Buyers are Googling less.
They're asking AI instead.
ChatGPT, Claude, Gemini, and Perplexity now answer with a shortlist instead of a search results page. That shortlist is becoming the new front door to every category.
ChatGPT
Perplexity
Claude
Gemini
Google AIThe Problem
Excellent companies get left off the list. Not because they are worse.
Ask an AI system for recommendations and it answers in seconds, with two or three names. That short list is the consideration set. Most buyers never look further.
Appearing on it is not about having the best product. It is about being legible: clear positioning, recognizable references, language the AI can confidently repeat.
A company can have happy customers and a better product and still be missing. Most founders only notice once a deal mentions a competitor they did not expect.
No second page
There is no scrolling, no page two. If a company is not in the opening answer, it is not in the conversation.
Category default
Once AI maps a company to a category, that description repeats across related questions. Early clarity compounds.
Invisible exclusion
The risk is not low visibility. It is being excluded before a buyer ever reaches out, which is hard to notice from outside.
The Origin
Srinivas Siddhart
Founder, Category Intelligence
Srinivas kept hearing the same thing from founders: a weaker competitor kept showing up first when buyers asked ChatGPT for recommendations. Nobody could explain why.
That question became two years of running the same queries across ChatGPT, Gemini, Claude, and Perplexity, watching which companies appeared and what separated them. A pattern emerged: the companies that appeared consistently were not always the strongest products. They were the ones AI could describe with confidence.
That research became a methodology, and the methodology became Category Intelligence: a practice built to show any company exactly how AI systems describe them, and why.
The investigation
Two-plus years tracking how ChatGPT, Gemini, Claude, and Perplexity actually form recommendations.
The research
Hundreds of logged queries, comparing what AI systems said against what was actually true.
The methodology
A repeatable way to trace a company's AI description back to its actual cause.
The Pattern
Not random.
Learnable.
A competitor outranking you is rarely about product quality. It is about signal clarity, and that signal is buildable. The companies that learn to build it become the default recommendation in their category.
Category positioning
How consistently a company is described as a category leader across external sources. Mixed signals read as ambiguous to AI systems.
Third-party references
Coverage in editorial publications, analyst reports, and review platforms. A few well-linked mentions often outweigh better documentation alone.
Comparison content
Review pages and alternatives articles that place a company against others. Appearing in these explicitly improves consistency.
Market signals
Case studies and customer names that signal which market a company serves, regardless of the actual customer mix.
Positioning consistency
Whether a company's own language matches what is said about it externally. When they diverge, AI defaults to the external version.
The Four Systems
ChatGPT
Patterns emerge from what it cites most: indexed web content, brand mentions, and reviews. Volume reveals who appears reliably.
Key signals: Web authority · Brand mentions · Category keywords
Gemini
G2 and Capterra reviews, Search rankings, and structured business data appear to weigh heavily on results.
Key signals: Search signals · Review platforms · Knowledge Graph
Claude
More conservative about naming companies, but when it does, results reflect positioning clarity over raw mention volume.
Key signals: Editorial coverage · Positioning clarity · Documentation
Perplexity
Cites sources directly, the most transparent of the four. Source authority and recency matter most here.
Key signals: Citation quality · Recency · Source diversity
Outputs vary across sessions, so a single prompt tells you little. Patterns only become visible at volume.
The Evidence
The first product
built from this research.
Below: an annotated AI response, a category recommendation map, and an anonymized case. Your report covers your company and category specifically.
Annotated AI response
Query
What are the best AI-powered sales intelligence tools for early-stage B2B startups in 2025?
Response
For early-stage B2B startups, the most recommended AI sales intelligence tools are:
Sources: G2, TechCrunch, Founder testimonials (2025)D
Category ownership
Apollo owns this use case in AI results. First, consistent, across all four systems. Your report shows what is driving that and where you sit relative to it.
Qualifier language
Notice how Clay is qualified immediately: 'steep learning curve,' 'technical teams.' That shrinks the audience who would consider them. Your report shows what qualifiers AI attaches to you.
Feature reduction
Instantly is reduced to cold email. One feature, no depth. This happens when AI does not have enough signal to say more. Your report tells you if this is happening to you.
Source influence
G2, TechCrunch, 'Founder testimonials.' These are the sources the AI drew from. Your report maps which sources are shaping what is being said about you.
Category recommendation map
Illustrative. Your report maps 4 systems across 20+ queries in your specific category.
Case narrative (anonymized)
What they expected
A workflow automation company in the revenue operations space. Active content program, strong G2 reviews, recent funding announcement. They assumed they would appear alongside two established players in their category. They expected Claude and ChatGPT to reflect their enterprise positioning.
What the report found
Across 24 queries tested, the company appeared in two responses, both on Perplexity, both in passing. On ChatGPT and Gemini, three competitors were named consistently. Claude returned the company once, with a description that reflected their 2022 positioning, not their current one.
Key findings
Competitors surfaced
Three named consistently across ChatGPT and Gemini. One named on Perplexity and Claude only.
Primary gap
No analyst coverage. No third-party editorial links. The company appeared in G2 reviews but these were not cited by any system tested.
