Newark, DE, Dec. 10, 2025 (GLOBE NEWSWIRE) — AI platforms returned incorrect pricing or feature information for B2B software products in 62% of simulated buyer queries, according to a new study from Kodec AI. The firm analyzed over 200 query cycles across Series B+ SaaS companies in the tech and financial services sectors.
A buyer’s first impression: AI-generated pricing pulled entirely from third-party reseller listings.
The research team tested leading AI search platforms using automated engines to simulate common buyer questions, including pricing lookups, feature comparisons, and product capabilities. The study found AI platforms frequently ignored official company sources. Instead, they prioritized outdated third-party content, reseller listings, and competitor-written “Best Alternatives” articles.
The “Rogue Sales Rep” Problem
The report categorizes this behavior as the “Rogue Sales Rep” problem, where AI models deliver wrong information without a verified source of truth. The data showed three primary failure types:
- Revenue Undercutting: AI quoted discontinued “Free Tier” limits. This sets buyer expectations at zero cost before sales teams can engage.
- Conflated Data: Models pulled pricing from third-party cloud marketplace reseller listings rather than official pricing pages.
- Fabricated Features: Platforms cited competitor-written ‘Best Alternatives’ articles, attributing capabilities to the wrong products.
“These are revenue leaks, not minor glitches,” a Kodec AI spokesperson said. “When an AI agent misquotes enterprise pricing, companies lose deals.”
The Fix: Machine-Readable Data
The study points to a specific technical gap: businesses lack machine-readable knowledge graphs that AI systems can treat as authoritative. Without structured data, models default to scraped web text.
This issue carries more weight as the web becomes “agentic.” In this phase, AI tools execute tasks on behalf of users, such as comparing vendors or initiating procurement. In that environment, unverified data can disqualify vendors from consideration entirely.
To secure this visibility, Kodec AI offers Search Infrastructure that allows enterprises to audit their AI presence and implement governed data structures that AI platforms can source reliably.
About Kodec AI
Kodec AI is an engineering firm building the search infrastructure of the future. Founded by engineers with deep experience in vector databases and high-scale retrieval systems, Kodec helps brands turn AI search into a predictable, controlled revenue channel.
Press inquiries
Kodec AI
https://www.kodec.net
Ganga Kharka
ganga@joinkodec.net
131 Continental Dr, Suite 305
A video accompanying this announcement is available here: https://youtube.com/watch?v=bdz2cqCFSIA
