Help
How scoring works, how to claim a profile, and how to read the leaderboard.
Verified Layer is an independent software intelligence organization that publishes monthly AI rankings for software vendors. We run 53 structured prompts across ChatGPT, Claude, and Gemini simultaneously, aggregate the responses across 8 scored dimensions, and publish a 0–100 Trust Score for every software vendor in 100+ categories. No paid placements. No vendor input. No analyst opinions. Just what AI already knows.
G2 and Gartner were built for the Google search era. G2 charges vendors up to $34,000 per year for better placement. Gartner charges vendors for analyst briefings that influence their reports. Both are structurally dependent on vendor revenue to survive, which means neither can be fully independent. Verified Layer uses AI as the judge, not as a tool. We do not ask vendors to submit reviews, pay for placement, or brief our analysts. We ask ChatGPT, Claude, and Gemini what they already know, and publish that answer independently. A vendor cannot buy a better score. They can only earn one.
Profound and Peec AI are marketing tools sold to vendor marketing teams. They help vendors monitor and improve what AI says about them; they are the PR agency. Verified Layer is the ratings agency. We publish what AI already knows before any optimization, the same way a bond rating is published before a company hires a PR firm. Vendors cannot purchase a better Verified Layer score. They can only earn one by genuinely improving their product, community presence, and evidence base.
Verified Layer is an independent organization. We are not funded by software vendors, not funded by advertising, and not owned by any software company with a conflict of interest. Vendors pay $150/month to access their diagnostic report, not to improve their score. The payment is completely isolated from the scoring pipeline by architectural design.
Three reasons. First, AI is already the primary discovery channel buyers use: if ChatGPT, Claude, and Gemini collectively know something about a vendor, that knowledge is exactly what determines whether buyers shortlist them. Second, three independent platforms built by three competing companies (Anthropic, OpenAI, and Google) are harder to simultaneously deceive than any human analyst team. Third, the entire methodology is published publicly: every prompt, every formula, every quality gate. You do not have to trust us. You can inspect exactly what we do.
The Trust Score is produced by running 159 LLM calls per vendor per month: 53 structured prompts × 3 platforms (Claude, ChatGPT, Gemini). Responses are scored across 8 dimensions, weighted by their influence on real enterprise purchasing decisions, and aggregated using inter-rater reliability methodology (Cronbach’s alpha). When all three platforms agree, the score is trusted. When they disagree, a conservative calculation is applied.
Product Quality (18%), Market Reputation (16%), Community Sentiment (15%), Customer Experience (14%), Technical Credibility (13%), Value for Money (11%), Innovation and Vision (8%), and Buying Safety (5%). All 8 sum to exactly 100%. Weights reflect how much each dimension influences actual enterprise purchase decisions based on buyer research.
Quality Score is rate-based and size-agnostic: it measures how good the product actually is, independent of company size. A four-person startup can score identically to a 4,000-person incumbent on Quality. Presence Score is scale-based: it measures absolute maturity signals like analyst recognition, review volume, and enterprise customer count, which naturally correlate with company size. The two axes are always shown separately and never blended into a single combined number.
A score of 74 means the weighted average of what Claude, ChatGPT, and Gemini collectively know about this vendor across 8 dimensions, after applying the evidence-first methodology and inter-rater reliability checks, is 74 out of 100. It does not mean the vendor is good or bad in absolute terms; it means they are positioned in the upper-mid range of verifiable public evidence. Scores above 75 with Presence above 65 qualify for the LEADER badge.
Yes, scores can reflect incomplete evidence if the vendor has a limited public presence, or if AI training data contains outdated information. Every score with low cross-platform agreement or insufficient evidence is published with a flag (Limited Evidence or Partial Evidence) so buyers understand the confidence context. A vendor who believes a specific data point is factually incorrect can submit a Score Correction.
