The continued claims chart I posted shows every recession that occurred since 1967. This is as long as this data series goes. The BLS started publishing the series in 1967. The CC series includes 59 years of market history, and all the latest recessions. And, you can see that every single recession since 1967 was predicted and coincidental with the Continued Claims trend going up. And there was no CC uptrend without a recession, and no recession without CC uptrend.
So, if Google says what you quoted above, I am curious to know what recession(s) is it referring to, cause it didn't happen in the last 60 years since the CC data was created and published. Before 1967, this data was not available to use as a quantitative indicator of unemployment. I wouldn't trust that Google vague statement without a specific quantitative example.
Also, look up the NBER definition of recession: basically economic contraction with rising unemployment. So, in a recession, continued claims rise by definition. Without CC rising, NBER won't declare a recession. For instance, in 2022, as the CC didn't change to trend up, we didn't have a recession, though the economy contracted.
So, to identify, check the trends of both S&P AND CC. If CC is on trend up and S&P on trend down, we will be having a recession.
How do you identify and qualify trends? I would use a long term MA like 12 months on a monthly chart. Just one method to identify trend changes.