The agency analyzes the issuer's financial condition, industry conditions, debt structure, and management quality, then assigns a rating on its scale and accompanies it with a forecast. The rating is subject to regular review.
The conflict of interest built into the model should be understood: in most cases, the rated issuer itself pays for the rating assignment. This is offset by reputation mechanisms and regulation of agency activities, but is not fully eliminated.
There are international agencies working on a global scale and national ones assessing issuers within a single country. The scales are not directly comparable: the same letter designation means different levels of risk.
For an investor, a rating is useful as a starting point, not as a conclusion. What matters is the assignment date, forecast, direction of recent changes, and which agency issued it. Special caution is required for issues with speculative-grade ratings: their high yield reflects the risk assessment, not an opportunity missed by the market.