Scores update monthly. Changes can be driven by new community discussions appearing in AI training updates, new evidence items verified or removed, changes in community sentiment across Reddit, GitHub, Glassdoor, or review platforms, or a Score Correction accepted in the prior cycle. Vendors with an Insights Utility subscription receive a detailed change report on the 2nd of each month explaining exactly which dimensions moved and why.
The range is 0–100. In practice, most active software vendors score between 35 and 90. Scores below 30 indicate very limited public evidence or significant documented problems. Scores above 85 are rare and require strong cross-platform agreement from all three AI judges across most dimensions.
Always integers. No decimals. The methodology explicitly prohibits decimal scores because false precision (a score of 73.6 vs 74) implies a measurement accuracy the evidence does not support.
Search for your company name on the leaderboard. If you appear, click "Claim Your Profile" on your vendor page. Claiming is free and does not change your score. It gives you the ability to submit score corrections and access the basic dashboard view.
No. Claiming your profile has zero effect on your score. The scoring engine runs independently and does not know which vendors have claimed profiles and which have not. Claiming is a one-way action: you gain access to see your score context. The score does not gain access to anything you submit.
Full diagnostic dashboard showing your raw dimension scores (D1–D8) and sub-scores, community signal breakdown across all 20 tracked channels, product evidence status across all 14 evidence items, competitor score comparison within your primary category, 2 score correction submissions per scoring cycle, and monthly score change notifications with specific dimension-level explanations.
No. Subscribing shows you the diagnostic intelligence behind your score. It does not change the score. The payment is architecturally isolated from the scoring pipeline: the engine that generates scores has no knowledge of which vendors are subscribers. This isolation is fundamental to Verified Layer’s independence and is documented in the Terms of Service.
Start with the diagnostic dashboard (Insights Utility, $150/month). It will show you exactly which dimensions are suppressing your score and which specific evidence items are missing. Most low scores come from three sources: limited community presence, missing security certifications, or insufficient independently verifiable customer references.
Subscribers can submit up to 2 Score Corrections per monthly scoring cycle through the dashboard. Each correction must identify the specific dimension (D1–D8) being disputed, the factual basis for the dispute, and a URL or documentary evidence supporting the correction. Accepted corrections trigger a targeted re-run of the relevant dimension within 48 business hours.
Corrections are for factual errors: a certification you hold that was not found, a customer reference that was miscategorized, a specific data point that is verifiably wrong. Corrections cannot change your overall ranking position (a mathematical output), scores derived from community sentiment that accurately reflects public opinion, LLM query responses (which reflect what AI knows from training data), or your category assignment.
A vendor that has not claimed its listing and is not a subscriber may request removal from the public leaderboard by submitting a written request from a company email domain matching the vendor’s registered domain. Removal from the public leaderboard does not prevent Verified Layer from continuing to score the vendor internally or from referencing the vendor in aggregate or anonymized analysis.
Yes. Vendors with an Active badge status are granted a limited, non-exclusive, non-transferable license to embed the Verified Layer Trust Badge on their website and marketing materials. The badge auto-updates when scores change: you embed it once and it always reflects your current status. The badge must not be modified, recolored, or used after it has been Revoked.
The Intent Marketplace is a buyer lead channel available to Insights Utility subscribers. Buyer traffic driven to the Verified Layer platform through embedded badges and leaderboard discovery is packaged as verified, high-intent leads and sold at $40 per lead. Each lead represents a buyer who visited Verified Layer, expressed specific category purchasing intent, and met Verified Layer’s verification criteria. No subscription commitment: pay only for leads delivered.
Category sizes vary. Major categories like CRM and Marketing Automation have 150+ scored vendors. Niche categories may have 20–40. You can see the full count on the category leaderboard page. New vendors are added to the database monthly as the scoring pipeline expands.
The scoring engine targets vendors with a verifiable public presence: a website, some community activity, and at least one public signal that allows the AI platforms to evaluate you. If you believe your company should be in the database and is not, submit a request through our contact page. We will add you to the next monthly scoring cycle if you meet the minimum presence threshold.
