Frequently Asked Questions

Cooper asks the same buyer-oriented questions through subscribed ChatGPT, Gemini, Claude and Perplexity web interfaces using fresh sessions and geographically distributed collection. Each question and assistant is observed 6–12 times, with more observations collected when answers drift. Verified positions are averaged within each cell before regional and assistant weights produce the category ranking. Read the methodology →

Cooper tracks ChatGPT, Gemini, Claude and Perplexity through authenticated subscriptions in their consumer web interfaces. Queries use fresh controlled sessions and a geographically distributed sample, which keeps the prompt and session conditions consistent while preserving interface-level search, routing, answer formatting and citation behavior.

For every software category, Cooper asks which products are best, rising, most affordable, and easiest to set up. Best, Most affordable, and Easiest to set up are the three core buyer intents combined into Overall AIQ. Rising is measured and normalized the same way but receives its own Rising AIQ and does not affect the Overall board. Competitor questions are part of the Cooper dashboard, where each vendor configures its own competitor set.

Each appearance receives a reciprocal position value, so a product named first contributes more than one named second. Cooper averages the 6–12 observations inside each question, assistant and region cell so variable cells do not receive extra influence, then applies the published regional hierarchy and assistant weights. The three core questions produce Overall AIQ, while the Rising question produces standalone Rising AIQ. Equal scores resolve by best placement and then total verified appearances.

AIQ is Cooper's 1 to 100 score for how strongly AI assistants recommend a product within its category and selected question scope. Overall AIQ combines Best, Most affordable, and Easiest to set up. Rising AIQ measures only the Rising question and does not contribute to Overall. A score of 100 means every tracked assistant named the product first for every question in that scope; the score is recalculated from raw measurements every week and never edited by hand. See how AIQ is calculated →

Cooper recalculates the index weekly using the newest snapshot. Earlier snapshots remain in the record, allowing the site to show movement and identify products that appear for the first time. A weekly change reflects a difference between observed snapshots rather than a judgment added by Cooper.

The label shows how many of the four tracked assistants named a product in the newest snapshot. It is a convergence measure rather than a separate score, so “4 of 4” means all four assistants included that product somewhere in their answers.

Cooper matches each name in an assistant's response to a real software product and vendor. Names that cannot be verified are excluded rather than being allowed to influence a ranking. This process also keeps spelling variations and ambiguous names from being treated as separate products.

Citations identify webpages used in the web-grounded portion of an assistant's response. Product-level cited pages are the subset Cooper verified as naming or representing that product; aggregate totals retain unattributed and inaccessible pages. They help show the evidence trail around a recommendation, but they do not expose model training data or establish the exact reason a product was included.

No. Cooper publishes verified weekly editions with their measurement window, methodology version, coverage, aggregate rankings, source summaries and correction history. Raw responses, session metadata, exact prompts, adaptive thresholds and anti-manipulation controls are retained under restricted access for quality assurance and controlled independent audits, rather than released in a form that could compromise accounts or allow the collection process to be gamed. Read the disclosure policy →

Cooper's current taxonomy covers 80 B2B software categories organized under Sales, Marketing, Support, HR, Finance, Productivity, IT & Security, and Engineering. Each top-level group contains ten software markets, and every market has its own ranking. The taxonomy can expand as Cooper's coverage grows.

Vendors cannot buy placement, manually adjust a score, or remove an unfavorable weekly result. Paid tools may provide deeper analysis and history, while the public ranking continues to reflect the same measured recommendations for every company.