Claude and Claude Code supply different responses to the aforesaid prompts, based on new data from Profound, a institution that monitors really brands look successful AI answers.
In their tests, Claude Code utilized web hunt successful 13% of responses, while Claude did truthful successful 93%. On average, the brands mentioned for the aforesaid punctual overlapped by astir 20%. Profound sells AI visibility search and promotes its level successful the report.
How The Responses Differ
Between July 13 and July 23, the institution tested 1,724 prompts connected some products, generating 24,135 responses pinch web hunt enabled for both. Interestingly, repetition runs from the aforesaid merchandise were much accordant than the 2 products were pinch each other.
For example, 2 responses from Claude to the aforesaid punctual shared astir half of their brands, while 2 responses from Claude Code shared astir 40%.
Despite searching little often, Claude Code mentioned astir the aforesaid number of brands per response, 6.6 compared to Claude’s 5.2. For coding prompts, Claude leaned toward codification editors and IDEs, whereas Claude Code much often surfaced code-quality and workflow tools.
Additionally, Claude Code’s responses tended to beryllium shorter and much structured, averaging 322 words versus Claude’s 459. Over half of Claude Code’s responses included a table, compared to 11% for Claude.
What Each Agent Reads
The 2nd dataset shows the apical 1,000 pages visited by each product’s supplier betwixt July 18 and August 18, crossed galore domains the institution tracks internally. Almost three-quarters of Claude Code’s observed visits went to documentation, informational, and pricing pages, while only 5% of Claude’s visits did the same.
In contrast, 60% of Claude’s visits were to robots.txt files, sitemaps, and homepages, compared to conscionable 4% for Claude Code. The institution interprets this arsenic Claude’s supplier exploring what a tract contains, whereas Claude Code’s supplier goes aft circumstantial specifications connected known page types.
For specified pages, the report’s proposal is to authorities clear facts for illustration “Supports Python 3.10–3.13, Node.js 20+, and Go 1.22+,” and lead pinch the reply nether question-shaped headings.
The institution mentions that an AI exemplary branded the page types, but the study doesn’t specify immoderate quality reappraisal aliases explicate really they distinguished betwixt the 2 agents’ traffic.
Why This Matters
The information suggests measuring Claude Code separately from Claude erstwhile your customers are utilizing it. In 1 dataset, agents displayed different page-visit patterns, and successful another, marque mentions overlapped by astir 20% connected average.
The report’s authors propose intelligibly listing method and pricing specifications connected pages visited by Claude Code’s agent, though they don’t explicitly trial whether these changes boost marque mentions aliases visibility.
Looking Ahead
AI products pinch shared models mightiness require their ain measurement arsenic they pull meaningful personification traffic. Claude Code is 1 illustration of that type of product. The chat astir page visits is based connected 1 period of supplier postulation crossed the domains Profound monitors internally.
Featured Image: Cast Of Thousands/Shutterstock
English (US) · 
Indonesian (ID) · 