Yes. The public leaderboard is completely free for buyers. All vendor scores, category ranks, badge statuses, and trend data are publicly accessible at no cost. No account required. No email capture required. No paywall.
The rankings are independent of vendor payment by structural design, not by policy. The scoring engine calls AI APIs directly and has no knowledge of which vendors have paid for anything. A vendor paying $150/month for the diagnostic subscription receives the same public score as a vendor who has never heard of Verified Layer.
Start with the category leaderboard for the software type you need. Sort by Quality Score to see which vendors have the strongest independent evidence of product quality. Sort by Presence Score to see which vendors are most established in the market. Use the competitor comparison on individual vendor profiles to do a side-by-side assessment of your shortlisted candidates.
Every month on the 1st. All leaderboards and vendor profile pages update simultaneously. Score change notifications go to subscribed vendors on the 2nd. Buyers always see the most recently scored data.
A LEADER badge means the vendor has a Quality Score of 75 or above AND a Presence Score of 65 or above in the current month’s scoring cycle. Both conditions must be met simultaneously. It is not a lifetime award; it is earned fresh each month from the same evidence-first process every other vendor goes through.
RISING is added to any badge when a vendor’s Momentum Score reaches +15 or above, meaning their Quality Score has improved significantly over the prior two quarters. A LEADER + RISING vendor is both at the top of the category and improving.
Three independent platforms, Claude (Anthropic), ChatGPT (OpenAI), and Gemini (Google), are built by three competing companies with three genuinely different training pipelines. If all three agree about a vendor, that agreement is meaningful. If one platform scored very differently, the inter-rater reliability calculation detects and responds to the disagreement conservatively. Using one platform would mean one company’s knowledge and biases determine every score.
Microsoft Copilot is built on the same underlying OpenAI technology that powers ChatGPT, making it the same judge wearing a different name, not an additional independent opinion. Perplexity is primarily a retrieval interface rather than a model with independent training. We require three genuinely distinct companies with genuinely distinct training pipelines: Anthropic, OpenAI, and Google.
Every prompt in the methodology requires the AI platform to collect and state evidence from public sources before considering any score, declare a confidence level (HIGH, MEDIUM, or LOW) based on that evidence before writing the score, and then assign the score constrained by what the evidence actually supports. An AI platform that assigns the score first and then generates supporting evidence is producing rationalization, not assessment.
A subset of prompts are specifically designed to find weaknesses, failures, complaints, and risks. A vendor who has manufactured only positive content will score well on standard prompts and poorly on inverse prompts. Both types feed the final score, which means silence about problems is not rewarded; the system actively searches for problems.
Zero would mean the vendor is terrible. One hundred would mean the vendor is excellent. Neither is correct for a vendor with no community presence; unknown is genuinely unknown. A score of 40 positions the vendor below the median (50) without penalizing them for being niche, and without rewarding them for silence.
Cronbach’s alpha measures how consistently the three AI platforms agree about which evidence is strong versus weak for a given vendor dimension. When all three agree (alpha above 0.80), the simple weighted mean of all scores is used. When they disagree sharply (alpha below 0.55), the system uses the 25th percentile of all scores rather than the average: a conservative score, not a papering-over average.
Four automated checks run before every score is published: inter-rater reliability (determines the aggregation method), evidence spot-check (validates citations against a live web search), month-over-month change gate (damps extreme score swings), and LOW confidence ratio (caps the overall score if too many individual calls returned LOW confidence). Scores that fail any check are still published (never suppressed) but with specific flags visible to buyers.
Nobody, the same as other independent ratings bodies. Authority is earned through methodology transparency, consistent application, and market trust, not through a certificate from a higher body. We are pursuing ISO 17065 accreditation, the international standard for conformity assessment bodies, a process verification confirming the scoring process is independent and reproducible, not a government endorsement of our scores.
