We propose a novel way to track inflation dynamics by identifying supply and demand shocks at a highly disaggregated level using electronic payments data. We estimate SVAR models and group historical decompositions at the product level into categories of the CPI. Our approach differs from others by explicitly estimating the shocks and retrieving their time-series dynamics. This information is valuable for monetary policy design, as it allows us to assess: (i) the type of shock driving any inflation category, (ii) whether shocks are generalized or driven by large shocks to specific items, and (iii) how the shocks evolve over time. An application to Chile suggests three distinct phases of inflation dynamics since COVID-19. In 2020, negative supply and demand shocks nearly offset each other. In 2021, demand shocks were boosted by massive liquidity injections. In 2022, global supply shocks introduced additional pressures on top of already elevated inflation.