I agree Tom Stoneham (npub1ttd…tvc8) . It's rather ambiguous what these two PWC reports are saying. Since we don't get to see their methodology, it's not very meaningful.
What strikes me as very odd is the use of the term *productivity*. Assuming they are not using this casually as a figure of speech, but as a technical term that through calculation has a specific value.
AFAIK, operations and production managers usually calculate productivity as either Single Factor Productivity or Total/Multi-Factor Productivity (i.e. output over input). The former is often used to compare change in processes based on partial measures of input or output (e.g. an individual employee with AI vs without, thereby excluding the effects of other input variables in this case).
While single factor measures of productivity are probably very useful for minor changes to processes with a lot of known elements, it might be more problematic for obtaining reliable numerical results when dealing with more fundament changes. (BTW, we are not even talking about wider socio-technical, societal, communal, and work relation aspects. It's purely about calculating the totality of output over input).
One more point that seems confusing to me is the claim about the increases in speed when looking at the table on potential increase in operating profit margin ([Greenstein et al. 2024, p. 3](https://www.pwc.com/gx/en/issues/technology/path-to-generative-ai-value.html )). Looking at the industry sectors with the highest potential increases (and yes also with the highest levels of uncertainty), they are not usually associated with a prime performance objective of speed. I think these sectors are usually more associated with prime performance objectives such as quality, flexibility or dependability.
So what can we make of these reports?