Algorithm changes affect the evidence AI platforms find about vendors. When that happens, scores update to reflect the new state of knowledge, which is the intended behavior. A vendor whose score drops after an algorithm change should investigate which dimension changed and why. Subscribers receive a monthly change report that explains exactly what moved.
We run across three platforms simultaneously. If one restricts structured queries, we score with the remaining platforms. Longer term, we plan to train a proprietary scoring model that can replicate the multi-platform consensus from community signals without calling external APIs, providing independence from any single platform’s decisions.
The methodology contains multiple structural anti-gaming controls. A share of prompts explicitly search for weaknesses: manufacturing positive content does not help on inverse prompts. Absence of community signals scores 40, not 100; silence provides no advantage. Marketing claims without independent attribution and self-attested certifications score 0. Multi-platform inter-rater reliability means gaming one platform’s outputs does not help if the others disagree. The fastest path to a better score is genuinely improving the product and company.
No. Verified Layer scores any software vendor with a publicly accessible online presence, regardless of whether they have registered on the platform or consented to being scored. This scoring activity constitutes algorithmic analysis of publicly available information, the same legal basis that applies to every rating agency, research firm, and analyst organization that publishes scores about companies without their advance consent.
Verified Layer collects publicly available information about software companies: community discussions, review platform data, certification registry information, press coverage, and the responses AI platforms provide when queried about vendors. We do not collect or store private or personal data about individual employees of scored vendors.
Subscriber diagnostic reports, including raw query breakdowns, gap analysis, correction history, and evidence item detail, are confidential to the subscribing vendor and will not be disclosed to any third party, including other vendors on the platform. Aggregate scores, dimension summaries, and category rankings derived from public signals are not confidential and may be displayed on public-facing leaderboards.
If your company is established in the European Economic Area or United Kingdom and you believe Verified Layer’s processing of data about your company engages your rights under the GDPR or UK GDPR, please refer to the Privacy Policy and submit a data subject request through our contact page. The right to erasure under GDPR applies to personal data, not to algorithmic scores about legal entities.
Start with a Score Correction: if a specific data point is factually incorrect, submit a correction with documentary evidence. If the issue is not a factual error but a dispute about the methodology or the independence of the process, submit a formal Score Dispute through our contact page. Legal action should be a last resort; the correction and dispute processes exist specifically to resolve factual disagreements efficiently and fairly.
Three revenue streams: Insights Utility subscriptions at $150/month give vendors access to their diagnostic report, Intent Marketplace leads at $40/lead sell verified buyer intent data to subscribed vendors, and Data API licensing (from Year 2 onward) makes the scoring dataset available to enterprise procurement tools and data platforms. Gross margin is approximately 87%: the scoring engine costs approximately $0.40 per vendor per month regardless of subscription status.
Global enterprise software spend is projected at $1.4 trillion in 2026. Buyers have moved from Google to AI for software research; the discovery channel has changed, but the market has not shrunk. There are approximately 150,000+ software vendors across major review platforms combined, of which roughly 18,000 are active (5+ reviews, 10–500 employees) and form the primary addressable market for the Insights Utility subscription at launch.
Profound and Peec AI are the two best-funded companies in the adjacent AI visibility space. Both are vendor-serving marketing tools: they help vendors improve how AI perceives them. Verified Layer is the independent ratings organization: we publish what AI already knows. The distinction is structural and permanent. Neither competitor can become independent without destroying their revenue model.
The target is based on scoring 10,000 vendors by Month 12 at a 5% conversion rate to the $150/month Insights Utility subscription (500 subscribers × $150 = $75,000/month subscription revenue) plus 350 leads/month from the Intent Marketplace ($14,000/month). Break-even is at approximately 65 subscribers (Month 3–4). Gross margin holds at approximately 87% at every volume point. The full financial model and assumptions are available in the investor brief on request.
Press and investor enquiries can be sent through our contact page, and we will route your message to the right team.
If your question is not answered here, contact us. We respond to every message.