Tackling climate change is one of the biggest challenges of today. Limiting climate change translates to drastically cutting carbon emissions to net zero as soon as possible. More and more commitments have been made by various authorities and companies to mitigate their GHG emissions accordingly, notably the Paris Agreement in 2015 that sets the ’well-below 2°C’ target. These energy targets generate the so-called ’transition risks’ and has impulsed a new type of financial risks assessment exercise: Climate Stress-Tests. However, the tools for these Stress-Tests remain limited. We propose a model that accounts for companies’ business model evolution in a given transition scenario for credit risk stress testing. Our model represents a single firm’s business model employing probabilistic modeling. We use stochastic control to derive the company’s intensity reduction strategy, as well as the resulting sales revenues and total emissions. We solve the minimization program using a numerical resolution method that we call Backward Sampling. We find that the intensity reduction strategy that would consist in following the same decrease rate as the sector inflates the company’s costs (up to 15.7% more expensive than the optimal strategy). Moreover, we show that investing the same amount as the total carbon cost paid at a given date is limited by its lack of a forward-looking feature, making it unable to provide a buffer for future carbon shocks in a disorderly transition scenario.