Positioning discrepancy
Internal language described the product as enterprise-grade. External coverage described it as a tool for small teams. AI systems reflected external coverage.
Near-term action
One Forbes Councils contributor article and two updated G2 review responses. Both were indexed within four weeks. Mention frequency on Perplexity increased.
Illustrative of report structure. Actual client findings are confidential. All names and data points are anonymized.
Signal vs. Noise
A few prompts
won't show you
this.
Anyone can ask ChatGPT what it recommends. The hard part was never running the prompt. It is knowing what the answer means and what would change it.
1
Query you'd run yourself, one system, one moment in time.
20–30 × 4
Queries run across all four AI systems for your report, so patterns replace guesswork.
Asking AI yourself
One query. One response. No baseline.
The report
20 to 30 queries, four systems. Consistent patterns separated from query-level variance.
Asking AI yourself
You see who appears. Not why.
The report
Competitor signals mapped: the specific sources, language, and references driving their recommendations.
Asking AI yourself
No competitive context. No comparison.
The report
Side-by-side analysis of what each competitor has indexed that you do not.
Asking AI yourself
No way to tell signal from noise.
The report
Pattern recognition across systems to identify what is consistent versus incidental.
Asking AI yourself
No path from output to action.
The report
Findings translated into specific, prioritized changes in positioning, PR, and category language.
Asking AI yourself
The report
One query. One response. No baseline.
20 to 30 queries, four systems. Consistent patterns separated from query-level variance.
You see who appears. Not why.
Competitor signals mapped: the specific sources, language, and references driving their recommendations.
No competitive context. No comparison.
Side-by-side analysis of what each competitor has indexed that you do not.
No way to tell signal from noise.
Pattern recognition across systems to identify what is consistent versus incidental.
No path from output to action.
Findings translated into specific, prioritized changes in positioning, PR, and category language.
Pattern recognition
One query produces one answer. Running 20 to 30 across four systems reveals what is consistent, which is the actual signal.
Competitor analysis
Knowing a competitor appears first is a data point. Knowing why, their sources, their language, their coverage, is the analysis.
Source attribution
AI systems do not explain their sources. Identifying the four or five shaping your description takes cross-referencing citation patterns.
Category interpretation
Knowing you appear less does not tell you what to change. Turning a pattern into a specific action takes judgment.
Strategic prioritization
Ten to fifteen factors typically drive recommendation patterns. The report identifies which three to prioritize first.
What You Get
Eight deliverables,
48 hours.
Everything is specific to your company and your category. The goal of each deliverable is not just to show you what is happening. It is to tell you what to do about it.
Category Placement Report
What category each AI system puts you in, and whether it is the right one. If they are filing you in an adjacent category, you will not surface on the queries your buyers are actually asking.
Competitive Positioning Analysis
Who appears instead of you, how consistently, and with what description. Most founders find two or three competitors here with weaker products but stronger positioning signals. That gap has a specific cause.
Query Coverage Map
20+ real queries across all four systems. Every appearance logged, every absence noted. The pattern of where you show up, and where you disappear, is usually where the problem lives.
Source & Citation Audit
The specific sources AI systems are using to form opinions about you. Usually four or five places. Often not what you'd guess. This is typically the most actionable finding.
Positioning Gap Analysis
The gap between what you say about yourself and what AI systems actually understand you to be. When these don't match, it tells you exactly what's not getting through.
Prioritized Action Plan
A ranked list of specific changes most likely to improve your competitive positioning, based on what is actually driving the gap. Not a checklist of fixes. Changes you can act on this week.
Loom Video Walkthrough
A full walkthrough of the findings and the action plan. So you can rewatch it, share it with your team, or come back to it in 60 days when you're running the follow-up.
30-Minute Review Call
A live call to work through the findings. Most useful for questions about the action plan and deciding what to tackle first.
The Process
From request
to report in 48 hours.
The analysis is thorough, but the timeline is fixed. Every report follows the same process, delivered within the same 48-hour window.
Request your report
Book a brief intake call via Calendly. Share your company URL, a short description of your market position, and your primary competitors.
Analysis begins
We run your company across four AI systems using 20+ targeted queries and a full competitor comparison. Every query is logged, analyzed, and cross-referenced.
Report compiled
Findings are compiled into all eight deliverables. The Loom walkthrough is recorded. Everything is checked against the raw query outputs before delivery.
Delivered and reviewed
You receive the complete report package. We then schedule your 30-minute review call to walk through findings and answer questions.
Pricing
$199
One-time. No retainer, no subscription, no ongoing fees. A follow-up report in three to six months is a separate engagement.
Category Perception Analysis across 4 AI systems
Competitive Discovery Analysis: who appears and how often
AI Recommendation Mapping across 20+ queries
Source and Citation Analysis
Positioning Gap Analysis
Strategic Action Plan
Loom video walkthrough
30-minute review call
Questions
Common
questions.
The next category leaders won't just build better products. They'll become the ones AI trusts enough to recommend.
See exactly how AI describes you today, and what would change it. Delivered in 48 hours.
$199 one-time
Newsletter
How AI discoverability works
Occasional dispatches on GEO, AI category positioning, and the signals that shift how AI systems recommend companies. No noise.