “how much does [X] cost” or “[X] pricing.” The brand is not simply mentioned less often. It is absent from the pricing prompt set altogether.
Two factors help explain this pattern:
1. AI engines rely on different sources for pricing queries
Discovery prompts often surface category leaders and authoritative summaries. Pricing prompts, however, are more likely to cite:
- Pricing pages
- Comparison-table articles
- Vendor documentation
- Third-party pricing aggregators
- Review platforms with pricing sections
Owning a pricing page is a baseline requirement, but it is rarely the only surface AI engines rely on. Brands whose pricing information exists only on their own /pricing page compete for a citation surface largely controlled by third-party sources.
2. Pricing questions occur at the highest-intent stage of the buyer journey
Pricing prompts may be fewer than discovery prompts, but they sit closest to conversion. Zero visibility at this stage means a brand is absent when buyers are making their final evaluation, even if it appeared earlier in the journey.
The response to a median-zero Pricing finding is not simply to publish more pricing content. Instead, audit the third-party sources AI uses to answer pricing questions in your category and evaluate how your brand is represented across those sources. The ghost traffic metric framing is a useful complement, since much of the buying attention at the AI-answer surface never reaches a website in the first place.
Where the funnel drop is steepest
The bucket-share drop varies materially by vertical:
| SaaS | 30.5% | 21.9% | 8.6pp |
| Services | 10.2% | 2.7% | 7.5pp |
| E-commerce | 30.0% | 0.0% | 30.0pp |
| Other | 31.3% | 28.9% | 2.4pp |
| Manufacturing | 40.0% | 44.4% | inverted |
Two verticals stand out, while one shows an interesting exception:
- SaaS starts with relatively high Discovery visibility (30.5%) but drops by 8.6 percentage points at Pricing, making it the clearest example of the funnel-drop pattern across the dataset.
- Services follow a similar pattern, with visibility falling by 7.5 percentage points. However, the decline starts from a much lower base, dropping from 10.2% at Discovery to just 2.7% at Pricing.
The remaining verticals tell a different story. E-commerce records the steepest decline, falling from 30.0% at Discovery to 0.0% at Pricing. Because the Pricing result is based on a single audit, it should be treated as directional rather than conclusive. Manufacturing shows the opposite pattern, with Pricing visibility (44.4%) exceeding Discovery visibility (40.0%). Since this finding is based on a single audit (n=1), it should be viewed as an interesting observation rather than a broader trend.
The Other category records the smallest decline at 2.4 percentage points, highlighting that funnel-stage visibility varies across industries. The engine-level divergence analysis showed a similar pattern of variation across verticals using a different measurement approach.
The engines diverge most at pricing
The three AI engines do not distribute citations evenly across the buyer journey.
- ChatGPT receives the largest share of mentions at the Discovery and Feature stages (around 41% each).
- Perplexity is strongest for Comparisons, Persona, and Pricing (around 36–37% each).
- Google AI Overviews ranks third across most buckets but reaches its highest share at Pricing (30.2%), where the gap between the three engines is the smallest.
These differences have practical implications:
- Pricing is the stage where investing across all three engines is likely to deliver the broadest coverage.
- Discovery is the stage where a ChatGPT-first strategy is least likely to leave major visibility gaps.
- Engine priorities should be aligned with the stage of the buyer journey you are trying to improve.
The intent-driven citation shifts documented in the earlier piece describe the underlying source-class dynamics that produce this bucket-level pattern.
How to think about your own funnel stage visibility
The most useful thing an operator can do with this data is refuse the aggregate-visibility conversation and force a bucket-level one.
- Where in the funnel do you actually have zero visibility?
For most audited brands, the answer is Pricing, and often Persona. The first useful diagnostic is not “what is my average share of voice” but “which bucket am I invisible in.” If the median-brand-Pricing-visibility number in the aggregate is 0%, you cannot be confident your own brand is different until you check. ReSO’s Visibility Audit runs that check bucket by bucket across the three engines and returns the per-bucket picture directly.
- What third-party sources answer the queries you are invisible in?
Pricing prompts are usually answered by third-party aggregators, review-platform pricing sections, or category-comparison articles. Whatever surface AI reaches for at your absence stage is the surface to prioritise for coverage. Semrush’s content-optimization study covers the signal weights AI applies to different content classes, useful for prioritising which surfaces to target first.
- What does your buyer’s actual journey look like against these buckets?
If your buyer’s category deliberation clusters at Comparisons and Persona, the top-of-funnel Discovery share of voice is less strategically important than a mid-funnel operator would think. Prioritise the bucket that matches the deliberation stage where your buyer’s decision gets made.
- What is the miss-rate cost you are absorbing?
Miss rate is not the inverse of share of voice. It is the rate of prompts where you appear in zero engines. A brand with 15% Pricing share of voice and a 65% miss rate is not just being cited less often at Pricing. It is completely silent on two-thirds of pricing prompts. That silence is the more actionable number.
How the funnel-stage data was captured
The analysis is based on ReSO’s proprietary AI visibility audits across 212 brands conducted between March and June 2025. Each audit ran approximately 35 brand-related prompts against ChatGPT, Perplexity, and Google AI Overviews, distributed across the five buyer-journey buckets: roughly Discovery (8), Comparisons (5), Feature (12), Persona (6), and Pricing (3), with some variation by category.
Each response was reviewed for brand mentions, and every brand’s appearance across the three AI engines was recorded. Share of voice was calculated as the average visibility across prompts within each bucket, while miss rate measured the percentage of prompts where the audited brand was absent from all three engines.
Two scope notes are worth keeping in mind:
- Audits using older prompt taxonomies, such as “Evaluation” or “Comparison” (singular), were excluded from the aggregated results.
- Pricing results for E-commerce and Manufacturing are based on small sample sizes and should be treated as directional rather than conclusive.
Get your funnel-stage AI search visibility audit
A single share-of-voice score cannot tell you where your brand disappears in AI-generated answers. The real insight comes from measuring visibility at each stage of the buyer journey.
ReSO’s AI Visibility Audit tracks your brand across Discovery, Comparisons, Feature, Persona, and Pricing prompts in ChatGPT, Perplexity, and Google AI. It highlights the stages where your brand is missing, the third-party sources AI relies on instead, and the opportunities to strengthen your visibility when buying intent is highest.
Every audit includes a stage-by-stage visibility breakdown and a prioritised action plan tailored to your category, helping you focus on the gaps that matter most.
Request your AI Visibility Audit to see where your brand stands across the AI buying journey.
FAQs
Why is the median much lower than the mean for Pricing visibility?
A small number of brands have strong Pricing visibility, while more than half receive none at all. The median reflects the typical brand, making it a better indicator of overall performance.
Are Discovery and Pricing measured using the same prompts?
No. Each audit uses prompts tailored to the brand’s category, but every prompt is consistently classified into one of the five buyer-journey buckets.
Does the funnel drop happen across every industry?
The overall pattern is consistent, but the size of the drop varies by industry. SaaS and Services show clear declines, while smaller vertical samples should be treated as directional.
What does a 71.2% Pricing miss rate mean?
It means that 71.2% of Pricing prompts did not mention the audited brand in ChatGPT, Perplexity, or Google AI. A high miss rate indicates that buyers are unlikely to encounter the brand when asking pricing-related questions.
