When salespeople recommend products, which information sources should they rely on?

When an AI agent recommends a certain tool, application programming interface, software as a service product or service, it almost never makes decisions based on a single accurate information source. It may come from: \- Official websites \- Documentation \- Price pages \- User reviews \- Forums \- GitHub activities \- Third-party rankings \- Social posts \- Historical usage data \- Information submitted by developers However, each source has its own limitations. The official website might be overly glorified to the point of being completely useless. Reviews can become outdated. Forums can be noisy but lack representativeness. Prices can change. Rankings can be bought and sold. GitHub activities can tell you some information, but it is not entirely the same as the actual product quality reflected. So the real question is not just "What does the agent know?" It is: How does the agent decide who to trust? Should different information sources have different weights? Should recommendations come with relevant evidence to support them? Should the agent mark information as outdated, incomplete, or biased towards business when it does? Then, are there any methods that can effectively assess the reliability of the information, and these methods should not become another form of search engine optimization? I'm curious about how others will design this layer.